{
  "dataset": "benchmark-leaderboard",
  "title": "OpenRouter benchmark leaderboard",
  "description": "One row per benchmark, model version, and provider: the aggregate values displayed on the OpenRouter leaderboard pages (GPQA Diamond, tau2-bench airline). Derived from the benchmark-runs dataset.",
  "schema_version": "v1",
  "snapshot_date": "2026-10-10",
  "retrieved_at": "2026-10-10T01:45:23.153Z",
  "update_cadence": "daily",
  "license": {
    "name": "CC BY 4.0",
    "url": "https://creativecommons.org/licenses/by/4.0/"
  },
  "attribution": "Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter.",
  "harness": {
    "name": "OpenRouter benchmark harness",
    "url": "https://github.com/OpenRouterTeam/benchmark-harness"
  },
  "source_pages": [
    "https://openrouter.ai/benchmarks/gpqa-diamond",
    "https://openrouter.ai/benchmarks/tau2-bench-airline"
  ],
  "methodology": "Rows aggregate the benchmark-runs dataset exactly as the leaderboard pages do, over each benchmark's non-zero-score runs from the newest 90 days: accuracy is the task-weighted mean of per-run accuracy, accuracy_std_dev is the population standard deviation of per-run accuracy (null with a single run), and the per-task averages divide summed cost, generation time, and output tokens by summed tasks. Rows are ordered by accuracy descending, then model name.",
  "readme_url": "https://openrouter.ai/api/v1/datasets/exports/README.md",
  "columns": [
    {
      "name": "benchmark",
      "type": "string",
      "nullable": false,
      "description": "Benchmark identifier."
    },
    {
      "name": "benchmark_page_url",
      "type": "string",
      "nullable": false,
      "description": "openrouter.ai leaderboard page that displays this row."
    },
    {
      "name": "model_permaslug",
      "type": "string",
      "nullable": false,
      "description": "Immutable OpenRouter model version identifier."
    },
    {
      "name": "model_slug",
      "type": "string",
      "nullable": true,
      "description": "Current OpenRouter model slug, null when the version is no longer listed."
    },
    {
      "name": "model_name",
      "type": "string",
      "nullable": false,
      "description": "Display name as shown on the leaderboard page."
    },
    {
      "name": "provider_name",
      "type": "string",
      "nullable": false,
      "description": "Inference provider that served the aggregated runs."
    },
    {
      "name": "accuracy",
      "type": "number",
      "nullable": false,
      "description": "Task-weighted mean accuracy across runs, 0 to 1."
    },
    {
      "name": "accuracy_std_dev",
      "type": "number",
      "nullable": true,
      "description": "Population standard deviation of per-run accuracy; null when only one run exists."
    },
    {
      "name": "run_count",
      "type": "integer",
      "nullable": false,
      "description": "Number of runs aggregated."
    },
    {
      "name": "total_questions",
      "type": "integer",
      "nullable": false,
      "description": "Tasks attempted across all aggregated runs (sample size)."
    },
    {
      "name": "avg_cost_per_task_usd",
      "type": "number",
      "nullable": true,
      "description": "Summed run cost divided by summed tasks, in USD."
    },
    {
      "name": "avg_duration_per_task_ms",
      "type": "number",
      "nullable": true,
      "description": "Summed generation time divided by summed tasks, in milliseconds."
    },
    {
      "name": "avg_output_tokens_per_task",
      "type": "number",
      "nullable": true,
      "description": "Summed completion tokens divided by summed tasks."
    },
    {
      "name": "last_run_timestamp",
      "type": "string",
      "nullable": false,
      "description": "ISO 8601 UTC timestamp of the most recent aggregated run."
    }
  ],
  "row_count": 2157,
  "rows": [
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.8-flash-20260902",
      "model_slug": "google/gemini-3.8-flash",
      "model_name": "Google: Gemini 3.8 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.95637,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 191,
      "avg_cost_per_task_usd": 0.06822796858638744,
      "avg_duration_per_task_ms": 132197.8062827225,
      "avg_output_tokens_per_task": 18049.068062827224,
      "last_run_timestamp": "2026-10-01T08:31:33.481Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-pro-20260903",
      "model_slug": "openai/gpt-6-astra-pro",
      "model_name": "OpenAI: GPT-6 Astra Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.954545,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.332555398989899,
      "avg_duration_per_task_ms": 75446.5505050505,
      "avg_output_tokens_per_task": 4530.388888888889,
      "last_run_timestamp": "2026-10-01T08:27:35.414Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.8-flash-20260902",
      "model_slug": "google/gemini-3.8-flash",
      "model_name": "Google: Gemini 3.8 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.952862,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.07423223737373738,
      "avg_duration_per_task_ms": 86748.28787878787,
      "avg_output_tokens_per_task": 19646.409090909092,
      "last_run_timestamp": "2026-10-01T08:29:11.347Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "Google",
      "accuracy": 0.95202,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20364760606060606,
      "avg_duration_per_task_ms": 135706.18686868687,
      "avg_output_tokens_per_task": 16847.588383838385,
      "last_run_timestamp": "2026-10-01T19:14:26.157Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-pro-20260903",
      "model_slug": "openai/gpt-6-astra-pro",
      "model_name": "OpenAI: GPT-6 Astra Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.9511785,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.33620659090909094,
      "avg_duration_per_task_ms": 70544.52272727272,
      "avg_output_tokens_per_task": 4588.830808080808,
      "last_run_timestamp": "2026-10-02T23:23:01.262Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-pro-20260903",
      "model_slug": "openai/gpt-6-astra-pro",
      "model_name": "OpenAI: GPT-6 Astra Pro",
      "provider_name": "Azure",
      "accuracy": 0.949495,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.35495546717171716,
      "avg_duration_per_task_ms": 72102.17676767676,
      "avg_output_tokens_per_task": 4622.60101010101,
      "last_run_timestamp": "2026-10-01T08:29:03.480Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "Google AI Studio",
      "accuracy": 0.9483723333333332,
      "accuracy_std_dev": 0.004827144451490511,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.20306945454545455,
      "avg_duration_per_task_ms": 134835.03535353535,
      "avg_output_tokens_per_task": 16798.409090909092,
      "last_run_timestamp": "2026-10-01T19:12:42.155Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.947811,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.021931015151515153,
      "avg_duration_per_task_ms": 68600.9191919192,
      "avg_output_tokens_per_task": 1835.4242424242425,
      "last_run_timestamp": "2026-10-03T01:13:05.002Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "OpenAI",
      "accuracy": 0.9472503333333333,
      "accuracy_std_dev": 0.0015872190215034077,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.11162585016835018,
      "avg_duration_per_task_ms": 48234.978114478115,
      "avg_output_tokens_per_task": 2076.191919191919,
      "last_run_timestamp": "2026-10-02T23:16:05.967Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.8-flash-20260902",
      "model_slug": "google/gemini-3.8-flash",
      "model_name": "Google: Gemini 3.8 Flash",
      "provider_name": "Google",
      "accuracy": 0.947009,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.07090916923076923,
      "avg_duration_per_task_ms": 142045.61538461538,
      "avg_output_tokens_per_task": 18762.907692307694,
      "last_run_timestamp": "2026-10-01T08:30:23.202Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sakana/fugu-ultra",
      "model_slug": "sakana/fugu-ultra",
      "model_name": "Fugu Ultra",
      "provider_name": "auto-routing",
      "accuracy": 0.9461279999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 1.148739303030303,
      "avg_duration_per_task_ms": 328779.10101010103,
      "avg_output_tokens_per_task": 33271.606060606064,
      "last_run_timestamp": "2026-09-15T01:07:47.987Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "auto-routing",
      "accuracy": 0.945286,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20203182828282829,
      "avg_duration_per_task_ms": 133595.55303030304,
      "avg_output_tokens_per_task": 16712.858585858587,
      "last_run_timestamp": "2026-10-01T19:16:14.342Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.945286,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.11987038888888889,
      "avg_duration_per_task_ms": 35572.169191919194,
      "avg_output_tokens_per_task": 2020.9419191919192,
      "last_run_timestamp": "2026-10-01T08:24:02.411Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "auto-routing",
      "accuracy": 0.944444,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.11333201010101009,
      "avg_duration_per_task_ms": 57056.792929292926,
      "avg_output_tokens_per_task": 2108.121212121212,
      "last_run_timestamp": "2026-10-01T08:25:16.514Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.944444,
      "accuracy_std_dev": 0.0027491439879836154,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.020607563973063972,
      "avg_duration_per_task_ms": 47519.37542087542,
      "avg_output_tokens_per_task": 1902.2373737373737,
      "last_run_timestamp": "2026-10-02T23:20:22.253Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "auto-routing",
      "accuracy": 0.944444,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02136650505050505,
      "avg_duration_per_task_ms": 52411.32323232323,
      "avg_output_tokens_per_task": 1978.1313131313132,
      "last_run_timestamp": "2026-10-01T08:24:51.870Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "Azure",
      "accuracy": 0.9431713315363881,
      "accuracy_std_dev": 0.006147978549897494,
      "run_count": 4,
      "total_questions": 742,
      "avg_cost_per_task_usd": 0.10982969407008086,
      "avg_duration_per_task_ms": 117613.10242587602,
      "avg_output_tokens_per_task": 1949.2371967654988,
      "last_run_timestamp": "2026-10-01T08:26:36.876Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.9427606666666666,
      "accuracy_std_dev": 0.0036369299629708685,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.029872336700336698,
      "avg_duration_per_task_ms": 28492.575757575756,
      "avg_output_tokens_per_task": 7819.474747474747,
      "last_run_timestamp": "2026-10-01T16:16:03.631Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "typesafe/jev-router",
      "model_slug": "typesafe/jev-router",
      "model_name": "TypeSafe: Jev Router",
      "provider_name": "auto-routing",
      "accuracy": 0.939394,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.015926459595959598,
      "avg_duration_per_task_ms": 24665.858585858587,
      "avg_output_tokens_per_task": 2219.030303030303,
      "last_run_timestamp": "2026-09-28T18:11:34.031Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.9393935,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.29793838636363634,
      "avg_duration_per_task_ms": 106861.7196969697,
      "avg_output_tokens_per_task": 11692.656565656565,
      "last_run_timestamp": "2026-10-01T16:10:33.208Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.938528,
      "accuracy_std_dev": 0.004357499999999959,
      "run_count": 2,
      "total_questions": 385,
      "avg_cost_per_task_usd": 0.0293307974025974,
      "avg_duration_per_task_ms": 59248.22337662338,
      "avg_output_tokens_per_task": 7683.532467532467,
      "last_run_timestamp": "2026-10-01T16:17:32.289Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "Azure",
      "accuracy": 0.9382713333333333,
      "accuracy_std_dev": 0.004627357405210382,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.022108563973063967,
      "avg_duration_per_task_ms": 32928.50084175084,
      "avg_output_tokens_per_task": 1914.9974747474748,
      "last_run_timestamp": "2026-10-01T08:40:50.203Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google",
      "accuracy": 0.937943329787234,
      "accuracy_std_dev": 0.011200313606720516,
      "run_count": 3,
      "total_questions": 564,
      "avg_cost_per_task_usd": 0.023761370567375886,
      "avg_duration_per_task_ms": 52418.78723404255,
      "avg_output_tokens_per_task": 7636.52304964539,
      "last_run_timestamp": "2026-10-01T16:16:37.307Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.9371493333333333,
      "accuracy_std_dev": 0.004199386807089248,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.30925494444444446,
      "avg_duration_per_task_ms": 152328.37037037036,
      "avg_output_tokens_per_task": 10179.092592592593,
      "last_run_timestamp": "2026-10-01T19:12:11.505Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "OpenAI",
      "accuracy": 0.9360266666666668,
      "accuracy_std_dev": 0.0036369299629708854,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.3141816919191919,
      "avg_duration_per_task_ms": 142474.1195286195,
      "avg_output_tokens_per_task": 10342.412457912458,
      "last_run_timestamp": "2026-10-02T23:25:41.614Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "Azure",
      "accuracy": 0.9337823333333333,
      "accuracy_std_dev": 0.008398773614178557,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.3842955404040404,
      "avg_duration_per_task_ms": 82918.1936026936,
      "avg_output_tokens_per_task": 11996.59090909091,
      "last_run_timestamp": "2026-10-01T16:36:19.657Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.9332206666666667,
      "accuracy_std_dev": 0.00965467639137749,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.24476150168350166,
      "avg_duration_per_task_ms": 101605.01851851853,
      "avg_output_tokens_per_task": 11508.424242424242,
      "last_run_timestamp": "2026-10-02T23:21:48.863Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Parasail",
      "accuracy": 0.932951424427481,
      "accuracy_std_dev": 0.006875320605511528,
      "run_count": 7,
      "total_questions": 1310,
      "avg_cost_per_task_usd": 0.13016139770992366,
      "avg_duration_per_task_ms": 164770.279389313,
      "avg_output_tokens_per_task": 8479.913740458016,
      "last_run_timestamp": "2026-10-09T00:00:06.212Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sakana/fugu-ultra-v2",
      "model_slug": "sakana/fugu-ultra-v2",
      "model_name": "Fugu Ultra V2",
      "provider_name": "auto-routing",
      "accuracy": 0.93266,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.5025571414141414,
      "avg_duration_per_task_ms": 194915.41919191918,
      "avg_output_tokens_per_task": 13806.227272727272,
      "last_run_timestamp": "2026-09-15T01:07:48.052Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Azure",
      "accuracy": 0.9326596250000001,
      "accuracy_std_dev": 0.00862529761714779,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.32610833964646463,
      "avg_duration_per_task_ms": 161836.85416666666,
      "avg_output_tokens_per_task": 10112.390782828283,
      "last_run_timestamp": "2026-10-01T19:14:07.472Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.9326595000000001,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02474653787878788,
      "avg_duration_per_task_ms": 20002.444444444445,
      "avg_output_tokens_per_task": 2234.5,
      "last_run_timestamp": "2026-10-03T09:05:12.762Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.9325395714285714,
      "accuracy_std_dev": 0.004401689845956867,
      "run_count": 4,
      "total_questions": 756,
      "avg_cost_per_task_usd": 0.06429072222222224,
      "avg_duration_per_task_ms": 63071.10317460317,
      "avg_output_tokens_per_task": 11529.871693121693,
      "last_run_timestamp": "2026-10-01T19:12:42.162Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Google",
      "accuracy": 0.931257,
      "accuracy_std_dev": 0.00924235327175929,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.026367194444444447,
      "avg_duration_per_task_ms": 19064.280303030304,
      "avg_output_tokens_per_task": 2249.5185185185187,
      "last_run_timestamp": "2026-10-03T09:39:59.864Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "Google",
      "accuracy": 0.930415,
      "accuracy_std_dev": 0.004827415699246697,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.06216489141414142,
      "avg_duration_per_task_ms": 64761.851851851854,
      "avg_output_tokens_per_task": 11730.429292929293,
      "last_run_timestamp": "2026-10-01T19:12:42.153Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.928451,
      "accuracy_std_dev": 0.004209000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.12326839393939394,
      "avg_duration_per_task_ms": 361758.7878787879,
      "avg_output_tokens_per_task": 20322.222222222223,
      "last_run_timestamp": "2026-10-01T16:09:08.142Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Fireworks",
      "accuracy": 0.9278394565075213,
      "accuracy_std_dev": 0.008041238261262993,
      "run_count": 25,
      "total_questions": 4587,
      "avg_cost_per_task_usd": 0.1608165642031829,
      "avg_duration_per_task_ms": 149318.05646391978,
      "avg_output_tokens_per_task": 8864.42119032047,
      "last_run_timestamp": "2026-10-03T10:48:13.331Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Makora",
      "accuracy": 0.9277417729032257,
      "accuracy_std_dev": 0.015966818205265555,
      "run_count": 5,
      "total_questions": 775,
      "avg_cost_per_task_usd": 0.09847772,
      "avg_duration_per_task_ms": 244962.84516129032,
      "avg_output_tokens_per_task": 7582.7083870967745,
      "last_run_timestamp": "2026-10-03T10:05:38.767Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.9264870000000001,
      "accuracy_std_dev": 0.004418715424192887,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.05958464141414141,
      "avg_duration_per_task_ms": 58983.43771043771,
      "avg_output_tokens_per_task": 11740.584175084176,
      "last_run_timestamp": "2026-10-01T19:11:04.522Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Relace",
      "accuracy": 0.926400137295082,
      "accuracy_std_dev": 0.007132075389394045,
      "run_count": 5,
      "total_questions": 976,
      "avg_cost_per_task_usd": 0.09721093032786886,
      "avg_duration_per_task_ms": 197311.04713114753,
      "avg_output_tokens_per_task": 8911.119877049181,
      "last_run_timestamp": "2026-10-03T10:09:17.601Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.925926,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.025347974747474744,
      "avg_duration_per_task_ms": 21521.87373737374,
      "avg_output_tokens_per_task": 2298.931818181818,
      "last_run_timestamp": "2026-10-03T09:35:22.953Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.925926,
      "accuracy_std_dev": 0.005498287975967322,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.1441699393939394,
      "avg_duration_per_task_ms": 61658.986531986535,
      "avg_output_tokens_per_task": 15894.8367003367,
      "last_run_timestamp": "2026-10-01T19:12:44.398Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.925926,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.14454752525252526,
      "avg_duration_per_task_ms": 66158.0808080808,
      "avg_output_tokens_per_task": 9383.212121212122,
      "last_run_timestamp": "2026-10-03T19:50:41.468Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.9259256666666668,
      "accuracy_std_dev": 0.0036366213562713575,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.1413088872053872,
      "avg_duration_per_task_ms": 82468.60606060606,
      "avg_output_tokens_per_task": 15584.752525252525,
      "last_run_timestamp": "2026-10-01T19:18:40.706Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9259255000000001,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.025620611111111115,
      "avg_duration_per_task_ms": 19140.497474747473,
      "avg_output_tokens_per_task": 2323.35101010101,
      "last_run_timestamp": "2026-10-03T09:34:06.834Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9255440309278352,
      "accuracy_std_dev": 0.0047577793839843644,
      "run_count": 3,
      "total_questions": 582,
      "avg_cost_per_task_usd": 0.3407643092783505,
      "avg_duration_per_task_ms": 132309.46391752578,
      "avg_output_tokens_per_task": 10197.671821305841,
      "last_run_timestamp": "2026-10-01T19:13:06.559Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.9248735,
      "accuracy_std_dev": 0.00862277742377712,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.10364464204545455,
      "avg_duration_per_task_ms": 55123.33648989899,
      "avg_output_tokens_per_task": 3517.8642676767677,
      "last_run_timestamp": "2026-10-01T19:42:48.407Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "xAI",
      "accuracy": 0.9246375695652174,
      "accuracy_std_dev": 0.006863146440316182,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.12292555942028985,
      "avg_duration_per_task_ms": 340480.1579710145,
      "avg_output_tokens_per_task": 19662.684057971015,
      "last_run_timestamp": "2026-10-01T16:09:31.184Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.924599700084246,
      "accuracy_std_dev": 0.006819818440072705,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.02751449789385004,
      "avg_duration_per_task_ms": 24957.139848357205,
      "avg_output_tokens_per_task": 2420.7127211457455,
      "last_run_timestamp": "2026-10-01T08:24:52.927Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "Google",
      "accuracy": 0.9242421666666667,
      "accuracy_std_dev": 0.0048599613653014136,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.14878865656565657,
      "avg_duration_per_task_ms": 85040.50420875421,
      "avg_output_tokens_per_task": 15641.649831649831,
      "last_run_timestamp": "2026-10-01T19:13:26.156Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "unbiased/pareto",
      "model_slug": "unbiased/pareto",
      "model_name": "Pareto",
      "provider_name": "auto-routing",
      "accuracy": 0.9242420000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.03377419696969697,
      "avg_duration_per_task_ms": 114763.54545454546,
      "avg_output_tokens_per_task": 4358.363636363636,
      "last_run_timestamp": "2026-09-29T03:11:33.733Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.9234005000000001,
      "accuracy_std_dev": 0.008625346413333228,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.05533751388888889,
      "avg_duration_per_task_ms": 61705.84722222222,
      "avg_output_tokens_per_task": 3564.286616161616,
      "last_run_timestamp": "2026-10-02T23:20:36.509Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9227054347826086,
      "accuracy_std_dev": 0.0037342422507786015,
      "run_count": 4,
      "total_questions": 759,
      "avg_cost_per_task_usd": 0.09927257048748353,
      "avg_duration_per_task_ms": 52555.73781291173,
      "avg_output_tokens_per_task": 3495.6416337285905,
      "last_run_timestamp": "2026-10-01T19:12:57.119Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Modal",
      "accuracy": 0.9219427970033833,
      "accuracy_std_dev": 0.006990580536078893,
      "run_count": 11,
      "total_questions": 2069,
      "avg_cost_per_task_usd": 0.15209735766070562,
      "avg_duration_per_task_ms": 132397.76655389077,
      "avg_output_tokens_per_task": 9993.401159980667,
      "last_run_timestamp": "2026-10-03T10:21:23.522Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Decart",
      "accuracy": 0.9212965,
      "accuracy_std_dev": 0.0012624999999999997,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13693505808080808,
      "avg_duration_per_task_ms": 189958.2702020202,
      "avg_output_tokens_per_task": 8925.60101010101,
      "last_run_timestamp": "2026-10-03T09:18:30.358Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "auto-routing",
      "accuracy": 0.9208750000000001,
      "accuracy_std_dev": 0.004761657064510223,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.06043574579124578,
      "avg_duration_per_task_ms": 62971.38047138047,
      "avg_output_tokens_per_task": 3586.8703703703704,
      "last_run_timestamp": "2026-10-01T19:12:28.573Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.920875,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.025261131313131315,
      "avg_duration_per_task_ms": 21541.484848484848,
      "avg_output_tokens_per_task": 2294.9949494949497,
      "last_run_timestamp": "2026-10-03T09:35:43.283Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.9208749999999999,
      "accuracy_std_dev": 0.002749143987983661,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.025590255892255893,
      "avg_duration_per_task_ms": 33502.22222222222,
      "avg_output_tokens_per_task": 2400.506734006734,
      "last_run_timestamp": "2026-10-02T23:17:42.236Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Sail Research",
      "accuracy": 0.9205012222222222,
      "accuracy_std_dev": 0.00850227984821038,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.1576766610549944,
      "avg_duration_per_task_ms": 236410.12177328844,
      "avg_output_tokens_per_task": 12595.254208754208,
      "last_run_timestamp": "2026-09-26T08:55:21.775Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Alibaba",
      "accuracy": 0.920314,
      "accuracy_std_dev": 0.007316769300722847,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1692038207070707,
      "avg_duration_per_task_ms": 320867.6077441077,
      "avg_output_tokens_per_task": 9628.20622895623,
      "last_run_timestamp": "2026-10-03T09:31:32.854Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "auto-routing",
      "accuracy": 0.9191919999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02528019191919192,
      "avg_duration_per_task_ms": 32477.28787878788,
      "avg_output_tokens_per_task": 2369.5,
      "last_run_timestamp": "2026-10-01T08:24:24.034Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9191915,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.026391550505050505,
      "avg_duration_per_task_ms": 26392.646464646463,
      "avg_output_tokens_per_task": 2240.712121212121,
      "last_run_timestamp": "2026-10-01T08:37:17.903Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "NextBit",
      "accuracy": 0.918981375,
      "accuracy_std_dev": 0.00679620392089402,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.12805398863636364,
      "avg_duration_per_task_ms": 541652.8712121212,
      "avg_output_tokens_per_task": 37116.541666666664,
      "last_run_timestamp": "2026-10-03T09:05:48.324Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "auto-routing",
      "accuracy": 0.9185433141891891,
      "accuracy_std_dev": 0.011935651348422994,
      "run_count": 9,
      "total_questions": 1776,
      "avg_cost_per_task_usd": 0.13230360641891895,
      "avg_duration_per_task_ms": 318469.4380630631,
      "avg_output_tokens_per_task": 10386.162725225226,
      "last_run_timestamp": "2026-10-03T09:44:31.992Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Wafer",
      "accuracy": 0.9181816,
      "accuracy_std_dev": 0.007721961838807562,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0816559797979798,
      "avg_duration_per_task_ms": 327979.47777777776,
      "avg_output_tokens_per_task": 24710.48888888889,
      "last_run_timestamp": "2026-10-03T11:44:39.356Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "BaseTen",
      "accuracy": 0.9173829326647566,
      "accuracy_std_dev": 0.013022336741090185,
      "run_count": 11,
      "total_questions": 2094,
      "avg_cost_per_task_usd": 0.12426453438395414,
      "avg_duration_per_task_ms": 122014.65329512894,
      "avg_output_tokens_per_task": 8324.876313276027,
      "last_run_timestamp": "2026-10-03T09:17:37.556Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "AkashML",
      "accuracy": 0.9171245769230769,
      "accuracy_std_dev": 0.013071755977424514,
      "run_count": 6,
      "total_questions": 728,
      "avg_cost_per_task_usd": 0.1094587857142857,
      "avg_duration_per_task_ms": 198227.6401098901,
      "avg_output_tokens_per_task": 7265.833791208791,
      "last_run_timestamp": "2026-10-09T16:12:14.447Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Baidu",
      "accuracy": 0.9161053333333333,
      "accuracy_std_dev": 0.004900048423797018,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.10286850841750843,
      "avg_duration_per_task_ms": 630668.6658249159,
      "avg_output_tokens_per_task": 39966.40488215488,
      "last_run_timestamp": "2026-10-03T09:15:20.023Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Venice",
      "accuracy": 0.9158830997830802,
      "accuracy_std_dev": 0.0087081660666812,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.2615176702819957,
      "avg_duration_per_task_ms": 379032.86912509036,
      "avg_output_tokens_per_task": 37727.92263195951,
      "last_run_timestamp": "2026-10-03T09:17:56.331Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Moonshot AI",
      "accuracy": 0.9153725336438923,
      "accuracy_std_dev": 0.007489152329202574,
      "run_count": 10,
      "total_questions": 1932,
      "avg_cost_per_task_usd": 0.14073902795031057,
      "avg_duration_per_task_ms": 293484.7536231884,
      "avg_output_tokens_per_task": 9227.590062111802,
      "last_run_timestamp": "2026-10-03T10:17:13.505Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "InferenceNet",
      "accuracy": 0.9148461928609483,
      "accuracy_std_dev": 0.040300990523310966,
      "run_count": 10,
      "total_questions": 1877,
      "avg_cost_per_task_usd": 0.11435862759722962,
      "avg_duration_per_task_ms": 215504.55833777305,
      "avg_output_tokens_per_task": 9109.937133724028,
      "last_run_timestamp": "2026-10-03T10:08:37.048Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Azure",
      "accuracy": 0.9145619999999999,
      "accuracy_std_dev": 0.003005841230005351,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.025954439393939394,
      "avg_duration_per_task_ms": 19982.071969696968,
      "avg_output_tokens_per_task": 2234.438131313131,
      "last_run_timestamp": "2026-10-03T10:12:04.782Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Minimax",
      "accuracy": 0.9141412727272726,
      "accuracy_std_dev": 0.01313873065053996,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.0326499334251607,
      "avg_duration_per_task_ms": 226257.22359963268,
      "avg_output_tokens_per_task": 27026.729568411385,
      "last_run_timestamp": "2026-10-03T16:51:25.531Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "AtlasCloud",
      "accuracy": 0.9132995,
      "accuracy_std_dev": 0.010546877796295947,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.15848487878787879,
      "avg_duration_per_task_ms": 520006.99494949495,
      "avg_output_tokens_per_task": 39704.125,
      "last_run_timestamp": "2026-10-03T09:15:53.269Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "GMICloud",
      "accuracy": 0.9117844,
      "accuracy_std_dev": 0.006079420781620585,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.03790721616161616,
      "avg_duration_per_task_ms": 129456.97070707071,
      "avg_output_tokens_per_task": 28484.87575757576,
      "last_run_timestamp": "2026-10-03T09:28:48.247Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Morph",
      "accuracy": 0.9115787584210527,
      "accuracy_std_dev": 0.03150249392869556,
      "run_count": 20,
      "total_questions": 3800,
      "avg_cost_per_task_usd": 0.14629828526315786,
      "avg_duration_per_task_ms": 443992.2734210526,
      "avg_output_tokens_per_task": 8936.269473684211,
      "last_run_timestamp": "2026-10-03T10:26:01.662Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Chutes",
      "accuracy": 0.9098249433962265,
      "accuracy_std_dev": 0.01556456113130081,
      "run_count": 10,
      "total_questions": 1961,
      "avg_cost_per_task_usd": 0.18924138296787354,
      "avg_duration_per_task_ms": 516166.16930137685,
      "avg_output_tokens_per_task": 12560.821009688934,
      "last_run_timestamp": "2026-10-03T08:12:32.366Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "CoreWeave",
      "accuracy": 0.9095505370786516,
      "accuracy_std_dev": 0.004402180380080482,
      "run_count": 9,
      "total_questions": 1780,
      "avg_cost_per_task_usd": 0.02965778258426966,
      "avg_duration_per_task_ms": 474993.9230337079,
      "avg_output_tokens_per_task": 30670.006179775282,
      "last_run_timestamp": "2026-10-03T17:37:28.270Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.9094274,
      "accuracy_std_dev": 0.002474202384608023,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07117444141414142,
      "avg_duration_per_task_ms": 135218.17474747475,
      "avg_output_tokens_per_task": 27695.128282828282,
      "last_run_timestamp": "2026-10-01T16:11:23.394Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "NextBit",
      "accuracy": 0.9094272,
      "accuracy_std_dev": 0.00547072746168187,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.025930728282828285,
      "avg_duration_per_task_ms": 177886.11212121212,
      "avg_output_tokens_per_task": 28121.050505050505,
      "last_run_timestamp": "2026-10-03T09:34:35.533Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Reka",
      "accuracy": 0.9090909999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.1305062777777778,
      "avg_duration_per_task_ms": 540279.6767676767,
      "avg_output_tokens_per_task": 38448.40404040404,
      "last_run_timestamp": "2026-10-01T09:14:23.001Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "GLM 5.3 Flash",
      "provider_name": "digitalocean",
      "accuracy": 0.9090909999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0054125,
      "avg_duration_per_task_ms": 235839.91414141413,
      "avg_output_tokens_per_task": 10749.717171717171,
      "last_run_timestamp": "2026-10-01T23:50:45.024Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.7-max-20260520",
      "model_slug": "qwen/qwen3.7-max",
      "model_name": "Qwen: Qwen3.7 Max",
      "provider_name": "AtlasCloud",
      "accuracy": 0.9090909999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.2221867676767677,
      "avg_duration_per_task_ms": 298883.5202020202,
      "avg_output_tokens_per_task": 29352.99494949495,
      "last_run_timestamp": "2026-09-27T00:06:19.757Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Relace",
      "accuracy": 0.9075092967032967,
      "accuracy_std_dev": 0.0008110000000000062,
      "run_count": 2,
      "total_questions": 364,
      "avg_cost_per_task_usd": 0.15048144505494504,
      "avg_duration_per_task_ms": 345399.21428571426,
      "avg_output_tokens_per_task": 38131.52197802198,
      "last_run_timestamp": "2026-10-03T10:04:03.257Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "Tencent",
      "accuracy": 0.907407,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.11648837373737374,
      "avg_duration_per_task_ms": 581221.1060606061,
      "avg_output_tokens_per_task": 46418.58585858586,
      "last_run_timestamp": "2026-10-03T09:31:25.870Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Alibaba",
      "accuracy": 0.9072202222222222,
      "accuracy_std_dev": 0.017575364676335258,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.13183432884399549,
      "avg_duration_per_task_ms": 553846.1520763187,
      "avg_output_tokens_per_task": 39640.39955106622,
      "last_run_timestamp": "2026-10-09T21:00:06.373Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/switchyard",
      "model_slug": "nvidia/switchyard",
      "model_name": "NVIDIA: Switchyard",
      "provider_name": "auto-routing",
      "accuracy": 0.9071381724573594,
      "accuracy_std_dev": 0.03371213456235009,
      "run_count": 8,
      "total_questions": 1583,
      "avg_cost_per_task_usd": 0.015826173089071385,
      "avg_duration_per_task_ms": 64047.12571067593,
      "avg_output_tokens_per_task": 3604.423246999368,
      "last_run_timestamp": "2026-09-30T03:11:21.812Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Together",
      "accuracy": 0.9071160662921349,
      "accuracy_std_dev": 0.01732392826250572,
      "run_count": 9,
      "total_questions": 1780,
      "avg_cost_per_task_usd": 0.03509260337078652,
      "avg_duration_per_task_ms": 322629.44325842697,
      "avg_output_tokens_per_task": 29224.32752808989,
      "last_run_timestamp": "2026-10-03T19:03:35.841Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Novita",
      "accuracy": 0.9063971,
      "accuracy_std_dev": 0.022689051418911293,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.03762983484848485,
      "avg_duration_per_task_ms": 268783.10353535356,
      "avg_output_tokens_per_task": 31583.5601010101,
      "last_run_timestamp": "2026-10-03T16:39:40.370Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.9060606,
      "accuracy_std_dev": 0.006060466668500054,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.023695626262626262,
      "avg_duration_per_task_ms": 214413.58686868686,
      "avg_output_tokens_per_task": 29199.484848484848,
      "last_run_timestamp": "2026-10-03T09:22:46.333Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "GMICloud",
      "accuracy": 0.9059472448132779,
      "accuracy_std_dev": 0.025925182844816933,
      "run_count": 11,
      "total_questions": 2169,
      "avg_cost_per_task_usd": 0.03841674366067312,
      "avg_duration_per_task_ms": 255822.2120792992,
      "avg_output_tokens_per_task": 30855.921161825725,
      "last_run_timestamp": "2026-10-03T16:57:18.499Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.90572375,
      "accuracy_std_dev": 0.009744176680843784,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.03296931691919192,
      "avg_duration_per_task_ms": 228601.25883838383,
      "avg_output_tokens_per_task": 27296.915404040403,
      "last_run_timestamp": "2026-10-03T10:17:41.561Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Venice",
      "accuracy": 0.905492966223132,
      "accuracy_std_dev": 0.012089569061385095,
      "run_count": 10,
      "total_questions": 1954,
      "avg_cost_per_task_usd": 0.02905465660184237,
      "avg_duration_per_task_ms": 262994.571136131,
      "avg_output_tokens_per_task": 24078.712896622314,
      "last_run_timestamp": "2026-10-03T16:51:44.428Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.9053872,
      "accuracy_std_dev": 0.006933048201188252,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.035172772727272726,
      "avg_duration_per_task_ms": 190581.2101010101,
      "avg_output_tokens_per_task": 29132.872727272726,
      "last_run_timestamp": "2026-10-03T09:10:43.528Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "InferenceNet",
      "accuracy": 0.905376,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 155,
      "avg_cost_per_task_usd": 0.09219292258064517,
      "avg_duration_per_task_ms": 892432.9419354838,
      "avg_output_tokens_per_task": 36790.02580645162,
      "last_run_timestamp": "2026-10-02T06:44:44.722Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.9053669934810953,
      "accuracy_std_dev": 0.014124276768541987,
      "run_count": 8,
      "total_questions": 1534,
      "avg_cost_per_task_usd": 0.012634106910039114,
      "avg_duration_per_task_ms": 286058.7731421121,
      "avg_output_tokens_per_task": 17810.844198174706,
      "last_run_timestamp": "2026-10-03T10:17:33.683Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "auto-routing",
      "accuracy": 0.9051339848275863,
      "accuracy_std_dev": 0.019662141676895572,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.033351784367816086,
      "avg_duration_per_task_ms": 333114.86988505744,
      "avg_output_tokens_per_task": 29114.81333333333,
      "last_run_timestamp": "2026-10-03T18:02:12.491Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "StreamLake",
      "accuracy": 0.903872,
      "accuracy_std_dev": 0.028683231540396564,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.04186602929292929,
      "avg_duration_per_task_ms": 291657.99141414143,
      "avg_output_tokens_per_task": 32019.437373737375,
      "last_run_timestamp": "2026-10-03T16:15:35.791Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.903508802631579,
      "accuracy_std_dev": 0.0984941779678147,
      "run_count": 3,
      "total_questions": 456,
      "avg_cost_per_task_usd": 0.11285829166666667,
      "avg_duration_per_task_ms": 423954.7916666667,
      "avg_output_tokens_per_task": 16846.184210526317,
      "last_run_timestamp": "2026-10-01T16:04:59.818Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Phala",
      "accuracy": 0.903424887142055,
      "accuracy_std_dev": 0.056166981894488544,
      "run_count": 9,
      "total_questions": 1781,
      "avg_cost_per_task_usd": 0.14229890510948906,
      "avg_duration_per_task_ms": 308136.8270634475,
      "avg_output_tokens_per_task": 10189.49241998877,
      "last_run_timestamp": "2026-10-03T10:42:35.637Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Baidu",
      "accuracy": 0.9031985,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.033975815656565655,
      "avg_duration_per_task_ms": 141490.74494949495,
      "avg_output_tokens_per_task": 28136.507575757576,
      "last_run_timestamp": "2026-10-03T09:38:38.471Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "auto-routing",
      "accuracy": 0.902918,
      "accuracy_std_dev": 0.0007933738084913016,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.14672305387205387,
      "avg_duration_per_task_ms": 133620.78956228957,
      "avg_output_tokens_per_task": 9652.131313131313,
      "last_run_timestamp": "2026-10-01T19:14:23.097Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.9028655747851002,
      "accuracy_std_dev": 0.042913716403447925,
      "run_count": 9,
      "total_questions": 1745,
      "avg_cost_per_task_usd": 0.010968459025787967,
      "avg_duration_per_task_ms": 489384.2842406877,
      "avg_output_tokens_per_task": 20816.36446991404,
      "last_run_timestamp": "2026-10-03T09:16:27.631Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Alibaba",
      "accuracy": 0.9027775,
      "accuracy_std_dev": 0.015400856096334393,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.026259813131313134,
      "avg_duration_per_task_ms": 203049.0542929293,
      "avg_output_tokens_per_task": 26132.907828282827,
      "last_run_timestamp": "2026-10-09T21:02:25.386Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Novita",
      "accuracy": 0.9024267704752275,
      "accuracy_std_dev": 0.008260579897319528,
      "run_count": 5,
      "total_questions": 989,
      "avg_cost_per_task_usd": 0.03290473609706774,
      "avg_duration_per_task_ms": 147318.88675429727,
      "avg_output_tokens_per_task": 29037.182002022244,
      "last_run_timestamp": "2026-10-03T09:38:12.362Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.9023569999999999,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.05417848484848485,
      "avg_duration_per_task_ms": 36580.51515151515,
      "avg_output_tokens_per_task": 2499.1969696969695,
      "last_run_timestamp": "2026-10-03T09:27:36.763Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "Novita",
      "accuracy": 0.9023565,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.032186345959595956,
      "avg_duration_per_task_ms": 777530.196969697,
      "avg_output_tokens_per_task": 36633.931818181816,
      "last_run_timestamp": "2026-10-03T10:59:16.091Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "SiliconFlow",
      "accuracy": 0.9021696666666666,
      "accuracy_std_dev": 0.023279077501768247,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.15281446520763184,
      "avg_duration_per_task_ms": 592828.4741863075,
      "avg_output_tokens_per_task": 38634.81593714927,
      "last_run_timestamp": "2026-10-03T10:13:07.289Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Phala",
      "accuracy": 0.9017359124999998,
      "accuracy_std_dev": 0.023311648922467388,
      "run_count": 8,
      "total_questions": 1440,
      "avg_cost_per_task_usd": 0.15431451527777776,
      "avg_duration_per_task_ms": 567983.5055555556,
      "avg_output_tokens_per_task": 37740.19097222222,
      "last_run_timestamp": "2026-10-03T09:09:53.172Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Together",
      "accuracy": 0.9016834000000001,
      "accuracy_std_dev": 0.04222344006165297,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.1424910212121212,
      "avg_duration_per_task_ms": 211878.58636363636,
      "avg_output_tokens_per_task": 9531.086363636363,
      "last_run_timestamp": "2026-10-03T10:10:50.167Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.9016834,
      "accuracy_std_dev": 0.01216780761846602,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.028648670707070705,
      "avg_duration_per_task_ms": 147445.17373737373,
      "avg_output_tokens_per_task": 28162.69292929293,
      "last_run_timestamp": "2026-10-03T09:28:15.105Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "StreamLake",
      "accuracy": 0.9014598562605278,
      "accuracy_std_dev": 0.03309502797836693,
      "run_count": 9,
      "total_questions": 1781,
      "avg_cost_per_task_usd": 0.10678652554744525,
      "avg_duration_per_task_ms": 569188.9988770353,
      "avg_output_tokens_per_task": 39083.0421111735,
      "last_run_timestamp": "2026-10-03T10:32:36.645Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Azure",
      "accuracy": 0.9007667777777778,
      "accuracy_std_dev": 0.005624838482180366,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.20504940852974188,
      "avg_duration_per_task_ms": 98780.56790123456,
      "avg_output_tokens_per_task": 7889.076038159372,
      "last_run_timestamp": "2026-10-03T09:47:38.062Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Parasail",
      "accuracy": 0.900635386539607,
      "accuracy_std_dev": 0.009720906385723498,
      "run_count": 10,
      "total_questions": 1679,
      "avg_cost_per_task_usd": 0.022405076235854672,
      "avg_duration_per_task_ms": 187927.48243001787,
      "avg_output_tokens_per_task": 18532.87313877308,
      "last_run_timestamp": "2026-10-03T16:51:49.235Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9004328214285713,
      "accuracy_std_dev": 0.005913167322978487,
      "run_count": 28,
      "total_questions": 5544,
      "avg_cost_per_task_usd": 0.21622311616161613,
      "avg_duration_per_task_ms": 94665.56475468975,
      "avg_output_tokens_per_task": 7881.3840187590185,
      "last_run_timestamp": "2026-10-03T10:10:31.743Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.9001119999999999,
      "accuracy_std_dev": 0.0034590473640392213,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.053618132996633,
      "avg_duration_per_task_ms": 27125.141414141413,
      "avg_output_tokens_per_task": 2464.4410774410776,
      "last_run_timestamp": "2026-10-03T09:16:11.366Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Baidu",
      "accuracy": 0.8999998999999999,
      "accuracy_std_dev": 0.005644066804175849,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.012073083333333333,
      "avg_duration_per_task_ms": 179513.8106060606,
      "avg_output_tokens_per_task": 26194.345454545455,
      "last_run_timestamp": "2026-10-03T09:29:36.040Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Azure",
      "accuracy": 0.8999248888888889,
      "accuracy_std_dev": 0.006354339085213833,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.15479229124579125,
      "avg_duration_per_task_ms": 130090.1857463524,
      "avg_output_tokens_per_task": 9547.675645342311,
      "last_run_timestamp": "2026-10-01T19:18:24.536Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "Xiaomi",
      "accuracy": 0.8998314999999999,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0343437398989899,
      "avg_duration_per_task_ms": 823628.1161616162,
      "avg_output_tokens_per_task": 39114.656565656565,
      "last_run_timestamp": "2026-10-03T11:02:12.320Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepSeek",
      "accuracy": 0.8996633999999999,
      "accuracy_std_dev": 0.005792805834826492,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.007343192929292929,
      "avg_duration_per_task_ms": 219337.0404040404,
      "avg_output_tokens_per_task": 25809.762626262625,
      "last_run_timestamp": "2026-08-08T09:30:52.456Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.8995779088607595,
      "accuracy_std_dev": 0.007113645637083682,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.016700413502109705,
      "avg_duration_per_task_ms": 439475.2455696203,
      "avg_output_tokens_per_task": 29074.91983122363,
      "last_run_timestamp": "2026-10-08T23:08:23.653Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "OpenAI",
      "accuracy": 0.899551,
      "accuracy_std_dev": 0.006198077174952426,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.14109027272727273,
      "avg_duration_per_task_ms": 131303,
      "avg_output_tokens_per_task": 9278.422558922559,
      "last_run_timestamp": "2026-10-01T19:14:55.028Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.899551,
      "accuracy_std_dev": 0.006198077174952426,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0412280202020202,
      "avg_duration_per_task_ms": 141835.40572390571,
      "avg_output_tokens_per_task": 27211.210437710437,
      "last_run_timestamp": "2026-10-01T16:12:34.647Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "Novita",
      "accuracy": 0.8994372877637131,
      "accuracy_std_dev": 0.01977995405637288,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.03544181097046414,
      "avg_duration_per_task_ms": 152645.21687763714,
      "avg_output_tokens_per_task": 26597.288607594935,
      "last_run_timestamp": "2026-10-03T09:28:39.969Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8994104999999998,
      "accuracy_std_dev": 0.004817079950550935,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.22572103156565654,
      "avg_duration_per_task_ms": 108993.70707070707,
      "avg_output_tokens_per_task": 8821.869949494949,
      "last_run_timestamp": "2026-10-03T09:42:54.478Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "NextBit",
      "accuracy": 0.8993638888888887,
      "accuracy_std_dev": 0.004029367042778795,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.030411617283950615,
      "avg_duration_per_task_ms": 237553.6509539843,
      "avg_output_tokens_per_task": 26633.1290684624,
      "last_run_timestamp": "2026-10-03T10:19:34.320Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.8993264,
      "accuracy_std_dev": 0.006511664598242117,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.13052546969696968,
      "avg_duration_per_task_ms": 30874.333333333332,
      "avg_output_tokens_per_task": 2403.130303030303,
      "last_run_timestamp": "2026-10-01T08:24:30.121Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Google",
      "accuracy": 0.8992219655172418,
      "accuracy_std_dev": 0.0066569301119999745,
      "run_count": 29,
      "total_questions": 5742,
      "avg_cost_per_task_usd": 0.22289186381051904,
      "avg_duration_per_task_ms": 94799.24817136886,
      "avg_output_tokens_per_task": 8115.898293277603,
      "last_run_timestamp": "2026-10-03T10:16:05.445Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8991767777777778,
      "accuracy_std_dev": 0.005290828285875197,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.058499294612794614,
      "avg_duration_per_task_ms": 32080.961840628508,
      "avg_output_tokens_per_task": 2531.0488215488217,
      "last_run_timestamp": "2026-10-03T10:16:57.881Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.8989897499999999,
      "accuracy_std_dev": 0.005188652155184398,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.03249949747474747,
      "avg_duration_per_task_ms": 137621.66666666666,
      "avg_output_tokens_per_task": 26156.715909090908,
      "last_run_timestamp": "2026-10-02T23:26:10.109Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "GPT-6-Luna",
      "provider_name": "auto-routing",
      "accuracy": 0.898477,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 197,
      "avg_cost_per_task_usd": 0.003442197969543147,
      "avg_duration_per_task_ms": 73312.30456852792,
      "avg_output_tokens_per_task": 6831.9390862944165,
      "last_run_timestamp": "2026-10-08T13:29:57.437Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.898148,
      "accuracy_std_dev": 0.012012972737836369,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.03667170117845118,
      "avg_duration_per_task_ms": 157797.89814814815,
      "avg_output_tokens_per_task": 27519.53787878788,
      "last_run_timestamp": "2026-10-03T09:28:24.400Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Venice",
      "accuracy": 0.8980803802559414,
      "accuracy_std_dev": 0.010966029181066412,
      "run_count": 6,
      "total_questions": 1094,
      "avg_cost_per_task_usd": 0.006137671846435101,
      "avg_duration_per_task_ms": 419028.80987202923,
      "avg_output_tokens_per_task": 12654.702925045704,
      "last_run_timestamp": "2026-10-03T10:14:40.227Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Google",
      "accuracy": 0.8980545555555557,
      "accuracy_std_dev": 0.0033774129188909115,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.057495572951739626,
      "avg_duration_per_task_ms": 25176.90740740741,
      "avg_output_tokens_per_task": 2488.7480359147025,
      "last_run_timestamp": "2026-10-03T09:48:22.257Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Azure",
      "accuracy": 0.8978673333333335,
      "accuracy_std_dev": 0.005382399950042904,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.057627444444444446,
      "avg_duration_per_task_ms": 26162.272727272728,
      "avg_output_tokens_per_task": 2526.015993265993,
      "last_run_timestamp": "2026-10-03T10:06:47.904Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8977272500000001,
      "accuracy_std_dev": 0.008615029218609743,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.19720289898989898,
      "avg_duration_per_task_ms": 95317.51073232324,
      "avg_output_tokens_per_task": 7684.227904040404,
      "last_run_timestamp": "2026-10-03T09:39:18.026Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "auto-routing",
      "accuracy": 0.8975330208728654,
      "accuracy_std_dev": 0.015763870376592163,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.11425380961416824,
      "avg_duration_per_task_ms": 628521.0765338393,
      "avg_output_tokens_per_task": 36072.062618595824,
      "last_run_timestamp": "2026-10-03T10:20:39.400Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "auto-routing",
      "accuracy": 0.8974932222222222,
      "accuracy_std_dev": 0.012382476271622079,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.20160183445566776,
      "avg_duration_per_task_ms": 97321.57126823794,
      "avg_output_tokens_per_task": 7848.616722783389,
      "last_run_timestamp": "2026-10-03T09:45:23.245Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8968568964143425,
      "accuracy_std_dev": 0.005278733293840445,
      "run_count": 4,
      "total_questions": 753,
      "avg_cost_per_task_usd": 0.15754132403718457,
      "avg_duration_per_task_ms": 84099.57768924303,
      "avg_output_tokens_per_task": 9420.674634794157,
      "last_run_timestamp": "2026-10-01T19:12:25.243Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DeepInfra",
      "accuracy": 0.8967969786476867,
      "accuracy_std_dev": 0.018458871221177085,
      "run_count": 10,
      "total_questions": 1967,
      "avg_cost_per_task_usd": 0.11560783629893238,
      "avg_duration_per_task_ms": 391555.7447890188,
      "avg_output_tokens_per_task": 7906.456532791052,
      "last_run_timestamp": "2026-10-08T22:32:00.376Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Azure",
      "accuracy": 0.8967218206896553,
      "accuracy_std_dev": 0.007385261912078655,
      "run_count": 20,
      "total_questions": 3915,
      "avg_cost_per_task_usd": 0.12559540842911876,
      "avg_duration_per_task_ms": 57752.34942528736,
      "avg_output_tokens_per_task": 4741.332567049809,
      "last_run_timestamp": "2026-10-03T10:05:55.795Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8966562765957448,
      "accuracy_std_dev": 0.007814351796534355,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.018003054711246202,
      "avg_duration_per_task_ms": 581928.556231003,
      "avg_output_tokens_per_task": 17806.63627152989,
      "last_run_timestamp": "2026-10-05T20:49:39.857Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-flash-20260826",
      "model_slug": "qwen/qwen3.8-flash",
      "model_name": "Qwen: Qwen3.8 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.8962022556962025,
      "accuracy_std_dev": 0.0022689999999999655,
      "run_count": 2,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.020123096202531644,
      "avg_duration_per_task_ms": 653583.2962025317,
      "avg_output_tokens_per_task": 42659.516455696205,
      "last_run_timestamp": "2026-10-10T00:00:04.637Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "ModelRun",
      "accuracy": 0.8960818181818182,
      "accuracy_std_dev": 0.021738909414052866,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.08613982369146005,
      "avg_duration_per_task_ms": 211123.70798898072,
      "avg_output_tokens_per_task": 28935.424242424244,
      "last_run_timestamp": "2026-10-05T18:10:18.108Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Novita",
      "accuracy": 0.8956340674536255,
      "accuracy_std_dev": 0.02132427030288766,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.13012930916245083,
      "avg_duration_per_task_ms": 507416.96458684653,
      "avg_output_tokens_per_task": 37964.30185497471,
      "last_run_timestamp": "2026-10-03T10:12:16.841Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepSeek",
      "accuracy": 0.8956230000000002,
      "accuracy_std_dev": 0.007273551310513056,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.00771163468013468,
      "avg_duration_per_task_ms": 254636.3367003367,
      "avg_output_tokens_per_task": 27268.55387205387,
      "last_run_timestamp": "2026-08-08T10:06:09.471Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Anthropic",
      "accuracy": 0.8956227499999999,
      "accuracy_std_dev": 0.005791223203046117,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.20478679124579122,
      "avg_duration_per_task_ms": 99940.12121212122,
      "avg_output_tokens_per_task": 7979.755892255892,
      "last_run_timestamp": "2026-10-03T09:40:16.985Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Wafer",
      "accuracy": 0.8952778724516466,
      "accuracy_std_dev": 0.027063673618015708,
      "run_count": 11,
      "total_questions": 1913,
      "avg_cost_per_task_usd": 0.2081439644537376,
      "avg_duration_per_task_ms": 408679.8609513853,
      "avg_output_tokens_per_task": 45592.24150548876,
      "last_run_timestamp": "2026-10-08T06:00:12.150Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Anthropic",
      "accuracy": 0.8951176,
      "accuracy_std_dev": 0.00756834082477789,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.12048651767676768,
      "avg_duration_per_task_ms": 57579.538383838386,
      "avg_output_tokens_per_task": 4603.499494949495,
      "last_run_timestamp": "2026-10-03T09:25:15.659Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8950130708661417,
      "accuracy_std_dev": 0.00891334037763057,
      "run_count": 4,
      "total_questions": 762,
      "avg_cost_per_task_usd": 0.048151682414698166,
      "avg_duration_per_task_ms": 43476.24934383202,
      "avg_output_tokens_per_task": 3503.036745406824,
      "last_run_timestamp": "2026-10-01T19:11:43.444Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "DeepInfra",
      "accuracy": 0.8949103109177213,
      "accuracy_std_dev": 0.025790262867514752,
      "run_count": 7,
      "total_questions": 1264,
      "avg_cost_per_task_usd": 0.02656089082278481,
      "avg_duration_per_task_ms": 720996.8117088608,
      "avg_output_tokens_per_task": 24102.247626582277,
      "last_run_timestamp": "2026-09-12T16:17:57.063Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Novita",
      "accuracy": 0.8948596495327101,
      "accuracy_std_dev": 0.0171199232602798,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.022189622274143304,
      "avg_duration_per_task_ms": 231984.90109034267,
      "avg_output_tokens_per_task": 26737.70015576324,
      "last_run_timestamp": "2026-10-03T09:42:10.726Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8945527678958785,
      "accuracy_std_dev": 0.010412875331697246,
      "run_count": 14,
      "total_questions": 2766,
      "avg_cost_per_task_usd": 0.012522617136659437,
      "avg_duration_per_task_ms": 291164.3582791034,
      "avg_output_tokens_per_task": 27224.147143890095,
      "last_run_timestamp": "2026-10-03T09:48:09.405Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "auto-routing",
      "accuracy": 0.8942841547799696,
      "accuracy_std_dev": 0.008875212006481873,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.22388517248356093,
      "avg_duration_per_task_ms": 107507.08952959029,
      "avg_output_tokens_per_task": 8731.16894284269,
      "last_run_timestamp": "2026-10-03T09:18:11.817Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Azure",
      "accuracy": 0.8941264444444444,
      "accuracy_std_dev": 0.009636502052221474,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.05178651795735129,
      "avg_duration_per_task_ms": 45445.343434343435,
      "avg_output_tokens_per_task": 3671.5897867564536,
      "last_run_timestamp": "2026-10-01T19:13:06.514Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Google",
      "accuracy": 0.8938831666666668,
      "accuracy_std_dev": 0.007122347684499023,
      "run_count": 30,
      "total_questions": 5940,
      "avg_cost_per_task_usd": 0.20402719309764314,
      "avg_duration_per_task_ms": 70642.39461279461,
      "avg_output_tokens_per_task": 7430.633333333333,
      "last_run_timestamp": "2026-10-03T10:09:58.833Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "OpenAI",
      "accuracy": 0.89267675,
      "accuracy_std_dev": 0.006979218773437292,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.037834488636363635,
      "avg_duration_per_task_ms": 50710.07575757576,
      "avg_output_tokens_per_task": 3480.1502525252527,
      "last_run_timestamp": "2026-10-02T23:21:06.556Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepInfra",
      "accuracy": 0.89267063172226,
      "accuracy_std_dev": 0.013806389148989432,
      "run_count": 15,
      "total_questions": 2938,
      "avg_cost_per_task_usd": 0.00379094996596324,
      "avg_duration_per_task_ms": 310296.7300884956,
      "avg_output_tokens_per_task": 20769.534717494895,
      "last_run_timestamp": "2026-10-08T23:06:17.060Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "BaseTen",
      "accuracy": 0.892567903763697,
      "accuracy_std_dev": 0.02022831814050264,
      "run_count": 11,
      "total_questions": 2099,
      "avg_cost_per_task_usd": 0.15797696188661267,
      "avg_duration_per_task_ms": 338559.5883754169,
      "avg_output_tokens_per_task": 39638.616484040016,
      "last_run_timestamp": "2026-09-19T08:51:08.108Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.892548239520958,
      "accuracy_std_dev": 0.006915975724525221,
      "run_count": 3,
      "total_questions": 501,
      "avg_cost_per_task_usd": 0.013722602794411178,
      "avg_duration_per_task_ms": 355406.4910179641,
      "avg_output_tokens_per_task": 23503.43512974052,
      "last_run_timestamp": "2026-10-03T10:04:24.528Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "DeepInfra",
      "accuracy": 0.892405,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 158,
      "avg_cost_per_task_usd": 0.12755274683544304,
      "avg_duration_per_task_ms": 1596411.9683544303,
      "avg_output_tokens_per_task": 50780.917721518985,
      "last_run_timestamp": "2026-10-08T23:19:46.552Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8922556666666667,
      "accuracy_std_dev": 0.0023810642345154617,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.05459258585858586,
      "avg_duration_per_task_ms": 34041.639730639734,
      "avg_output_tokens_per_task": 2518.429292929293,
      "last_run_timestamp": "2026-10-03T09:50:45.552Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "DeepInfra",
      "accuracy": 0.8909751541501977,
      "accuracy_std_dev": 0.003207768937369934,
      "run_count": 3,
      "total_questions": 506,
      "avg_cost_per_task_usd": 0.0325840256916996,
      "avg_duration_per_task_ms": 936484.3814229249,
      "avg_output_tokens_per_task": 37243.156126482216,
      "last_run_timestamp": "2026-10-08T22:57:13.219Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "Google",
      "accuracy": 0.8905723333333333,
      "accuracy_std_dev": 0.021339306148253502,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.13216584343434343,
      "avg_duration_per_task_ms": 251067.54713804714,
      "avg_output_tokens_per_task": 43934.003367003366,
      "last_run_timestamp": "2026-10-01T19:18:35.588Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "SiliconFlow",
      "accuracy": 0.890572,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.12533664393939395,
      "avg_duration_per_task_ms": 808381.0631313132,
      "avg_output_tokens_per_task": 49868.60353535353,
      "last_run_timestamp": "2026-10-03T09:33:15.909Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DekaLLM",
      "accuracy": 0.8904339810366624,
      "accuracy_std_dev": 0.010766495576556007,
      "run_count": 4,
      "total_questions": 791,
      "avg_cost_per_task_usd": 0.020936142857142858,
      "avg_duration_per_task_ms": 384871.00505689,
      "avg_output_tokens_per_task": 23969.699115044248,
      "last_run_timestamp": "2026-10-09T02:24:39.455Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "Anthropic",
      "accuracy": 0.8900113333333334,
      "accuracy_std_dev": 0.005722799567422281,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.19800758417508418,
      "avg_duration_per_task_ms": 96141.14141414141,
      "avg_output_tokens_per_task": 7695.558922558923,
      "last_run_timestamp": "2026-10-01T19:11:18.427Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.8900113333333334,
      "accuracy_std_dev": 0.0015872190215034077,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.008230473063973063,
      "avg_duration_per_task_ms": 80128.80303030302,
      "avg_output_tokens_per_task": 12935.998316498317,
      "last_run_timestamp": "2026-10-02T23:23:43.005Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "auto-routing",
      "accuracy": 0.8900110000000001,
      "accuracy_std_dev": 0.0055552329084086844,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.2164912121212121,
      "avg_duration_per_task_ms": 104660.01851851853,
      "avg_output_tokens_per_task": 8434.90404040404,
      "last_run_timestamp": "2026-10-01T19:10:42.213Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "GMICloud",
      "accuracy": 0.8898507142857144,
      "accuracy_std_dev": 0.013463734020527524,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.013538019841269842,
      "avg_duration_per_task_ms": 210147.86976911977,
      "avg_output_tokens_per_task": 24623.09126984127,
      "last_run_timestamp": "2026-10-03T09:29:34.537Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Google",
      "accuracy": 0.8898480650076179,
      "accuracy_std_dev": 0.00936748461292215,
      "run_count": 30,
      "total_questions": 5907,
      "avg_cost_per_task_usd": 0.13325874741831725,
      "avg_duration_per_task_ms": 59111.40138818351,
      "avg_output_tokens_per_task": 4767.174030810902,
      "last_run_timestamp": "2026-10-03T09:22:12.774Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Parasail",
      "accuracy": 0.8895201249999999,
      "accuracy_std_dev": 0.018150193638343783,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.1383345025252525,
      "avg_duration_per_task_ms": 580496.523989899,
      "avg_output_tokens_per_task": 34909.79734848485,
      "last_run_timestamp": "2026-10-03T10:33:13.241Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.88945,
      "accuracy_std_dev": 0.00975207530050226,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.024394900673400673,
      "avg_duration_per_task_ms": 319196.3552188552,
      "avg_output_tokens_per_task": 27886,
      "last_run_timestamp": "2026-10-03T10:36:49.762Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Anthropic",
      "accuracy": 0.8894498333333334,
      "accuracy_std_dev": 0.02853141346198763,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.136280962962963,
      "avg_duration_per_task_ms": 40331.19781144781,
      "avg_output_tokens_per_task": 2516.2390572390573,
      "last_run_timestamp": "2026-10-01T08:24:17.840Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.8894498333333334,
      "accuracy_std_dev": 0.009896244159893304,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.008765856902356902,
      "avg_duration_per_task_ms": 59664.589225589225,
      "avg_output_tokens_per_task": 12901.691919191919,
      "last_run_timestamp": "2026-10-01T08:29:53.989Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Anthropic",
      "accuracy": 0.8894080654205608,
      "accuracy_std_dev": 0.00941594536414648,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.21110783605919006,
      "avg_duration_per_task_ms": 101949.81853582554,
      "avg_output_tokens_per_task": 8216.18808411215,
      "last_run_timestamp": "2026-10-03T09:19:25.458Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8892253999999999,
      "accuracy_std_dev": 0.008511315206241623,
      "run_count": 20,
      "total_questions": 3960,
      "avg_cost_per_task_usd": 0.24073908914141412,
      "avg_duration_per_task_ms": 84900.21515151515,
      "avg_output_tokens_per_task": 8956.617424242424,
      "last_run_timestamp": "2026-10-03T09:30:01.468Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cohere",
      "accuracy": 0.8888886666666669,
      "accuracy_std_dev": 0.007653072992087717,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.008966420875420876,
      "avg_duration_per_task_ms": 148767.4797979798,
      "avg_output_tokens_per_task": 31547.74074074074,
      "last_run_timestamp": "2026-10-03T10:11:42.574Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Morph",
      "accuracy": 0.8883276666666667,
      "accuracy_std_dev": 0.0028614651647170097,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.008124902356902356,
      "avg_duration_per_task_ms": 491163.50673400675,
      "avg_output_tokens_per_task": 22629.957912457914,
      "last_run_timestamp": "2026-10-01T11:20:35.509Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Reka",
      "accuracy": 0.8881577165991904,
      "accuracy_std_dev": 0.008283430245979028,
      "run_count": 10,
      "total_questions": 1976,
      "avg_cost_per_task_usd": 0.015942041497975706,
      "avg_duration_per_task_ms": 164141.9762145749,
      "avg_output_tokens_per_task": 26128.10374493927,
      "last_run_timestamp": "2026-10-03T13:59:16.761Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "Novita",
      "accuracy": 0.8880469999999999,
      "accuracy_std_dev": 0.004209000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.11352663888888888,
      "avg_duration_per_task_ms": 879783.9520202021,
      "avg_output_tokens_per_task": 45200.48484848485,
      "last_run_timestamp": "2026-10-03T10:42:46.764Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Azure",
      "accuracy": 0.8880445319148936,
      "accuracy_std_dev": 0.009804096960454846,
      "run_count": 10,
      "total_questions": 1974,
      "avg_cost_per_task_usd": 0.20574746403242147,
      "avg_duration_per_task_ms": 100391.19402228977,
      "avg_output_tokens_per_task": 8003.147416413374,
      "last_run_timestamp": "2026-10-03T09:42:47.873Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Cloudflare",
      "accuracy": 0.887854866328257,
      "accuracy_std_dev": 0.036201773425286826,
      "run_count": 9,
      "total_questions": 1773,
      "avg_cost_per_task_usd": 0.13917155047941343,
      "avg_duration_per_task_ms": 799755.2588832488,
      "avg_output_tokens_per_task": 34841.37958262832,
      "last_run_timestamp": "2026-10-03T10:47:57.305Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "auto-routing",
      "accuracy": 0.8877663333333334,
      "accuracy_std_dev": 0.005555502277522277,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03440870875420875,
      "avg_duration_per_task_ms": 50195.98148148148,
      "avg_output_tokens_per_task": 3309.2491582491584,
      "last_run_timestamp": "2026-10-01T19:13:07.806Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "NextBit",
      "accuracy": 0.887537924012158,
      "accuracy_std_dev": 0.020250680116973837,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.12033997061803446,
      "avg_duration_per_task_ms": 580755.4113475177,
      "avg_output_tokens_per_task": 34357.5663627153,
      "last_run_timestamp": "2026-10-03T16:44:56.199Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Azure",
      "accuracy": 0.8872051666666666,
      "accuracy_std_dev": 0.024761493125527695,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1434989882154882,
      "avg_duration_per_task_ms": 42121.643939393936,
      "avg_output_tokens_per_task": 2660.465488215488,
      "last_run_timestamp": "2026-10-01T08:24:34.535Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "SiliconFlow",
      "accuracy": 0.887192300254453,
      "accuracy_std_dev": 0.021070431267668825,
      "run_count": 6,
      "total_questions": 1179,
      "avg_cost_per_task_usd": 0.19019277692960135,
      "avg_duration_per_task_ms": 1311560.4325699746,
      "avg_output_tokens_per_task": 55844.176420695505,
      "last_run_timestamp": "2026-10-03T17:43:47.691Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.8867842499999999,
      "accuracy_std_dev": 0.003834422340783562,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.027946079545454545,
      "avg_duration_per_task_ms": 154476.702020202,
      "avg_output_tokens_per_task": 23124.396464646463,
      "last_run_timestamp": "2026-10-03T10:53:39.908Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Inceptron",
      "accuracy": 0.8867689502093997,
      "accuracy_std_dev": 0.008568113097385388,
      "run_count": 11,
      "total_questions": 2149,
      "avg_cost_per_task_usd": 0.09704398464402049,
      "avg_duration_per_task_ms": 550284.3764541647,
      "avg_output_tokens_per_task": 29436.060027919964,
      "last_run_timestamp": "2026-10-03T17:04:25.569Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Fireworks",
      "accuracy": 0.8867051248554914,
      "accuracy_std_dev": 0.021208571037200977,
      "run_count": 5,
      "total_questions": 865,
      "avg_cost_per_task_usd": 0.16137210289017342,
      "avg_duration_per_task_ms": 430828.78150289017,
      "avg_output_tokens_per_task": 40500.787283237,
      "last_run_timestamp": "2026-09-19T08:53:57.162Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.8863635000000001,
      "accuracy_std_dev": 0.004208500000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.031905555555555555,
      "avg_duration_per_task_ms": 767151.1060606061,
      "avg_output_tokens_per_task": 36810.59595959596,
      "last_run_timestamp": "2026-10-03T10:37:19.082Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.8855853621621622,
      "accuracy_std_dev": 0.025702889954244423,
      "run_count": 3,
      "total_questions": 555,
      "avg_cost_per_task_usd": 0.012490181981981983,
      "avg_duration_per_task_ms": 346127.65405405406,
      "avg_output_tokens_per_task": 22279.807207207206,
      "last_run_timestamp": "2026-10-03T10:21:30.026Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "Google AI Studio",
      "accuracy": 0.8855220000000001,
      "accuracy_std_dev": 0.012598160421267844,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.1415003787878788,
      "avg_duration_per_task_ms": 184472.2676767677,
      "avg_output_tokens_per_task": 47043.42424242424,
      "last_run_timestamp": "2026-10-01T19:12:30.168Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sakana/fugu-max",
      "model_slug": "sakana/fugu-max",
      "model_name": "Fugu Max",
      "provider_name": "auto-routing",
      "accuracy": 0.885522,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.07555469191919191,
      "avg_duration_per_task_ms": 225091.94949494948,
      "avg_output_tokens_per_task": 12279.550505050505,
      "last_run_timestamp": "2026-09-15T01:07:47.957Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-flash-20260826",
      "model_slug": "qwen/qwen3.8-flash",
      "model_name": "Qwen: Qwen3.8 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.885522,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02146120202020202,
      "avg_duration_per_task_ms": 674351.9393939395,
      "avg_output_tokens_per_task": 45464.17676767677,
      "last_run_timestamp": "2026-10-01T08:31:57.634Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8853236853832442,
      "accuracy_std_dev": 0.023583528877870113,
      "run_count": 6,
      "total_questions": 1122,
      "avg_cost_per_task_usd": 0.1068478440285205,
      "avg_duration_per_task_ms": 683979.679144385,
      "avg_output_tokens_per_task": 29631.141711229946,
      "last_run_timestamp": "2026-10-05T21:52:29.584Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8853190481927712,
      "accuracy_std_dev": 0.005795847495189989,
      "run_count": 4,
      "total_questions": 747,
      "avg_cost_per_task_usd": 0.0097919718875502,
      "avg_duration_per_task_ms": 50542.31994645247,
      "avg_output_tokens_per_task": 7272.5394912985275,
      "last_run_timestamp": "2026-10-01T19:09:12.892Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Relace",
      "accuracy": 0.8846155538461539,
      "accuracy_std_dev": 0.0024989999999999735,
      "run_count": 2,
      "total_questions": 390,
      "avg_cost_per_task_usd": 0.10645926410256411,
      "avg_duration_per_task_ms": 514880.74102564104,
      "avg_output_tokens_per_task": 34685.94102564103,
      "last_run_timestamp": "2026-10-03T17:17:05.949Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Parasail",
      "accuracy": 0.8845953462138955,
      "accuracy_std_dev": 0.012778996436692273,
      "run_count": 13,
      "total_questions": 2562,
      "avg_cost_per_task_usd": 0.007367392271662762,
      "avg_duration_per_task_ms": 467182.3048399688,
      "avg_output_tokens_per_task": 26353.96487119438,
      "last_run_timestamp": "2026-10-03T09:22:28.114Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Sail Research",
      "accuracy": 0.8845707161572053,
      "accuracy_std_dev": 0.02143221713073307,
      "run_count": 6,
      "total_questions": 1145,
      "avg_cost_per_task_usd": 0.16085208820960697,
      "avg_duration_per_task_ms": 756655.7283842795,
      "avg_output_tokens_per_task": 40813.194759825325,
      "last_run_timestamp": "2026-09-16T00:37:42.883Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Novita",
      "accuracy": 0.8843432,
      "accuracy_std_dev": 0.016512865255914843,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.11556713585858584,
      "avg_duration_per_task_ms": 607287.5232323232,
      "avg_output_tokens_per_task": 30218.88484848485,
      "last_run_timestamp": "2026-10-03T16:57:22.957Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.884269595505618,
      "accuracy_std_dev": 0.010583520080008851,
      "run_count": 9,
      "total_questions": 1780,
      "avg_cost_per_task_usd": 0.018860816853932583,
      "avg_duration_per_task_ms": 74171.62584269662,
      "avg_output_tokens_per_task": 7862.741573033708,
      "last_run_timestamp": "2026-10-01T19:12:42.994Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Together",
      "accuracy": 0.884259,
      "accuracy_std_dev": 0.003834285135458768,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.028357917929292934,
      "avg_duration_per_task_ms": 90698.91287878787,
      "avg_output_tokens_per_task": 23491.551767676767,
      "last_run_timestamp": "2026-10-03T10:07:51.005Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Phala",
      "accuracy": 0.8841843306451613,
      "accuracy_std_dev": 0.016935757620711884,
      "run_count": 4,
      "total_questions": 744,
      "avg_cost_per_task_usd": 0.04011120698924731,
      "avg_duration_per_task_ms": 259031.17204301077,
      "avg_output_tokens_per_task": 34712.397849462366,
      "last_run_timestamp": "2026-10-03T11:48:18.674Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "CoreWeave",
      "accuracy": 0.8840485,
      "accuracy_std_dev": 0.014246189262395764,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.18107499936868682,
      "avg_duration_per_task_ms": 298705.5252525252,
      "avg_output_tokens_per_task": 45444.8851010101,
      "last_run_timestamp": "2026-10-03T10:13:34.085Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Moonshot AI",
      "accuracy": 0.8839788953995158,
      "accuracy_std_dev": 0.013479642880336476,
      "run_count": 21,
      "total_questions": 4130,
      "avg_cost_per_task_usd": 0.19186280629539954,
      "avg_duration_per_task_ms": 375394.6331719128,
      "avg_output_tokens_per_task": 31206.65593220339,
      "last_run_timestamp": "2026-10-03T16:59:23.370Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "auto-routing",
      "accuracy": 0.8838383333333335,
      "accuracy_std_dev": 0.001374572014200141,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.13180945791245793,
      "avg_duration_per_task_ms": 248970.3754208754,
      "avg_output_tokens_per_task": 43817.097643097644,
      "last_run_timestamp": "2026-10-01T19:15:31.426Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.883838,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008301368686868687,
      "avg_duration_per_task_ms": 72985.27777777778,
      "avg_output_tokens_per_task": 13098.969696969696,
      "last_run_timestamp": "2026-10-01T08:24:38.949Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Phala",
      "accuracy": 0.8829965000000001,
      "accuracy_std_dev": 0.009135988652028878,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.02256884242424242,
      "avg_duration_per_task_ms": 361325.05656565656,
      "avg_output_tokens_per_task": 25803.312121212122,
      "last_run_timestamp": "2026-10-03T11:18:47.195Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.882652041571654,
      "accuracy_std_dev": 0.010061769676464596,
      "run_count": 30,
      "total_questions": 5701,
      "avg_cost_per_task_usd": 0.13625698438870373,
      "avg_duration_per_task_ms": 62308.83844939484,
      "avg_output_tokens_per_task": 4947.5786704087,
      "last_run_timestamp": "2026-10-03T09:33:58.131Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Novita",
      "accuracy": 0.8823811431670283,
      "accuracy_std_dev": 0.01015848669631376,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.00803116413593637,
      "avg_duration_per_task_ms": 247598.04772234274,
      "avg_output_tokens_per_task": 28350.216919739698,
      "last_run_timestamp": "2026-10-03T16:57:09.267Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Parasail",
      "accuracy": 0.8822876450184503,
      "accuracy_std_dev": 0.011212925575387701,
      "run_count": 7,
      "total_questions": 1355,
      "avg_cost_per_task_usd": 0.07029816752767527,
      "avg_duration_per_task_ms": 312193.415498155,
      "avg_output_tokens_per_task": 19426.869372693727,
      "last_run_timestamp": "2026-10-03T18:59:42.075Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Relace",
      "accuracy": 0.882236963045085,
      "accuracy_std_dev": 0.03276554586072114,
      "run_count": 7,
      "total_questions": 1353,
      "avg_cost_per_task_usd": 0.006029221729490022,
      "avg_duration_per_task_ms": 323803.38359201775,
      "avg_output_tokens_per_task": 25311.461197339246,
      "last_run_timestamp": "2026-10-03T09:49:39.330Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Wafer",
      "accuracy": 0.8816051722365038,
      "accuracy_std_dev": 0.07456878886603124,
      "run_count": 13,
      "total_questions": 2334,
      "avg_cost_per_task_usd": 0.010089964010282775,
      "avg_duration_per_task_ms": 149337.25192802056,
      "avg_output_tokens_per_task": 23612.06083976007,
      "last_run_timestamp": "2026-10-03T10:12:41.974Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.8815936666666665,
      "accuracy_std_dev": 0.013561239749947471,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.005247140572390572,
      "avg_duration_per_task_ms": 250251.0606060606,
      "avg_output_tokens_per_task": 28487.505892255893,
      "last_run_timestamp": "2026-10-03T17:07:12.901Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.8815936666666665,
      "accuracy_std_dev": 0.00975199857579064,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.002981575757575758,
      "avg_duration_per_task_ms": 52993.44276094276,
      "avg_output_tokens_per_task": 5804.632996632997,
      "last_run_timestamp": "2026-10-02T23:20:31.396Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Relace",
      "accuracy": 0.8815301729019211,
      "accuracy_std_dev": 0.0059611119566738745,
      "run_count": 5,
      "total_questions": 989,
      "avg_cost_per_task_usd": 0.014240532861476238,
      "avg_duration_per_task_ms": 332365.4651162791,
      "avg_output_tokens_per_task": 26073.906976744187,
      "last_run_timestamp": "2026-10-03T10:46:54.027Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.8813523333333333,
      "accuracy_std_dev": 0.012418655370047113,
      "run_count": 5,
      "total_questions": 774,
      "avg_cost_per_task_usd": 0.007396838501291989,
      "avg_duration_per_task_ms": 996564.0413436693,
      "avg_output_tokens_per_task": 29356.904392764856,
      "last_run_timestamp": "2026-10-03T09:53:30.020Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8813128999999998,
      "accuracy_std_dev": 0.018445584861695216,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.024727196464646464,
      "avg_duration_per_task_ms": 216333.7505050505,
      "avg_output_tokens_per_task": 25254.168181818182,
      "last_run_timestamp": "2026-10-03T09:13:55.562Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "DeepInfra",
      "accuracy": 0.8811927142857143,
      "accuracy_std_dev": 0.007888829783113527,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.015891973304473304,
      "avg_duration_per_task_ms": 279407.10173160175,
      "avg_output_tokens_per_task": 24276.554834054834,
      "last_run_timestamp": "2026-10-08T22:18:51.642Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Morph",
      "accuracy": 0.8810132745849297,
      "accuracy_std_dev": 0.029062612893329138,
      "run_count": 4,
      "total_questions": 783,
      "avg_cost_per_task_usd": 0.03828848659003832,
      "avg_duration_per_task_ms": 742925.4495530013,
      "avg_output_tokens_per_task": 25587.997445721583,
      "last_run_timestamp": "2026-08-22T10:03:34.994Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Inceptron",
      "accuracy": 0.880903752293578,
      "accuracy_std_dev": 0.010625848964200456,
      "run_count": 10,
      "total_questions": 1962,
      "avg_cost_per_task_usd": 0.005325050458715596,
      "avg_duration_per_task_ms": 400241.4791029562,
      "avg_output_tokens_per_task": 18619.075433231395,
      "last_run_timestamp": "2026-10-03T10:33:52.640Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Novita",
      "accuracy": 0.8808888960000001,
      "accuracy_std_dev": 0.009359000000000006,
      "run_count": 2,
      "total_questions": 375,
      "avg_cost_per_task_usd": 0.10911519200000001,
      "avg_duration_per_task_ms": 1097247.664,
      "avg_output_tokens_per_task": 43639.90933333334,
      "last_run_timestamp": "2026-08-01T17:46:48.312Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Modal",
      "accuracy": 0.8804712499999999,
      "accuracy_std_dev": 0.006838655510222777,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.028478232323232322,
      "avg_duration_per_task_ms": 124855.71338383839,
      "avg_output_tokens_per_task": 23539.155303030304,
      "last_run_timestamp": "2026-10-03T10:18:21.349Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Decart",
      "accuracy": 0.880471,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.006361709595959596,
      "avg_duration_per_task_ms": 516019.3207070707,
      "avg_output_tokens_per_task": 34951.01515151515,
      "last_run_timestamp": "2026-10-03T11:41:19.088Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.880471,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0032475000000000004,
      "avg_duration_per_task_ms": 58949.292929292926,
      "avg_output_tokens_per_task": 5746.025252525253,
      "last_run_timestamp": "2026-09-23T06:09:34.299Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.8801906666666667,
      "accuracy_std_dev": 0.011226983784119793,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0227881835016835,
      "avg_duration_per_task_ms": 234700.68434343435,
      "avg_output_tokens_per_task": 25264.190235690236,
      "last_run_timestamp": "2026-10-03T09:30:18.263Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Decart",
      "accuracy": 0.8798644568527919,
      "accuracy_std_dev": 0.004527248931623837,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.0029938477157360403,
      "avg_duration_per_task_ms": 316818.90016920475,
      "avg_output_tokens_per_task": 18321.88663282572,
      "last_run_timestamp": "2026-08-22T10:20:26.941Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8796295,
      "accuracy_std_dev": 0.015267782951365261,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.028045234848484847,
      "avg_duration_per_task_ms": 222260.06818181818,
      "avg_output_tokens_per_task": 27117.741161616163,
      "last_run_timestamp": "2026-10-03T09:35:14.478Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openrouter/auto",
      "model_slug": "openrouter/auto",
      "model_name": "Auto Router",
      "provider_name": "auto-routing",
      "accuracy": 0.8795957999999999,
      "accuracy_std_dev": 0.026351833738091142,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.08917041111111111,
      "avg_duration_per_task_ms": 196788.44343434344,
      "avg_output_tokens_per_task": 37445.42121212121,
      "last_run_timestamp": "2026-09-29T14:28:17.204Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8794991295774648,
      "accuracy_std_dev": 0.01942029328961192,
      "run_count": 11,
      "total_questions": 2130,
      "avg_cost_per_task_usd": 0.12990812910798122,
      "avg_duration_per_task_ms": 63213.040375586854,
      "avg_output_tokens_per_task": 4970.991549295774,
      "last_run_timestamp": "2026-10-03T10:06:49.792Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Venice",
      "accuracy": 0.8793958967310551,
      "accuracy_std_dev": 0.04684221744738129,
      "run_count": 7,
      "total_questions": 1346,
      "avg_cost_per_task_usd": 0.05736572734026745,
      "avg_duration_per_task_ms": 634171.9687964339,
      "avg_output_tokens_per_task": 19702.684249628528,
      "last_run_timestamp": "2026-10-03T17:12:28.043Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "OpenInference",
      "accuracy": 0.879386,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 76,
      "avg_cost_per_task_usd": 0.03120878947368421,
      "avg_duration_per_task_ms": 274583.63157894736,
      "avg_output_tokens_per_task": 9741.526315789473,
      "last_run_timestamp": "2026-10-09T00:00:06.206Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Reka",
      "accuracy": 0.879189,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 189,
      "avg_cost_per_task_usd": 0.020723571428571428,
      "avg_duration_per_task_ms": 141095.54497354498,
      "avg_output_tokens_per_task": 17734.21164021164,
      "last_run_timestamp": "2026-09-15T00:33:26.568Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Fireworks",
      "accuracy": 0.8790546552363299,
      "accuracy_std_dev": 0.030844451113161732,
      "run_count": 6,
      "total_questions": 1079,
      "avg_cost_per_task_usd": 0.1346450278035218,
      "avg_duration_per_task_ms": 284974.0537534754,
      "avg_output_tokens_per_task": 31126.302131603337,
      "last_run_timestamp": "2026-09-19T09:14:09.704Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "Deepseek V4.1 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.878788,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.010643959595959596,
      "avg_duration_per_task_ms": 127627.76767676767,
      "avg_output_tokens_per_task": 14277.949494949495,
      "last_run_timestamp": "2026-10-08T13:29:57.446Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-pro",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Mimo V2.6 Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.878788,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01249009595959596,
      "avg_duration_per_task_ms": 315799.5808080808,
      "avg_output_tokens_per_task": 14226.873737373737,
      "last_run_timestamp": "2026-10-08T01:55:44.175Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.2-20251211",
      "model_slug": "openai/gpt-5.2",
      "model_name": "OpenAI: GPT-5.2",
      "provider_name": "OpenAI",
      "accuracy": 0.878788,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.13332395959595958,
      "avg_duration_per_task_ms": 138478.24242424243,
      "avg_output_tokens_per_task": 9426.934343434343,
      "last_run_timestamp": "2026-08-01T08:50:28.136Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.7-plus-20260602",
      "model_slug": "qwen/qwen3.7-plus",
      "model_name": "Qwen: Qwen3.7 Plus",
      "provider_name": "auto-routing",
      "accuracy": 0.878788,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.043995565656565656,
      "avg_duration_per_task_ms": 588519.7272727273,
      "avg_output_tokens_per_task": 34166.82828282828,
      "last_run_timestamp": "2026-10-01T08:47:14.818Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "auto-routing",
      "accuracy": 0.8787876666666666,
      "accuracy_std_dev": 0.005991714074916775,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.007907055555555554,
      "avg_duration_per_task_ms": 79561.398989899,
      "avg_output_tokens_per_task": 7834.653198653199,
      "last_run_timestamp": "2026-10-01T19:14:22.294Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "StreamLake",
      "accuracy": 0.8787876363636363,
      "accuracy_std_dev": 0.006091232690182722,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.005402353535353536,
      "avg_duration_per_task_ms": 380915.31404958677,
      "avg_output_tokens_per_task": 29413.866391184572,
      "last_run_timestamp": "2026-10-03T09:51:11.430Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Ionstream",
      "accuracy": 0.8787874999999999,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04341706060606061,
      "avg_duration_per_task_ms": 244250.4595959596,
      "avg_output_tokens_per_task": 37643.704545454544,
      "last_run_timestamp": "2026-10-03T10:51:03.639Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Fireworks",
      "accuracy": 0.8787801424227066,
      "accuracy_std_dev": 0.016401176693323544,
      "run_count": 11,
      "total_questions": 1973,
      "avg_cost_per_task_usd": 0.011341015205271159,
      "avg_duration_per_task_ms": 261740.6791687785,
      "avg_output_tokens_per_task": 24571.48302078054,
      "last_run_timestamp": "2026-09-19T08:54:58.652Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Venice",
      "accuracy": 0.8772726000000001,
      "accuracy_std_dev": 0.03313293645362571,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.14829554040404042,
      "avg_duration_per_task_ms": 354670.0207070707,
      "avg_output_tokens_per_task": 34320.64595959596,
      "last_run_timestamp": "2026-10-03T17:43:07.447Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Reka",
      "accuracy": 0.8772070529695024,
      "accuracy_std_dev": 0.03893659390197737,
      "run_count": 4,
      "total_questions": 623,
      "avg_cost_per_task_usd": 0.09425666452648475,
      "avg_duration_per_task_ms": 461375.2568218299,
      "avg_output_tokens_per_task": 23922.329052969504,
      "last_run_timestamp": "2026-09-30T11:37:23.219Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.8771043333333335,
      "accuracy_std_dev": 0.007896528133018712,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.003143374579124579,
      "avg_duration_per_task_ms": 35450.33585858586,
      "avg_output_tokens_per_task": 5733.669191919192,
      "last_run_timestamp": "2026-10-01T08:25:15.234Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "GMICloud",
      "accuracy": 0.877104,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.07509321212121212,
      "avg_duration_per_task_ms": 263175.32828282827,
      "avg_output_tokens_per_task": 20738.459595959597,
      "last_run_timestamp": "2026-08-29T19:12:43.653Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8770865174506829,
      "accuracy_std_dev": 0.007621402236465436,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.11726762670713202,
      "avg_duration_per_task_ms": 751432.9919069296,
      "avg_output_tokens_per_task": 30746.914516944868,
      "last_run_timestamp": "2026-10-03T16:47:49.662Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Together",
      "accuracy": 0.876843640328698,
      "accuracy_std_dev": 0.014698550138959288,
      "run_count": 8,
      "total_questions": 1582,
      "avg_cost_per_task_usd": 0.08638480594184576,
      "avg_duration_per_task_ms": 222833.85398230088,
      "avg_output_tokens_per_task": 21940.641592920354,
      "last_run_timestamp": "2026-10-03T09:40:59.529Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AkashML",
      "accuracy": 0.8765966933797908,
      "accuracy_std_dev": 0.020189124617118887,
      "run_count": 6,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.004125730836236933,
      "avg_duration_per_task_ms": 453615.28397212544,
      "avg_output_tokens_per_task": 16054.564459930314,
      "last_run_timestamp": "2026-09-05T08:24:32.148Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Google",
      "accuracy": 0.8765431666666668,
      "accuracy_std_dev": 0.03217293038884639,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.11818556060606061,
      "avg_duration_per_task_ms": 28260.9031986532,
      "avg_output_tokens_per_task": 2161.995791245791,
      "last_run_timestamp": "2026-10-01T08:25:56.890Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Parasail",
      "accuracy": 0.8761782480719793,
      "accuracy_std_dev": 0.00788648157846706,
      "run_count": 4,
      "total_questions": 778,
      "avg_cost_per_task_usd": 0.0930879910025707,
      "avg_duration_per_task_ms": 616690.5848329049,
      "avg_output_tokens_per_task": 27406.845758354757,
      "last_run_timestamp": "2026-08-22T09:51:24.721Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Wafer",
      "accuracy": 0.8757659457720588,
      "accuracy_std_dev": 0.09251418420879137,
      "run_count": 11,
      "total_questions": 2176,
      "avg_cost_per_task_usd": 0.2399917977941176,
      "avg_duration_per_task_ms": 482095.9535845588,
      "avg_output_tokens_per_task": 59848.47564338235,
      "last_run_timestamp": "2026-10-03T10:18:13.963Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Novita",
      "accuracy": 0.8757225447976879,
      "accuracy_std_dev": 0.004672056689969172,
      "run_count": 7,
      "total_questions": 1384,
      "avg_cost_per_task_usd": 0.12312977673410404,
      "avg_duration_per_task_ms": 937524.2485549133,
      "avg_output_tokens_per_task": 43259.596098265894,
      "last_run_timestamp": "2026-10-03T17:24:45.959Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Novita",
      "accuracy": 0.8757013333333332,
      "accuracy_std_dev": 0.025185257232930715,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.09241545538720539,
      "avg_duration_per_task_ms": 486271.6456228956,
      "avg_output_tokens_per_task": 35368.11616161616,
      "last_run_timestamp": "2026-10-03T17:33:23.858Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Crusoe",
      "accuracy": 0.875421,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.04712759090909091,
      "avg_duration_per_task_ms": 154613.29292929292,
      "avg_output_tokens_per_task": 39139.681818181816,
      "last_run_timestamp": "2026-10-08T18:00:04.415Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "WandB",
      "accuracy": 0.875421,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.1538974696969697,
      "avg_duration_per_task_ms": 363407.6313131313,
      "avg_output_tokens_per_task": 34764.520202020205,
      "last_run_timestamp": "2026-07-18T08:26:40.271Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Ionstream",
      "accuracy": 0.8754206666666666,
      "accuracy_std_dev": 0.00901385812081723,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.061844909090909096,
      "avg_duration_per_task_ms": 262173.4595959596,
      "avg_output_tokens_per_task": 22965.14478114478,
      "last_run_timestamp": "2026-09-26T09:19:06.449Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.8753905728900255,
      "accuracy_std_dev": 0.012094409094012564,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.035846069053708435,
      "avg_duration_per_task_ms": 251997.9505541347,
      "avg_output_tokens_per_task": 29905.09377664109,
      "last_run_timestamp": "2026-10-03T12:29:29.078Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Alibaba",
      "accuracy": 0.8749397142857143,
      "accuracy_std_dev": 0.007830033729434499,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.049728212842712845,
      "avg_duration_per_task_ms": 278314.67604617606,
      "avg_output_tokens_per_task": 21125.652958152958,
      "last_run_timestamp": "2026-10-03T17:26:10.527Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "GMICloud",
      "accuracy": 0.8748596666666666,
      "accuracy_std_dev": 0.03480145848157005,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.10809361952861953,
      "avg_duration_per_task_ms": 516538.27384960715,
      "avg_output_tokens_per_task": 34956.31930415264,
      "last_run_timestamp": "2026-10-03T09:34:05.323Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Inceptron",
      "accuracy": 0.8747592697495183,
      "accuracy_std_dev": 0.04128530357388087,
      "run_count": 4,
      "total_questions": 519,
      "avg_cost_per_task_usd": 0.012099612716763006,
      "avg_duration_per_task_ms": 465229.36030828516,
      "avg_output_tokens_per_task": 26222.763005780347,
      "last_run_timestamp": "2026-10-03T09:28:13.430Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Parasail",
      "accuracy": 0.8747029593082402,
      "accuracy_std_dev": 0.008749466763180506,
      "run_count": 5,
      "total_questions": 983,
      "avg_cost_per_task_usd": 0.1150072502543235,
      "avg_duration_per_task_ms": 564566.1881993896,
      "avg_output_tokens_per_task": 32828.28179043744,
      "last_run_timestamp": "2026-10-03T16:53:50.088Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "auto-routing",
      "accuracy": 0.8746725555555557,
      "accuracy_std_dev": 0.018426695846040054,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.13061521717171717,
      "avg_duration_per_task_ms": 63177.591470258136,
      "avg_output_tokens_per_task": 5005.783950617284,
      "last_run_timestamp": "2026-10-03T09:54:43.475Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Makora",
      "accuracy": 0.8745317312013828,
      "accuracy_std_dev": 0.011186512454241005,
      "run_count": 6,
      "total_questions": 1157,
      "avg_cost_per_task_usd": 0.0043274416594641305,
      "avg_duration_per_task_ms": 185028.52031114954,
      "avg_output_tokens_per_task": 21870.51080380294,
      "last_run_timestamp": "2026-09-26T08:53:59.525Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8744108,
      "accuracy_std_dev": 0.01990792291928014,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.008221054545454545,
      "avg_duration_per_task_ms": 359745.57575757575,
      "avg_output_tokens_per_task": 29012.32424242424,
      "last_run_timestamp": "2026-10-03T17:03:58.997Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Decart",
      "accuracy": 0.8743913028237584,
      "accuracy_std_dev": 0.02063518118543075,
      "run_count": 6,
      "total_questions": 1027,
      "avg_cost_per_task_usd": 0.1379305988315482,
      "avg_duration_per_task_ms": 596483.0662122688,
      "avg_output_tokens_per_task": 50109.39727361246,
      "last_run_timestamp": "2026-10-03T17:01:24.374Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "CoreWeave",
      "accuracy": 0.8743494545454547,
      "accuracy_std_dev": 0.0150964078771537,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.1106280968778696,
      "avg_duration_per_task_ms": 234771.5486685032,
      "avg_output_tokens_per_task": 31461.72359963269,
      "last_run_timestamp": "2026-10-03T16:40:49.490Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8742763324913168,
      "accuracy_std_dev": 0.03081876645203374,
      "run_count": 16,
      "total_questions": 3167,
      "avg_cost_per_task_usd": 0.14868513672245026,
      "avg_duration_per_task_ms": 406084.6277233975,
      "avg_output_tokens_per_task": 11240.50994632144,
      "last_run_timestamp": "2026-10-05T20:36:24.348Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Phala",
      "accuracy": 0.8742182857142856,
      "accuracy_std_dev": 0.005225109204991012,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.07788331168831168,
      "avg_duration_per_task_ms": 336872.55194805196,
      "avg_output_tokens_per_task": 22144.56132756133,
      "last_run_timestamp": "2026-10-03T19:00:32.081Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "WandB",
      "accuracy": 0.873737,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008290222222222223,
      "avg_duration_per_task_ms": 529447,
      "avg_output_tokens_per_task": 29241.171717171717,
      "last_run_timestamp": "2026-07-18T08:31:23.912Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "auto-routing",
      "accuracy": 0.873737,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0029526111111111116,
      "avg_duration_per_task_ms": 49722.71717171717,
      "avg_output_tokens_per_task": 5746.707070707071,
      "last_run_timestamp": "2026-10-01T08:23:37.659Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.7-plus-20260602",
      "model_slug": "qwen/qwen3.7-plus",
      "model_name": "Qwen: Qwen3.7 Plus",
      "provider_name": "Alibaba",
      "accuracy": 0.873737,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.045554454545454544,
      "avg_duration_per_task_ms": 605480.6616161616,
      "avg_output_tokens_per_task": 35385.22727272727,
      "last_run_timestamp": "2026-10-01T08:40:06.225Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "GMICloud",
      "accuracy": 0.8736602893401014,
      "accuracy_std_dev": 0.018240759785588822,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.033190517766497465,
      "avg_duration_per_task_ms": 731117.2707275804,
      "avg_output_tokens_per_task": 38434.73942470389,
      "last_run_timestamp": "2026-10-03T11:12:58.942Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "auto-routing",
      "accuracy": 0.873140968503937,
      "accuracy_std_dev": 0.008141500000000024,
      "run_count": 2,
      "total_questions": 381,
      "avg_cost_per_task_usd": 0.10504513910761155,
      "avg_duration_per_task_ms": 1003143.2047244095,
      "avg_output_tokens_per_task": 42008.77427821522,
      "last_run_timestamp": "2026-08-01T17:53:10.275Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Z.AI",
      "accuracy": 0.8727401757049892,
      "accuracy_std_dev": 0.012263522167761748,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.00852685032537961,
      "avg_duration_per_task_ms": 434756.32754880696,
      "avg_output_tokens_per_task": 21739.99276934201,
      "last_run_timestamp": "2026-10-03T10:06:09.405Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "ModelRun",
      "accuracy": 0.8727272475524476,
      "accuracy_std_dev": 0.029233686851601524,
      "run_count": 8,
      "total_questions": 1430,
      "avg_cost_per_task_usd": 0.09700304615384615,
      "avg_duration_per_task_ms": 194874.37272727274,
      "avg_output_tokens_per_task": 26397.99300699301,
      "last_run_timestamp": "2026-09-26T09:19:13.012Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8726482966714906,
      "accuracy_std_dev": 0.007320320889029898,
      "run_count": 7,
      "total_questions": 1382,
      "avg_cost_per_task_usd": 0.0685544536903039,
      "avg_duration_per_task_ms": 284232.5897250362,
      "avg_output_tokens_per_task": 11158.96454413893,
      "last_run_timestamp": "2026-10-08T22:30:47.735Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.8724745,
      "accuracy_std_dev": 0.008858340547190563,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.008325782828282828,
      "avg_duration_per_task_ms": 77307.88005050505,
      "avg_output_tokens_per_task": 7842.704545454545,
      "last_run_timestamp": "2026-10-02T23:19:10.742Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Moonshot AI",
      "accuracy": 0.8723343333333334,
      "accuracy_std_dev": 0.015226503403240287,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.2220637457912458,
      "avg_duration_per_task_ms": 974700.601010101,
      "avg_output_tokens_per_task": 55783.88468013468,
      "last_run_timestamp": "2026-10-03T16:37:45.002Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.8722849626411816,
      "accuracy_std_dev": 0.03667207040635816,
      "run_count": 6,
      "total_questions": 1151,
      "avg_cost_per_task_usd": 0.01310653171155517,
      "avg_duration_per_task_ms": 596394.4969591659,
      "avg_output_tokens_per_task": 34928.014769765425,
      "last_run_timestamp": "2026-10-03T09:08:44.457Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Makora",
      "accuracy": 0.8722010494296577,
      "accuracy_std_dev": 0.0020781669055925365,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.045619747782002534,
      "avg_duration_per_task_ms": 189717.51711026617,
      "avg_output_tokens_per_task": 39672.14828897338,
      "last_run_timestamp": "2026-10-03T10:44:50.348Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8720540000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.004176666666666667,
      "avg_duration_per_task_ms": 40447.64646464647,
      "avg_output_tokens_per_task": 8116.434343434344,
      "last_run_timestamp": "2026-10-08T06:00:12.133Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Morph",
      "accuracy": 0.871875,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 160,
      "avg_cost_per_task_usd": 0.04507655,
      "avg_duration_per_task_ms": 1504372.61875,
      "avg_output_tokens_per_task": 18693.5375,
      "last_run_timestamp": "2026-08-27T15:00:05.074Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Together",
      "accuracy": 0.8715861506361324,
      "accuracy_std_dev": 0.009632287238761115,
      "run_count": 10,
      "total_questions": 1965,
      "avg_cost_per_task_usd": 0.008127365394402036,
      "avg_duration_per_task_ms": 324970.47837150126,
      "avg_output_tokens_per_task": 28742.415267175573,
      "last_run_timestamp": "2026-10-03T09:52:27.516Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8714926666666667,
      "accuracy_std_dev": 0.0028614651647170092,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.004327225589225589,
      "avg_duration_per_task_ms": 40230.3164983165,
      "avg_output_tokens_per_task": 7890.072390572391,
      "last_run_timestamp": "2026-10-08T06:00:12.145Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Chutes",
      "accuracy": 0.8714225,
      "accuracy_std_dev": 0.004885840587862051,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.06337124747474747,
      "avg_duration_per_task_ms": 284598.4861111111,
      "avg_output_tokens_per_task": 21034.015151515152,
      "last_run_timestamp": "2026-08-15T15:55:02.752Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Wafer",
      "accuracy": 0.8710353744745891,
      "accuracy_std_dev": 0.16260220955169502,
      "run_count": 15,
      "total_questions": 2617,
      "avg_cost_per_task_usd": 0.142784888803974,
      "avg_duration_per_task_ms": 265667.35957202903,
      "avg_output_tokens_per_task": 10833.450133740926,
      "last_run_timestamp": "2026-10-03T10:04:25.516Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "Together",
      "accuracy": 0.8710116425661915,
      "accuracy_std_dev": 0.01025225770257459,
      "run_count": 5,
      "total_questions": 982,
      "avg_cost_per_task_usd": 0.025059682281059068,
      "avg_duration_per_task_ms": 328330.6405295316,
      "avg_output_tokens_per_task": 20642.19348268839,
      "last_run_timestamp": "2026-09-12T08:54:29.103Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.8708512857142857,
      "accuracy_std_dev": 0.01450202738254489,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.026409178932178933,
      "avg_duration_per_task_ms": 705209.8095238095,
      "avg_output_tokens_per_task": 34652.24603174603,
      "last_run_timestamp": "2026-10-03T16:57:03.670Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8705104999999999,
      "accuracy_std_dev": 0.07295362229012346,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.12380340740740742,
      "avg_duration_per_task_ms": 1287771.0867003368,
      "avg_output_tokens_per_task": 54917.97643097643,
      "last_run_timestamp": "2026-10-03T17:53:34.568Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "BaseTen",
      "accuracy": 0.8703701666666666,
      "accuracy_std_dev": 0.006941288747215641,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.024583196127946128,
      "avg_duration_per_task_ms": 150981.7457912458,
      "avg_output_tokens_per_task": 20283.477272727272,
      "last_run_timestamp": "2026-09-19T08:49:51.735Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Venice",
      "accuracy": 0.8699495000000002,
      "accuracy_std_dev": 0.018722910197135487,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.030484161616161617,
      "avg_duration_per_task_ms": 134273.03535353535,
      "avg_output_tokens_per_task": 21527.786616161615,
      "last_run_timestamp": "2026-10-03T12:27:06.948Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Ionstream",
      "accuracy": 0.8699495,
      "accuracy_std_dev": 0.002104500000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.00502325,
      "avg_duration_per_task_ms": 389434.8611111111,
      "avg_output_tokens_per_task": 17944.10101010101,
      "last_run_timestamp": "2026-08-08T10:05:07.443Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Parasail",
      "accuracy": 0.8698091666666667,
      "accuracy_std_dev": 0.0056397381268006655,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.008173530303030302,
      "avg_duration_per_task_ms": 420480.94865319866,
      "avg_output_tokens_per_task": 28865.37626262626,
      "last_run_timestamp": "2026-10-03T16:58:32.515Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.8695285000000001,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.10538721464646464,
      "avg_duration_per_task_ms": 581981.0101010101,
      "avg_output_tokens_per_task": 34695.7398989899,
      "last_run_timestamp": "2026-08-08T09:03:54.762Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Together",
      "accuracy": 0.8695285000000001,
      "accuracy_std_dev": 0.010411879221190886,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07784478367003367,
      "avg_duration_per_task_ms": 107880.84595959596,
      "avg_output_tokens_per_task": 12619.909932659933,
      "last_run_timestamp": "2026-10-03T10:07:47.560Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DeepInfra",
      "accuracy": 0.8693955131578947,
      "accuracy_std_dev": 0.019648291273268125,
      "run_count": 7,
      "total_questions": 1368,
      "avg_cost_per_task_usd": 0.07663164254385965,
      "avg_duration_per_task_ms": 433978.8881578947,
      "avg_output_tokens_per_task": 23705.32821637427,
      "last_run_timestamp": "2026-10-08T23:14:17.356Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "CoreWeave",
      "accuracy": 0.8692988181818181,
      "accuracy_std_dev": 0.01574807678773497,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.008268114784205692,
      "avg_duration_per_task_ms": 218163.90817263545,
      "avg_output_tokens_per_task": 29305.03948576676,
      "last_run_timestamp": "2026-10-03T10:06:44.341Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Sail Research",
      "accuracy": 0.8692809647058825,
      "accuracy_std_dev": 0.009104176033557385,
      "run_count": 4,
      "total_questions": 765,
      "avg_cost_per_task_usd": 0.0047775359477124175,
      "avg_duration_per_task_ms": 482494.045751634,
      "avg_output_tokens_per_task": 26353.58431372549,
      "last_run_timestamp": "2026-08-22T16:48:48.690Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Venice",
      "accuracy": 0.869276,
      "accuracy_std_dev": 0.008908612349855623,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.006512984848484849,
      "avg_duration_per_task_ms": 255242.29242424242,
      "avg_output_tokens_per_task": 18235.289393939394,
      "last_run_timestamp": "2026-10-03T09:02:06.104Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "OpenAI",
      "accuracy": 0.8692479999999999,
      "accuracy_std_dev": 0.004827531943619518,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.22171807575757574,
      "avg_duration_per_task_ms": 260359.50673400675,
      "avg_output_tokens_per_task": 22076.11111111111,
      "last_run_timestamp": "2026-10-01T16:12:49.000Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "auto-routing",
      "accuracy": 0.8692478333333333,
      "accuracy_std_dev": 0.012955574635611068,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.08235681228956229,
      "avg_duration_per_task_ms": 202641.79545454544,
      "avg_output_tokens_per_task": 13365.021885521886,
      "last_run_timestamp": "2026-10-03T09:43:24.655Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Phala",
      "accuracy": 0.8691446346328195,
      "accuracy_std_dev": 0.025234963139792104,
      "run_count": 6,
      "total_questions": 1103,
      "avg_cost_per_task_usd": 0.01014118223028105,
      "avg_duration_per_task_ms": 374535.42611060746,
      "avg_output_tokens_per_task": 21668.27742520399,
      "last_run_timestamp": "2026-10-03T09:34:29.241Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.8689928181818182,
      "accuracy_std_dev": 0.05184823150279035,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.07711455371900827,
      "avg_duration_per_task_ms": 533320.4696969697,
      "avg_output_tokens_per_task": 38582.5886134068,
      "last_run_timestamp": "2026-10-08T22:52:10.694Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Ambient",
      "accuracy": 0.868687,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.13888562626262627,
      "avg_duration_per_task_ms": 1028019.2020202021,
      "avg_output_tokens_per_task": 46184.969696969696,
      "last_run_timestamp": "2026-07-18T16:16:27.436Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "Azure",
      "accuracy": 0.8686867500000001,
      "accuracy_std_dev": 0.009143975622643609,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.2541581641414141,
      "avg_duration_per_task_ms": 251149.7840909091,
      "avg_output_tokens_per_task": 24116.972222222223,
      "last_run_timestamp": "2026-10-01T16:14:35.237Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "DeepInfra",
      "accuracy": 0.868234624430956,
      "accuracy_std_dev": 0.009898395121653556,
      "run_count": 8,
      "total_questions": 1318,
      "avg_cost_per_task_usd": 0.05128331487101669,
      "avg_duration_per_task_ms": 415827.8520485584,
      "avg_output_tokens_per_task": 17002.027314112293,
      "last_run_timestamp": "2026-10-08T23:30:45.096Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Novita",
      "accuracy": 0.8679651428571429,
      "accuracy_std_dev": 0.004916957259716384,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.07019697691197692,
      "avg_duration_per_task_ms": 226467.18253968254,
      "avg_output_tokens_per_task": 11432.188311688311,
      "last_run_timestamp": "2026-10-03T09:14:11.063Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Alibaba",
      "accuracy": 0.8678913937007873,
      "accuracy_std_dev": 0.012999770226334775,
      "run_count": 8,
      "total_questions": 1524,
      "avg_cost_per_task_usd": 0.008103982283464567,
      "avg_duration_per_task_ms": 243353.82217847768,
      "avg_output_tokens_per_task": 29838.47900262467,
      "last_run_timestamp": "2026-10-09T21:00:06.195Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Parasail",
      "accuracy": 0.8678219449541285,
      "accuracy_std_dev": 0.011378331853132084,
      "run_count": 5,
      "total_questions": 981,
      "avg_cost_per_task_usd": 0.1803859877675841,
      "avg_duration_per_task_ms": 784706.3690112131,
      "avg_output_tokens_per_task": 52291.86748216106,
      "last_run_timestamp": "2026-10-03T17:02:58.648Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.8677916075949366,
      "accuracy_std_dev": 0.019070276800560625,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.057791014345991557,
      "avg_duration_per_task_ms": 510760.66835443035,
      "avg_output_tokens_per_task": 33916.00421940928,
      "last_run_timestamp": "2026-10-03T17:23:31.069Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "DeepInfra",
      "accuracy": 0.867634625,
      "accuracy_std_dev": 0.007431114989311851,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.025694838383838382,
      "avg_duration_per_task_ms": 248942.36553030304,
      "avg_output_tokens_per_task": 21167.008207070707,
      "last_run_timestamp": "2026-10-08T22:14:16.087Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Modal",
      "accuracy": 0.8675644999999997,
      "accuracy_std_dev": 0.008112634749779732,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07749805892255893,
      "avg_duration_per_task_ms": 84869.17171717172,
      "avg_output_tokens_per_task": 12670.825757575758,
      "last_run_timestamp": "2026-10-03T09:34:51.690Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Inceptron",
      "accuracy": 0.8670031666666665,
      "accuracy_std_dev": 0.018364664062783427,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1649758611111111,
      "avg_duration_per_task_ms": 1021779.7348484849,
      "avg_output_tokens_per_task": 53789.280303030304,
      "last_run_timestamp": "2026-10-03T17:43:48.193Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Novita",
      "accuracy": 0.8670031666666665,
      "accuracy_std_dev": 0.02310358935617773,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.19514486279461277,
      "avg_duration_per_task_ms": 1047935.0816498316,
      "avg_output_tokens_per_task": 57929.56986531986,
      "last_run_timestamp": "2026-10-03T17:03:53.018Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "WandB",
      "accuracy": 0.867003,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.11555696464646466,
      "avg_duration_per_task_ms": 292702.6111111111,
      "avg_output_tokens_per_task": 29334.075757575756,
      "last_run_timestamp": "2026-07-18T08:27:05.304Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Mancer 2",
      "accuracy": 0.866915421641791,
      "accuracy_std_dev": 0.02922832992755772,
      "run_count": 12,
      "total_questions": 2144,
      "avg_cost_per_task_usd": 0.010834175839552238,
      "avg_duration_per_task_ms": 328640.2140858209,
      "avg_output_tokens_per_task": 19735.994869402984,
      "last_run_timestamp": "2026-10-03T09:27:10.840Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Inceptron",
      "accuracy": 0.866685276559865,
      "accuracy_std_dev": 0.03454921866619338,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.11449315514333895,
      "avg_duration_per_task_ms": 640608.7127599776,
      "avg_output_tokens_per_task": 38548.483417650365,
      "last_run_timestamp": "2026-10-03T16:56:36.095Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepSeek",
      "accuracy": 0.8661615,
      "accuracy_std_dev": 0.012626499999999985,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.030201459595959598,
      "avg_duration_per_task_ms": 496509.8611111111,
      "avg_output_tokens_per_task": 34831.292929292926,
      "last_run_timestamp": "2026-08-08T09:24:32.269Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.8660412857142855,
      "accuracy_std_dev": 0.006542422852086991,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0036472380952380944,
      "avg_duration_per_task_ms": 423809.5887445887,
      "avg_output_tokens_per_task": 29733.171717171717,
      "last_run_timestamp": "2026-10-03T17:32:07.433Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Google",
      "accuracy": 0.8658809999999999,
      "accuracy_std_dev": 0.011527973831800053,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.004285917508417508,
      "avg_duration_per_task_ms": 35717.200336700334,
      "avg_output_tokens_per_task": 7821.843434343435,
      "last_run_timestamp": "2026-10-08T06:00:12.146Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8658007142857141,
      "accuracy_std_dev": 0.009679853093842238,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.06976018037518038,
      "avg_duration_per_task_ms": 223743.95815295816,
      "avg_output_tokens_per_task": 11357.300144300145,
      "last_run_timestamp": "2026-10-03T09:30:30.956Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Fireworks",
      "accuracy": 0.865689794871795,
      "accuracy_std_dev": 0.004877364201651907,
      "run_count": 3,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.005803672161172161,
      "avg_duration_per_task_ms": 153675.5,
      "avg_output_tokens_per_task": 20639.26923076923,
      "last_run_timestamp": "2026-08-08T09:25:45.716Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "auto-routing",
      "accuracy": 0.8656005,
      "accuracy_std_dev": 0.00803956708515244,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.02525451262626263,
      "avg_duration_per_task_ms": 233486.3627946128,
      "avg_output_tokens_per_task": 20811.0765993266,
      "last_run_timestamp": "2026-10-01T16:08:04.290Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Azure",
      "accuracy": 0.86532,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.004150199494949496,
      "avg_duration_per_task_ms": 35865.53535353536,
      "avg_output_tokens_per_task": 7670.747474747474,
      "last_run_timestamp": "2026-10-08T06:00:12.394Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "NextBit",
      "accuracy": 0.8653195,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.008587179292929294,
      "avg_duration_per_task_ms": 240899.30303030304,
      "avg_output_tokens_per_task": 30387.573232323233,
      "last_run_timestamp": "2026-08-29T18:38:15.907Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "BaseTen",
      "accuracy": 0.8651969388477632,
      "accuracy_std_dev": 0.0392226856144323,
      "run_count": 16,
      "total_questions": 3107,
      "avg_cost_per_task_usd": 0.004922328290955906,
      "avg_duration_per_task_ms": 103773.42162858062,
      "avg_output_tokens_per_task": 18562.228194399744,
      "last_run_timestamp": "2026-10-03T10:10:22.431Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.8649528761766837,
      "accuracy_std_dev": 0.01859749373173817,
      "run_count": 7,
      "total_questions": 1381,
      "avg_cost_per_task_usd": 0.0065336495293265755,
      "avg_duration_per_task_ms": 348157.4265025344,
      "avg_output_tokens_per_task": 21785.366401158582,
      "last_run_timestamp": "2026-10-03T10:20:35.518Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8649467344461305,
      "accuracy_std_dev": 0.10260638821189447,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.026698589782498738,
      "avg_duration_per_task_ms": 213357.39251390996,
      "avg_output_tokens_per_task": 23147.549317147194,
      "last_run_timestamp": "2026-10-03T17:06:35.452Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Reka",
      "accuracy": 0.864636,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 197,
      "avg_cost_per_task_usd": 0.07161002030456853,
      "avg_duration_per_task_ms": 868248.1269035533,
      "avg_output_tokens_per_task": 28363.54314720812,
      "last_run_timestamp": "2026-10-03T14:58:52.397Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "StreamLake",
      "accuracy": 0.8645714444444443,
      "accuracy_std_dev": 0.01709693694535898,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.08812040179573512,
      "avg_duration_per_task_ms": 199235.38664421998,
      "avg_output_tokens_per_task": 24344.7822671156,
      "last_run_timestamp": "2026-10-03T16:06:47.442Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.8643883344182262,
      "accuracy_std_dev": 0.01774344793127301,
      "run_count": 7,
      "total_questions": 1229,
      "avg_cost_per_task_usd": 0.0051325646867371845,
      "avg_duration_per_task_ms": 651023.4035801465,
      "avg_output_tokens_per_task": 28234.647681041497,
      "last_run_timestamp": "2026-10-08T23:29:09.758Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Inceptron",
      "accuracy": 0.8642566289198608,
      "accuracy_std_dev": 0.016122613680445513,
      "run_count": 6,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.08601719425087108,
      "avg_duration_per_task_ms": 391195.4764808362,
      "avg_output_tokens_per_task": 24802.604529616725,
      "last_run_timestamp": "2026-10-03T09:33:56.518Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Novita",
      "accuracy": 0.8642559853907962,
      "accuracy_std_dev": 0.01553435314759793,
      "run_count": 7,
      "total_questions": 1369,
      "avg_cost_per_task_usd": 0.07498870489408328,
      "avg_duration_per_task_ms": 273379.0087655223,
      "avg_output_tokens_per_task": 20706.99488677867,
      "last_run_timestamp": "2026-10-03T17:17:14.917Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Baidu",
      "accuracy": 0.8641975,
      "accuracy_std_dev": 0.020610892069890297,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.16472780976430976,
      "avg_duration_per_task_ms": 1088563.0942760943,
      "avg_output_tokens_per_task": 59591.08249158249,
      "last_run_timestamp": "2026-10-03T17:43:01.166Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8641171428571428,
      "accuracy_std_dev": 0.010825054462793655,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.008647630591630592,
      "avg_duration_per_task_ms": 270553.19336219336,
      "avg_output_tokens_per_task": 30573.133477633477,
      "last_run_timestamp": "2026-10-03T17:35:32.100Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "InferenceNet",
      "accuracy": 0.864098116850127,
      "accuracy_std_dev": 0.012040853347389158,
      "run_count": 6,
      "total_questions": 1181,
      "avg_cost_per_task_usd": 0.06984978577476715,
      "avg_duration_per_task_ms": 203374.89754445385,
      "avg_output_tokens_per_task": 24608.592718035565,
      "last_run_timestamp": "2026-10-03T09:19:06.149Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8639166666666666,
      "accuracy_std_dev": 0.022215769209180104,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1178505563973064,
      "avg_duration_per_task_ms": 576577.5420875421,
      "avg_output_tokens_per_task": 35099.228114478115,
      "last_run_timestamp": "2026-10-03T17:19:05.730Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.863636,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.003967691919191919,
      "avg_duration_per_task_ms": 39901.43939393939,
      "avg_output_tokens_per_task": 7695.671717171717,
      "last_run_timestamp": "2026-10-08T06:00:12.135Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Morph",
      "accuracy": 0.863636,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008496712121212122,
      "avg_duration_per_task_ms": 739364.4242424242,
      "avg_output_tokens_per_task": 30053.868686868685,
      "last_run_timestamp": "2026-07-18T08:30:06.336Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Nebius",
      "accuracy": 0.863636,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0938164393939394,
      "avg_duration_per_task_ms": 667365.3939393939,
      "avg_output_tokens_per_task": 23446.50505050505,
      "last_run_timestamp": "2026-09-10T01:44:54.935Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "OpenInference",
      "accuracy": 0.863636,
      "accuracy_std_dev": 0.007273551310513057,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.020974939393939393,
      "avg_duration_per_task_ms": 363755.6632996633,
      "avg_output_tokens_per_task": 34831.013468013465,
      "last_run_timestamp": "2026-10-03T11:53:46.126Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.863636,
      "accuracy_std_dev": 0.004761657064510171,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.22125699158249157,
      "avg_duration_per_task_ms": 259075.24915824915,
      "avg_output_tokens_per_task": 22030.511784511786,
      "last_run_timestamp": "2026-10-01T16:12:54.241Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Cloudflare",
      "accuracy": 0.8636106825127334,
      "accuracy_std_dev": 0.03982580471794763,
      "run_count": 9,
      "total_questions": 1767,
      "avg_cost_per_task_usd": 0.08851625127334464,
      "avg_duration_per_task_ms": 511731.79343520093,
      "avg_output_tokens_per_task": 22926.203735144314,
      "last_run_timestamp": "2026-10-03T17:07:22.491Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "Fireworks",
      "accuracy": 0.86321525,
      "accuracy_std_dev": 0.010127171825712258,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.022006194444444443,
      "avg_duration_per_task_ms": 265798.13636363635,
      "avg_output_tokens_per_task": 33108.22601010101,
      "last_run_timestamp": "2026-09-19T08:51:58.825Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8629148571428572,
      "accuracy_std_dev": 0.017903828987346902,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.07367307792207792,
      "avg_duration_per_task_ms": 277976.63924963924,
      "avg_output_tokens_per_task": 20972.66738816739,
      "last_run_timestamp": "2026-10-03T17:13:56.806Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "auto-routing",
      "accuracy": 0.8625952061068702,
      "accuracy_std_dev": 0.005736999999999992,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.043527526717557254,
      "avg_duration_per_task_ms": 365946.2213740458,
      "avg_output_tokens_per_task": 24956.239185750637,
      "last_run_timestamp": "2026-08-01T17:57:47.536Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "auto-routing",
      "accuracy": 0.8625138888888889,
      "accuracy_std_dev": 0.02716878807440808,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.00797921548821549,
      "avg_duration_per_task_ms": 504816.8608305275,
      "avg_output_tokens_per_task": 25080.10830527497,
      "last_run_timestamp": "2026-10-03T10:11:44.532Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Google",
      "accuracy": 0.8625137777777779,
      "accuracy_std_dev": 0.00880146315838422,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.8542647558922559,
      "avg_duration_per_task_ms": 483293.1627384961,
      "avg_output_tokens_per_task": 33979.86139169472,
      "last_run_timestamp": "2026-10-03T09:32:40.286Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DekaLLM",
      "accuracy": 0.8625002076923076,
      "accuracy_std_dev": 0.01982845351172571,
      "run_count": 3,
      "total_questions": 520,
      "avg_cost_per_task_usd": 0.017099969230769233,
      "avg_duration_per_task_ms": 517523.88653846155,
      "avg_output_tokens_per_task": 25671.957692307693,
      "last_run_timestamp": "2026-10-09T03:54:06.701Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Baidu",
      "accuracy": 0.8622331666666665,
      "accuracy_std_dev": 0.014944828898503846,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.047331268518518516,
      "avg_duration_per_task_ms": 475487.20454545453,
      "avg_output_tokens_per_task": 33266.92003367003,
      "last_run_timestamp": "2026-10-03T17:33:25.840Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.8619530000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008309641414141413,
      "avg_duration_per_task_ms": 730613.3737373737,
      "avg_output_tokens_per_task": 29358.666666666668,
      "last_run_timestamp": "2026-07-31T04:09:58.076Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Alibaba",
      "accuracy": 0.8618355304347826,
      "accuracy_std_dev": 0.022183205439048254,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.07637872463768115,
      "avg_duration_per_task_ms": 413540.6456521739,
      "avg_output_tokens_per_task": 32373.957971014494,
      "last_run_timestamp": "2026-10-10T00:00:05.170Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.8617657777777777,
      "accuracy_std_dev": 0.009167555942050057,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.8453705723905725,
      "avg_duration_per_task_ms": 415699.1722783389,
      "avg_output_tokens_per_task": 33623.734006734005,
      "last_run_timestamp": "2026-10-03T09:39:37.575Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.8617122857142857,
      "accuracy_std_dev": 0.015778023212360787,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.04849973448773449,
      "avg_duration_per_task_ms": 719703.6161616162,
      "avg_output_tokens_per_task": 33196.97546897547,
      "last_run_timestamp": "2026-10-03T17:03:53.250Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "BaseTen",
      "accuracy": 0.8616955969987624,
      "accuracy_std_dev": 0.021294415333698184,
      "run_count": 33,
      "total_questions": 6464,
      "avg_cost_per_task_usd": 0.2610163531868812,
      "avg_duration_per_task_ms": 338765.33369430696,
      "avg_output_tokens_per_task": 47192.89201732673,
      "last_run_timestamp": "2026-10-03T17:31:59.643Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8615786666666667,
      "accuracy_std_dev": 0.009382650075775203,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.8094348821548821,
      "avg_duration_per_task_ms": 461988.8821548821,
      "avg_output_tokens_per_task": 32205.768237934906,
      "last_run_timestamp": "2026-10-03T10:13:16.654Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Minimax",
      "accuracy": 0.8614377764705882,
      "accuracy_std_dev": 0.012431326538225896,
      "run_count": 4,
      "total_questions": 765,
      "avg_cost_per_task_usd": 0.051462403921568624,
      "avg_duration_per_task_ms": 413319.64183006535,
      "avg_output_tokens_per_task": 26795.732026143793,
      "last_run_timestamp": "2026-08-01T17:44:01.464Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Ambient",
      "accuracy": 0.8613915000000001,
      "accuracy_std_dev": 0.04603966955336814,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.004525512626262627,
      "avg_duration_per_task_ms": 565963.2584175084,
      "avg_output_tokens_per_task": 18978.082491582492,
      "last_run_timestamp": "2026-08-29T10:57:26.955Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "CoreWeave",
      "accuracy": 0.8613915000000001,
      "accuracy_std_dev": 0.015601910243620808,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.19432850925925926,
      "avg_duration_per_task_ms": 386703.5462962963,
      "avg_output_tokens_per_task": 56841.80892255892,
      "last_run_timestamp": "2026-10-03T16:58:25.704Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Morph",
      "accuracy": 0.8612792,
      "accuracy_std_dev": 0.03136197207702347,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.005120404040404041,
      "avg_duration_per_task_ms": 530889.794949495,
      "avg_output_tokens_per_task": 17542.636363636364,
      "last_run_timestamp": "2026-09-05T09:22:54.724Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Moonshot AI",
      "accuracy": 0.8609939949452401,
      "accuracy_std_dev": 0.034291003424934796,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.12772796882898063,
      "avg_duration_per_task_ms": 665372.3294018534,
      "avg_output_tokens_per_task": 42817.63689974726,
      "last_run_timestamp": "2026-08-29T18:25:54.650Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Together",
      "accuracy": 0.8609694285714284,
      "accuracy_std_dev": 0.023293052022008633,
      "run_count": 4,
      "total_questions": 784,
      "avg_cost_per_task_usd": 0.10234140178571428,
      "avg_duration_per_task_ms": 463624.993622449,
      "avg_output_tokens_per_task": 30066.539540816328,
      "last_run_timestamp": "2026-08-22T16:49:15.579Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "DekaLLM",
      "accuracy": 0.8608923543307087,
      "accuracy_std_dev": 0.017439499999999997,
      "run_count": 2,
      "total_questions": 254,
      "avg_cost_per_task_usd": 0.003837814960629921,
      "avg_duration_per_task_ms": 298144.1771653543,
      "avg_output_tokens_per_task": 11523.807086614173,
      "last_run_timestamp": "2026-10-03T17:04:09.464Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8607369444444446,
      "accuracy_std_dev": 0.008186229884332315,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.8829101122334456,
      "avg_duration_per_task_ms": 416465.39506172837,
      "avg_output_tokens_per_task": 33444.42985409652,
      "last_run_timestamp": "2026-10-03T09:27:25.920Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Fireworks",
      "accuracy": 0.8606576691394658,
      "accuracy_std_dev": 0.011001380598888357,
      "run_count": 24,
      "total_questions": 4044,
      "avg_cost_per_task_usd": 0.16706140232443123,
      "avg_duration_per_task_ms": 253403.07789317507,
      "avg_output_tokens_per_task": 30580.46612265084,
      "last_run_timestamp": "2026-09-26T08:54:22.466Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Relace",
      "accuracy": 0.8602695000000001,
      "accuracy_std_dev": 0.011784499999999976,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07283564141414141,
      "avg_duration_per_task_ms": 493206.3787878788,
      "avg_output_tokens_per_task": 56881.042929292926,
      "last_run_timestamp": "2026-10-03T17:41:27.941Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Baidu",
      "accuracy": 0.8602693333333333,
      "accuracy_std_dev": 0.008473596416057473,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0050054402356902354,
      "avg_duration_per_task_ms": 240643.9755892256,
      "avg_output_tokens_per_task": 28765.88973063973,
      "last_run_timestamp": "2026-10-03T16:57:52.806Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.8602692222222225,
      "accuracy_std_dev": 0.01392791126381983,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.7468858333333334,
      "avg_duration_per_task_ms": 493290.462962963,
      "avg_output_tokens_per_task": 29714.63973063973,
      "last_run_timestamp": "2026-10-03T09:34:14.060Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.8601822310030396,
      "accuracy_std_dev": 0.013547575585321513,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.007783161094224923,
      "avg_duration_per_task_ms": 526386.7112462006,
      "avg_output_tokens_per_task": 27793.246200607904,
      "last_run_timestamp": "2026-08-29T18:49:16.416Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ambient",
      "accuracy": 0.8599289999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 188,
      "avg_cost_per_task_usd": 0.023036143617021276,
      "avg_duration_per_task_ms": 1599986.2287234042,
      "avg_output_tokens_per_task": 29232.622340425532,
      "last_run_timestamp": "2026-07-18T08:57:54.026Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Venice",
      "accuracy": 0.8598355714285714,
      "accuracy_std_dev": 0.011856733262862344,
      "run_count": 6,
      "total_questions": 1176,
      "avg_cost_per_task_usd": 0.14287056972789117,
      "avg_duration_per_task_ms": 268723.07568027213,
      "avg_output_tokens_per_task": 28901.64880952381,
      "last_run_timestamp": "2026-10-03T09:42:23.512Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Z.AI",
      "accuracy": 0.8596616913043478,
      "accuracy_std_dev": 0.02724773313095075,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.10258134202898551,
      "avg_duration_per_task_ms": 343569.69492753624,
      "avg_output_tokens_per_task": 23077.884057971016,
      "last_run_timestamp": "2026-10-03T09:34:01.485Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8592168750000002,
      "accuracy_std_dev": 0.04194354929377548,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.17683030808080807,
      "avg_duration_per_task_ms": 45565.645833333336,
      "avg_output_tokens_per_task": 3337.2266414141413,
      "last_run_timestamp": "2026-10-03T09:18:12.856Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Alibaba",
      "accuracy": 0.8590668571428572,
      "accuracy_std_dev": 0.0075671013591461535,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.06653240404040404,
      "avg_duration_per_task_ms": 215829.40548340548,
      "avg_output_tokens_per_task": 10820.84126984127,
      "last_run_timestamp": "2026-10-03T10:11:33.574Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.858586,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.08651329797979797,
      "avg_duration_per_task_ms": 499241.5707070707,
      "avg_output_tokens_per_task": 35858.23737373737,
      "last_run_timestamp": "2026-10-04T03:06:10.317Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "WandB",
      "accuracy": 0.858586,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.2511099898989899,
      "avg_duration_per_task_ms": 290522.3080808081,
      "avg_output_tokens_per_task": 62590.06565656565,
      "last_run_timestamp": "2026-07-18T08:26:40.483Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.8578204545454546,
      "accuracy_std_dev": 0.010161020030789983,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.8425653397612488,
      "avg_duration_per_task_ms": 451495.49678604223,
      "avg_output_tokens_per_task": 33526.43617998163,
      "last_run_timestamp": "2026-10-03T10:12:25.141Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Baidu",
      "accuracy": 0.8578001999999999,
      "accuracy_std_dev": 0.012246153786393496,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.11574598114478116,
      "avg_duration_per_task_ms": 406292.1531986532,
      "avg_output_tokens_per_task": 41531.61616161616,
      "last_run_timestamp": "2026-10-03T17:10:11.381Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "StreamLake",
      "accuracy": 0.8576669736716133,
      "accuracy_std_dev": 0.03054982749271608,
      "run_count": 11,
      "total_questions": 2089,
      "avg_cost_per_task_usd": 0.025787480612733366,
      "avg_duration_per_task_ms": 476213.8056486357,
      "avg_output_tokens_per_task": 24583.015797032072,
      "last_run_timestamp": "2026-10-03T09:49:53.481Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "auto-routing",
      "accuracy": 0.8576504444444445,
      "accuracy_std_dev": 0.017188931406971766,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.09760986363636363,
      "avg_duration_per_task_ms": 514558.1818181818,
      "avg_output_tokens_per_task": 26225.028058361393,
      "last_run_timestamp": "2026-10-03T16:43:18.077Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "BaseTen",
      "accuracy": 0.857418,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 173,
      "avg_cost_per_task_usd": 0.08778269942196533,
      "avg_duration_per_task_ms": 317630.5202312139,
      "avg_output_tokens_per_task": 36575.57225433526,
      "last_run_timestamp": "2026-08-04T03:12:38.954Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Together",
      "accuracy": 0.8573280311284046,
      "accuracy_std_dev": 0.008776662705721345,
      "run_count": 4,
      "total_questions": 771,
      "avg_cost_per_task_usd": 0.11767329182879378,
      "avg_duration_per_task_ms": 355412.9636835279,
      "avg_output_tokens_per_task": 29743.00907911803,
      "last_run_timestamp": "2026-08-22T10:19:36.962Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Cloudflare",
      "accuracy": 0.8572655795356836,
      "accuracy_std_dev": 0.023144525715238454,
      "run_count": 6,
      "total_questions": 1163,
      "avg_cost_per_task_usd": 0.02012267153912296,
      "avg_duration_per_task_ms": 381128.8787618229,
      "avg_output_tokens_per_task": 23297.414445399827,
      "last_run_timestamp": "2026-10-04T03:08:02.396Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.8569023333333334,
      "accuracy_std_dev": 0.016379967131292483,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.01228687542087542,
      "avg_duration_per_task_ms": 639472.5816498316,
      "avg_output_tokens_per_task": 33004.78787878788,
      "last_run_timestamp": "2026-10-03T09:19:24.461Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "WandB",
      "accuracy": 0.8569019999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.006055888888888888,
      "avg_duration_per_task_ms": 490894.14646464644,
      "avg_output_tokens_per_task": 17030.60606060606,
      "last_run_timestamp": "2026-07-18T08:04:02.357Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "auto-routing",
      "accuracy": 0.8569019999999999,
      "accuracy_std_dev": 0.0033669999999999534,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20989276262626264,
      "avg_duration_per_task_ms": 263850.72222222225,
      "avg_output_tokens_per_task": 20893.070707070707,
      "last_run_timestamp": "2026-10-01T16:31:36.833Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Minimax",
      "accuracy": 0.8563952803943043,
      "accuracy_std_dev": 0.008474177580046515,
      "run_count": 14,
      "total_questions": 2739,
      "avg_cost_per_task_usd": 0.06252096166484118,
      "avg_duration_per_task_ms": 600270.9682365827,
      "avg_output_tokens_per_task": 34408.29098211026,
      "last_run_timestamp": "2026-10-03T17:02:38.592Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Novita",
      "accuracy": 0.8560605,
      "accuracy_std_dev": 0.012626499999999985,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02547615909090909,
      "avg_duration_per_task_ms": 266390.93434343435,
      "avg_output_tokens_per_task": 21055.325757575756,
      "last_run_timestamp": "2026-08-01T17:57:21.740Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "AkashML",
      "accuracy": 0.8557689653846154,
      "accuracy_std_dev": 0.016989351495046,
      "run_count": 3,
      "total_questions": 520,
      "avg_cost_per_task_usd": 0.06363424807692308,
      "avg_duration_per_task_ms": 459711.01153846155,
      "avg_output_tokens_per_task": 23017.521153846155,
      "last_run_timestamp": "2026-08-15T08:47:51.369Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "ModelRun",
      "accuracy": 0.855219,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.19856715151515153,
      "avg_duration_per_task_ms": 638184.691919192,
      "avg_output_tokens_per_task": 58287.33838383838,
      "last_run_timestamp": "2026-07-18T08:26:48.377Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Together",
      "accuracy": 0.8550479847715735,
      "accuracy_std_dev": 0.00436888279642389,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.24340484940778345,
      "avg_duration_per_task_ms": 425015.06429780036,
      "avg_output_tokens_per_task": 54319.2165820643,
      "last_run_timestamp": "2026-08-15T08:57:45.830Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Azure",
      "accuracy": 0.8550316666666666,
      "accuracy_std_dev": 0.01908702254878369,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.8377942901234569,
      "avg_duration_per_task_ms": 451775.2300785634,
      "avg_output_tokens_per_task": 33341.289001122335,
      "last_run_timestamp": "2026-10-03T09:45:17.956Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.8549221865284975,
      "accuracy_std_dev": 0.006241046279804878,
      "run_count": 3,
      "total_questions": 579,
      "avg_cost_per_task_usd": 0.02481670293609672,
      "avg_duration_per_task_ms": 142245.9913644214,
      "avg_output_tokens_per_task": 37982.34369602763,
      "last_run_timestamp": "2026-10-03T10:08:27.319Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Decart",
      "accuracy": 0.8548752091836734,
      "accuracy_std_dev": 0.01338517551144632,
      "run_count": 6,
      "total_questions": 1176,
      "avg_cost_per_task_usd": 0.1090017406462585,
      "avg_duration_per_task_ms": 191170.07142857142,
      "avg_output_tokens_per_task": 28383.128401360544,
      "last_run_timestamp": "2026-10-03T09:07:14.738Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Azure",
      "accuracy": 0.8548644736842105,
      "accuracy_std_dev": 0.024533883813741147,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.15887166959064328,
      "avg_duration_per_task_ms": 152796.97049441785,
      "avg_output_tokens_per_task": 15519.810738968634,
      "last_run_timestamp": "2026-10-03T09:27:53.801Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Ambient",
      "accuracy": 0.854800172815534,
      "accuracy_std_dev": 0.012641847777693532,
      "run_count": 8,
      "total_questions": 1545,
      "avg_cost_per_task_usd": 0.1002615708737864,
      "avg_duration_per_task_ms": 668327.793527508,
      "avg_output_tokens_per_task": 36663.679611650485,
      "last_run_timestamp": "2026-09-19T08:55:10.654Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Anthropic",
      "accuracy": 0.8547135999999999,
      "accuracy_std_dev": 0.03570492302526363,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.1789077186868687,
      "avg_duration_per_task_ms": 44801.785858585856,
      "avg_output_tokens_per_task": 3372.5843434343433,
      "last_run_timestamp": "2026-10-03T09:14:11.064Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Chutes",
      "accuracy": 0.854592,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.08009020408163266,
      "avg_duration_per_task_ms": 1565820.9795918367,
      "avg_output_tokens_per_task": 40916.48469387755,
      "last_run_timestamp": "2026-07-18T16:22:39.277Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "BaseTen",
      "accuracy": 0.8545491290566037,
      "accuracy_std_dev": 0.016485433040158266,
      "run_count": 21,
      "total_questions": 3975,
      "avg_cost_per_task_usd": 0.15811767823899373,
      "avg_duration_per_task_ms": 199794.28352201259,
      "avg_output_tokens_per_task": 28179.00402515723,
      "last_run_timestamp": "2026-10-03T09:59:27.517Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "auto-routing",
      "accuracy": 0.8545453999999999,
      "accuracy_std_dev": 0.04176428250119952,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.17536612575757574,
      "avg_duration_per_task_ms": 43901.6994949495,
      "avg_output_tokens_per_task": 3302.6388888888887,
      "last_run_timestamp": "2026-10-03T09:19:04.660Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.8545453999999999,
      "accuracy_std_dev": 0.008010288199559339,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.006016572727272727,
      "avg_duration_per_task_ms": 589898.7232323232,
      "avg_output_tokens_per_task": 28997.2,
      "last_run_timestamp": "2026-10-03T17:45:03.327Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Minimax",
      "accuracy": 0.8542019060913706,
      "accuracy_std_dev": 0.006301576321753829,
      "run_count": 12,
      "total_questions": 2364,
      "avg_cost_per_task_usd": 0.05433487690355329,
      "avg_duration_per_task_ms": 423724.30245346867,
      "avg_output_tokens_per_task": 27804.93189509306,
      "last_run_timestamp": "2026-08-29T19:17:08.288Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "auto-routing",
      "accuracy": 0.8540964999999999,
      "accuracy_std_dev": 0.014793979437032262,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.09252428872053871,
      "avg_duration_per_task_ms": 604262.7205387205,
      "avg_output_tokens_per_task": 27745.595117845118,
      "last_run_timestamp": "2026-10-03T09:59:46.432Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Azure",
      "accuracy": 0.8537222222222223,
      "accuracy_std_dev": 0.017912548902175803,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.7545602300785634,
      "avg_duration_per_task_ms": 501662.9904601571,
      "avg_output_tokens_per_task": 30067.62962962963,
      "last_run_timestamp": "2026-10-03T09:46:07.607Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cloudflare",
      "accuracy": 0.8535444426523295,
      "accuracy_std_dev": 0.03729714924788522,
      "run_count": 29,
      "total_questions": 5022,
      "avg_cost_per_task_usd": 0.012806513739546001,
      "avg_duration_per_task_ms": 408091.7891278375,
      "avg_output_tokens_per_task": 16584.42453205894,
      "last_run_timestamp": "2026-10-03T10:45:26.547Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Google",
      "accuracy": 0.8535352222222222,
      "accuracy_std_dev": 0.01802737382598276,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.8012126506734006,
      "avg_duration_per_task_ms": 478739.47053872055,
      "avg_output_tokens_per_task": 30372.589225589225,
      "last_run_timestamp": "2026-10-03T09:41:04.765Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8533247499999999,
      "accuracy_std_dev": 0.03034305490861954,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.8320753535353536,
      "avg_duration_per_task_ms": 427272.283459596,
      "avg_output_tokens_per_task": 33208.628787878784,
      "last_run_timestamp": "2026-10-03T09:44:19.593Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Azure",
      "accuracy": 0.8531986,
      "accuracy_std_dev": 0.030700851614246794,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.12264740303030303,
      "avg_duration_per_task_ms": 564764.8696969697,
      "avg_output_tokens_per_task": 32450.073737373736,
      "last_run_timestamp": "2026-10-03T16:54:37.572Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.8531398217391304,
      "accuracy_std_dev": 0.028934812364545037,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.012572355072463768,
      "avg_duration_per_task_ms": 541534.9311594203,
      "avg_output_tokens_per_task": 33634.00579710145,
      "last_run_timestamp": "2026-10-03T09:08:08.594Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.853024801369863,
      "accuracy_std_dev": 0.02604945551361019,
      "run_count": 6,
      "total_questions": 1168,
      "avg_cost_per_task_usd": 0.13878068065068494,
      "avg_duration_per_task_ms": 378053.25171232875,
      "avg_output_tokens_per_task": 31592.83133561644,
      "last_run_timestamp": "2026-10-03T09:40:06.763Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Together",
      "accuracy": 0.8528984782608696,
      "accuracy_std_dev": 0.014334577107719047,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.018874540579710147,
      "avg_duration_per_task_ms": 358751.04855072463,
      "avg_output_tokens_per_task": 37538.637681159424,
      "last_run_timestamp": "2026-10-03T10:13:52.682Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.8527699090909091,
      "accuracy_std_dev": 0.015435615118376213,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.7526788682277317,
      "avg_duration_per_task_ms": 500624.6831955923,
      "avg_output_tokens_per_task": 30015.131313131315,
      "last_run_timestamp": "2026-10-03T09:32:02.382Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Sail Research",
      "accuracy": 0.8522881577868852,
      "accuracy_std_dev": 0.009664339658766143,
      "run_count": 5,
      "total_questions": 976,
      "avg_cost_per_task_usd": 0.09954490881147542,
      "avg_duration_per_task_ms": 487265.8411885246,
      "avg_output_tokens_per_task": 30959.861680327867,
      "last_run_timestamp": "2026-08-29T11:19:17.684Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8522125357142856,
      "accuracy_std_dev": 0.025081070583327737,
      "run_count": 28,
      "total_questions": 5544,
      "avg_cost_per_task_usd": 0.1703613856421356,
      "avg_duration_per_task_ms": 157589.71464646465,
      "avg_output_tokens_per_task": 15942.862554112555,
      "last_run_timestamp": "2026-10-03T09:43:58.916Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "AtlasCloud",
      "accuracy": 0.851932886214442,
      "accuracy_std_dev": 0.02400418545476889,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.04229712910284464,
      "avg_duration_per_task_ms": 657747.1064916119,
      "avg_output_tokens_per_task": 35368.60831509847,
      "last_run_timestamp": "2026-10-03T16:51:59.968Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "BaseTen",
      "accuracy": 0.851852,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 144,
      "avg_cost_per_task_usd": 0.15792222916666668,
      "avg_duration_per_task_ms": 798429.2847222222,
      "avg_output_tokens_per_task": 42935.319444444445,
      "last_run_timestamp": "2026-07-18T08:29:37.313Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Ionstream",
      "accuracy": 0.851852,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.16585097979797978,
      "avg_duration_per_task_ms": 389107.60101010103,
      "avg_output_tokens_per_task": 37536.631313131315,
      "last_run_timestamp": "2026-08-04T03:12:10.622Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "Azure",
      "accuracy": 0.8518517499999999,
      "accuracy_std_dev": 0.0020620621201845155,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.21063902146464644,
      "avg_duration_per_task_ms": 251655.29545454544,
      "avg_output_tokens_per_task": 19971.159090909092,
      "last_run_timestamp": "2026-10-01T16:27:58.636Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Reka",
      "accuracy": 0.8518308614084509,
      "accuracy_std_dev": 0.02345464140331345,
      "run_count": 9,
      "total_questions": 1775,
      "avg_cost_per_task_usd": 0.006422030422535212,
      "avg_duration_per_task_ms": 229362.44,
      "avg_output_tokens_per_task": 12608.14704225352,
      "last_run_timestamp": "2026-10-03T13:30:57.186Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Near AI",
      "accuracy": 0.85145952259887,
      "accuracy_std_dev": 0.01855480961260448,
      "run_count": 5,
      "total_questions": 708,
      "avg_cost_per_task_usd": 0.003704364406779661,
      "avg_duration_per_task_ms": 411396.5324858757,
      "avg_output_tokens_per_task": 7770.766949152542,
      "last_run_timestamp": "2026-10-03T09:41:45.583Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Parasail",
      "accuracy": 0.8512626000000001,
      "accuracy_std_dev": 0.01008732084549707,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.040042293434343436,
      "avg_duration_per_task_ms": 396466.652020202,
      "avg_output_tokens_per_task": 40285.31212121212,
      "last_run_timestamp": "2026-10-03T19:06:53.371Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "GMICloud",
      "accuracy": 0.8508898571428573,
      "accuracy_std_dev": 0.059458570338462385,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.1082869906204906,
      "avg_duration_per_task_ms": 468673.2698412698,
      "avg_output_tokens_per_task": 27557.489898989897,
      "last_run_timestamp": "2026-10-03T17:44:09.402Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.8507314900265956,
      "accuracy_std_dev": 0.014339170843684643,
      "run_count": 8,
      "total_questions": 1504,
      "avg_cost_per_task_usd": 0.018270000664893617,
      "avg_duration_per_task_ms": 282286.72473404254,
      "avg_output_tokens_per_task": 34448.984707446805,
      "last_run_timestamp": "2026-10-03T09:44:00.880Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Mara",
      "accuracy": 0.850575,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 145,
      "avg_cost_per_task_usd": 0.015425489655172412,
      "avg_duration_per_task_ms": 62324.744827586204,
      "avg_output_tokens_per_task": 15913.096551724138,
      "last_run_timestamp": "2026-08-01T18:26:19.887Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8505745613793103,
      "accuracy_std_dev": 0.00859156025430222,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.05167572781609195,
      "avg_duration_per_task_ms": 279727.17655172414,
      "avg_output_tokens_per_task": 46309.96643678161,
      "last_run_timestamp": "2026-10-03T17:05:07.642Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "Novita",
      "accuracy": 0.8504390873786407,
      "accuracy_std_dev": 0.008326774223864035,
      "run_count": 11,
      "total_questions": 2163,
      "avg_cost_per_task_usd": 0.07101044845122514,
      "avg_duration_per_task_ms": 216557.44290337493,
      "avg_output_tokens_per_task": 22089.64678687009,
      "last_run_timestamp": "2026-10-03T09:46:11.970Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Chutes",
      "accuracy": 0.850427,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 156,
      "avg_cost_per_task_usd": 0.014483525641025641,
      "avg_duration_per_task_ms": 276126.60256410256,
      "avg_output_tokens_per_task": 12035.96794871795,
      "last_run_timestamp": "2026-07-18T16:53:38.018Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "Xiaomi",
      "accuracy": 0.8503366000000001,
      "accuracy_std_dev": 0.010921921719184751,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.028464721717171713,
      "avg_duration_per_task_ms": 704138.8641414142,
      "avg_output_tokens_per_task": 32456.946464646466,
      "last_run_timestamp": "2026-10-03T09:34:29.314Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "OpenAI",
      "accuracy": 0.8501685000000001,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20853091414141414,
      "avg_duration_per_task_ms": 280266.345959596,
      "avg_output_tokens_per_task": 20755.358585858587,
      "last_run_timestamp": "2026-10-01T16:20:44.419Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8499464590032153,
      "accuracy_std_dev": 0.022950616654380467,
      "run_count": 7,
      "total_questions": 1244,
      "avg_cost_per_task_usd": 0.010361853697749194,
      "avg_duration_per_task_ms": 360749.26848874596,
      "avg_output_tokens_per_task": 20512.316720257233,
      "last_run_timestamp": "2026-10-03T09:28:36.120Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Azure",
      "accuracy": 0.8498315000000001,
      "accuracy_std_dev": 0.03952326621940551,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.17816506616161618,
      "avg_duration_per_task_ms": 45707.15909090909,
      "avg_output_tokens_per_task": 3356.5939393939393,
      "last_run_timestamp": "2026-10-03T09:19:41.728Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Anthropic",
      "accuracy": 0.8498314,
      "accuracy_std_dev": 0.02688129444874262,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.15721394040404038,
      "avg_duration_per_task_ms": 152201.82676767677,
      "avg_output_tokens_per_task": 15506.534343434343,
      "last_run_timestamp": "2026-10-03T09:23:24.359Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Ambient",
      "accuracy": 0.849375142857143,
      "accuracy_std_dev": 0.05373653913902533,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.09692596960486322,
      "avg_duration_per_task_ms": 658494.2613981763,
      "avg_output_tokens_per_task": 27042.529888551166,
      "last_run_timestamp": "2026-08-29T11:31:24.094Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "auto-routing",
      "accuracy": 0.8493265454545452,
      "accuracy_std_dev": 0.019857803633761332,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.10014188016528924,
      "avg_duration_per_task_ms": 590927.261707989,
      "avg_output_tokens_per_task": 43625.35996326905,
      "last_run_timestamp": "2026-10-03T16:57:15.084Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8493151205479452,
      "accuracy_std_dev": 0.01972059190006447,
      "run_count": 3,
      "total_questions": 365,
      "avg_cost_per_task_usd": 0.06587333698630138,
      "avg_duration_per_task_ms": 279697.01369863015,
      "avg_output_tokens_per_task": 16490.610958904108,
      "last_run_timestamp": "2026-08-15T08:47:41.523Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Phala",
      "accuracy": 0.8492828305531167,
      "accuracy_std_dev": 0.030679553487986455,
      "run_count": 6,
      "total_questions": 1139,
      "avg_cost_per_task_usd": 0.1125236496927129,
      "avg_duration_per_task_ms": 549482.1018437225,
      "avg_output_tokens_per_task": 33342.05618964003,
      "last_run_timestamp": "2026-10-03T10:16:13.056Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8487653333333334,
      "accuracy_std_dev": 0.011791380929400174,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.01937495622895623,
      "avg_duration_per_task_ms": 442506.297979798,
      "avg_output_tokens_per_task": 38526.9595959596,
      "last_run_timestamp": "2026-10-03T09:49:43.878Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Baidu",
      "accuracy": 0.8486995349544073,
      "accuracy_std_dev": 0.009718197845279743,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.10196987639311045,
      "avg_duration_per_task_ms": 302885.430597771,
      "avg_output_tokens_per_task": 30304.676798378925,
      "last_run_timestamp": "2026-10-03T09:33:45.532Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Crusoe",
      "accuracy": 0.848597158186864,
      "accuracy_std_dev": 0.013754474089263532,
      "run_count": 6,
      "total_questions": 1081,
      "avg_cost_per_task_usd": 0.006980295097132284,
      "avg_duration_per_task_ms": 490407.8658649399,
      "avg_output_tokens_per_task": 17200.208140610546,
      "last_run_timestamp": "2026-10-03T16:56:50.425Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "auto-routing",
      "accuracy": 0.848485,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0012582575757575758,
      "avg_duration_per_task_ms": 88400.0101010101,
      "avg_output_tokens_per_task": 17501.79797979798,
      "last_run_timestamp": "2026-10-01T08:22:54.644Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Nebius",
      "accuracy": 0.848485,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.26182384343434345,
      "avg_duration_per_task_ms": 437826.7272727273,
      "avg_output_tokens_per_task": 59251.080808080806,
      "last_run_timestamp": "2026-10-06T00:05:36.431Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.8480139539435808,
      "accuracy_std_dev": 0.013068837655060239,
      "run_count": 9,
      "total_questions": 1737,
      "avg_cost_per_task_usd": 0.02671003742084053,
      "avg_duration_per_task_ms": 67728.42774899252,
      "avg_output_tokens_per_task": 19651.77605066206,
      "last_run_timestamp": "2026-10-03T09:55:20.843Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8478198075949368,
      "accuracy_std_dev": 0.015598435238474664,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.07492492320675105,
      "avg_duration_per_task_ms": 285938.56540084386,
      "avg_output_tokens_per_task": 19472.943459915612,
      "last_run_timestamp": "2026-10-03T10:12:41.882Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8478035581395349,
      "accuracy_std_dev": 0.061331422730297465,
      "run_count": 7,
      "total_questions": 1290,
      "avg_cost_per_task_usd": 0.010003304651162791,
      "avg_duration_per_task_ms": 534481.4193798449,
      "avg_output_tokens_per_task": 19807.151937984498,
      "last_run_timestamp": "2026-10-03T10:05:20.841Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.8477956427969672,
      "accuracy_std_dev": 0.03319167551259403,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.09577104128053918,
      "avg_duration_per_task_ms": 1415253.807919124,
      "avg_output_tokens_per_task": 42907.37742207245,
      "last_run_timestamp": "2026-08-29T18:36:10.818Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "Novita",
      "accuracy": 0.8477489772382398,
      "accuracy_std_dev": 0.010277326771101524,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.030415789074355083,
      "avg_duration_per_task_ms": 680083.3545776429,
      "avg_output_tokens_per_task": 31761.424380374305,
      "last_run_timestamp": "2026-10-03T10:03:21.013Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Parasail",
      "accuracy": 0.8476779139314369,
      "accuracy_std_dev": 0.012065198415953493,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.007001258205689278,
      "avg_duration_per_task_ms": 526965.3136396791,
      "avg_output_tokens_per_task": 17331.969365426696,
      "last_run_timestamp": "2026-10-03T17:01:29.670Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Decart",
      "accuracy": 0.8474553149480417,
      "accuracy_std_dev": 0.0229310901226483,
      "run_count": 13,
      "total_questions": 2502,
      "avg_cost_per_task_usd": 0.1667569224620304,
      "avg_duration_per_task_ms": 293740.35211830534,
      "avg_output_tokens_per_task": 41652.80575539568,
      "last_run_timestamp": "2026-10-03T17:08:51.266Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Venice",
      "accuracy": 0.8467171000000001,
      "accuracy_std_dev": 0.010298635807231945,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.09985031666666667,
      "avg_duration_per_task_ms": 586132.338888889,
      "avg_output_tokens_per_task": 31406.80808080808,
      "last_run_timestamp": "2026-10-03T18:05:42.597Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Novita",
      "accuracy": 0.8466377271407838,
      "accuracy_std_dev": 0.016590356861005438,
      "run_count": 7,
      "total_questions": 1378,
      "avg_cost_per_task_usd": 0.009304855587808419,
      "avg_duration_per_task_ms": 444689.5050798258,
      "avg_output_tokens_per_task": 28194.9412191582,
      "last_run_timestamp": "2026-10-03T10:04:48.522Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Sail Research",
      "accuracy": 0.8466045281660314,
      "accuracy_std_dev": 0.011422067985981737,
      "run_count": 12,
      "total_questions": 2361,
      "avg_cost_per_task_usd": 0.1267483756882677,
      "avg_duration_per_task_ms": 355086.75984752225,
      "avg_output_tokens_per_task": 33407.23930537908,
      "last_run_timestamp": "2026-10-03T09:40:39.339Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8464646,
      "accuracy_std_dev": 0.005966271485609766,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.10013606363636364,
      "avg_duration_per_task_ms": 581159.3505050505,
      "avg_output_tokens_per_task": 31964.29090909091,
      "last_run_timestamp": "2026-10-03T17:01:05.431Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Relace",
      "accuracy": 0.8461756369863013,
      "accuracy_std_dev": 0.003462455005724478,
      "run_count": 3,
      "total_questions": 584,
      "avg_cost_per_task_usd": 0.07262447602739726,
      "avg_duration_per_task_ms": 181056.27226027398,
      "avg_output_tokens_per_task": 31634.909246575342,
      "last_run_timestamp": "2026-10-03T09:37:27.212Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8461699999999999,
      "accuracy_std_dev": 0.03781276874945815,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.6916325820707071,
      "avg_duration_per_task_ms": 471363.1818181818,
      "avg_output_tokens_per_task": 27729.57702020202,
      "last_run_timestamp": "2026-10-03T09:31:25.951Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8458659999999999,
      "accuracy_std_dev": 0.07956358358082058,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.007138646464646464,
      "avg_duration_per_task_ms": 679204.691919192,
      "avg_output_tokens_per_task": 29375.537598204264,
      "last_run_timestamp": "2026-10-06T03:03:44.162Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Alibaba",
      "accuracy": 0.8454851888607595,
      "accuracy_std_dev": 0.015145407351405253,
      "run_count": 10,
      "total_questions": 1975,
      "avg_cost_per_task_usd": 0.08874519088607594,
      "avg_duration_per_task_ms": 411899.82784810127,
      "avg_output_tokens_per_task": 22073.469873417722,
      "last_run_timestamp": "2026-10-03T17:44:12.665Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8453983333333333,
      "accuracy_std_dev": 0.006102318839770835,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.009480049663299664,
      "avg_duration_per_task_ms": 418423.1557239057,
      "avg_output_tokens_per_task": 18452.791245791246,
      "last_run_timestamp": "2026-10-03T17:22:47.678Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "GMICloud",
      "accuracy": 0.8453581428571427,
      "accuracy_std_dev": 0.013659783740481936,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.10694367099567102,
      "avg_duration_per_task_ms": 329051.74314574315,
      "avg_output_tokens_per_task": 29197.39393939394,
      "last_run_timestamp": "2026-10-03T09:28:28.144Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Novita",
      "accuracy": 0.8449273065217391,
      "accuracy_std_dev": 0.015410014700227412,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.03735347173913043,
      "avg_duration_per_task_ms": 618214.4876811594,
      "avg_output_tokens_per_task": 34942.45652173913,
      "last_run_timestamp": "2026-10-03T17:17:28.936Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Parasail",
      "accuracy": 0.8447187666034155,
      "accuracy_std_dev": 0.033801436872417136,
      "run_count": 5,
      "total_questions": 527,
      "avg_cost_per_task_usd": 0.10369237191650856,
      "avg_duration_per_task_ms": 295443.4819734345,
      "avg_output_tokens_per_task": 23792.745730550283,
      "last_run_timestamp": "2026-08-22T17:29:49.714Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8446586363636364,
      "accuracy_std_dev": 0.08451124453669165,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.09526816483011938,
      "avg_duration_per_task_ms": 784606.3654729109,
      "avg_output_tokens_per_task": 42179.34618916437,
      "last_run_timestamp": "2026-10-05T21:07:37.081Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Google",
      "accuracy": 0.8444975263157894,
      "accuracy_std_dev": 0.05215908171451967,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.18362076209463052,
      "avg_duration_per_task_ms": 40742.116427432215,
      "avg_output_tokens_per_task": 3305.715842636895,
      "last_run_timestamp": "2026-10-03T09:56:46.597Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Together",
      "accuracy": 0.8444072953657173,
      "accuracy_std_dev": 0.011477740217046216,
      "run_count": 10,
      "total_questions": 1791,
      "avg_cost_per_task_usd": 0.015475027359017308,
      "avg_duration_per_task_ms": 603617.405360134,
      "avg_output_tokens_per_task": 16554.757118927973,
      "last_run_timestamp": "2026-08-29T18:29:51.235Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Crusoe",
      "accuracy": 0.8441076000000001,
      "accuracy_std_dev": 0.00842413538827578,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1428492090909091,
      "avg_duration_per_task_ms": 160872.74545454545,
      "avg_output_tokens_per_task": 32292.333333333332,
      "last_run_timestamp": "2026-10-03T09:56:48.670Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Azure",
      "accuracy": 0.8441076,
      "accuracy_std_dev": 0.036707935445077826,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.015406010101010103,
      "avg_duration_per_task_ms": 387191.47878787876,
      "avg_output_tokens_per_task": 27797.58787878788,
      "last_run_timestamp": "2026-10-03T16:57:48.932Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Novita",
      "accuracy": 0.8440029006622516,
      "accuracy_std_dev": 0.014256857466367755,
      "run_count": 7,
      "total_questions": 1359,
      "avg_cost_per_task_usd": 0.08230250993377485,
      "avg_duration_per_task_ms": 942574.9713024283,
      "avg_output_tokens_per_task": 41739.75423105225,
      "last_run_timestamp": "2026-10-04T00:11:46.036Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8439016603498544,
      "accuracy_std_dev": 0.05982632132724711,
      "run_count": 7,
      "total_questions": 1372,
      "avg_cost_per_task_usd": 0.07957356997084548,
      "avg_duration_per_task_ms": 563629.0634110788,
      "avg_output_tokens_per_task": 22172.81559766764,
      "last_run_timestamp": "2026-10-05T21:06:37.734Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.8438616803537586,
      "accuracy_std_dev": 0.021039203126247442,
      "run_count": 8,
      "total_questions": 1583,
      "avg_cost_per_task_usd": 0.021147979153506004,
      "avg_duration_per_task_ms": 446490.1522425774,
      "avg_output_tokens_per_task": 40034.66519267214,
      "last_run_timestamp": "2026-10-08T06:00:12.162Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8438314193103448,
      "accuracy_std_dev": 0.03520063790838207,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.05348304597701149,
      "avg_duration_per_task_ms": 217549.43770114944,
      "avg_output_tokens_per_task": 22183.33103448276,
      "last_run_timestamp": "2026-10-03T10:19:52.322Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Alibaba",
      "accuracy": 0.8435766681127982,
      "accuracy_std_dev": 0.022602212098108688,
      "run_count": 21,
      "total_questions": 4149,
      "avg_cost_per_task_usd": 0.2387051561822126,
      "avg_duration_per_task_ms": 512529.7343938298,
      "avg_output_tokens_per_task": 45605.27524704748,
      "last_run_timestamp": "2026-10-09T21:00:06.284Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Fireworks",
      "accuracy": 0.8435523956204379,
      "accuracy_std_dev": 0.013812717451238726,
      "run_count": 7,
      "total_questions": 1370,
      "avg_cost_per_task_usd": 0.13016191605839417,
      "avg_duration_per_task_ms": 317794.05401459854,
      "avg_output_tokens_per_task": 27511.544525547444,
      "last_run_timestamp": "2026-10-04T21:02:19.974Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Mistral",
      "accuracy": 0.8434342499999999,
      "accuracy_std_dev": 0.008881818827667008,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.14890805050505052,
      "avg_duration_per_task_ms": 162687.35606060605,
      "avg_output_tokens_per_task": 32857.1712962963,
      "last_run_timestamp": "2026-10-03T10:18:11.626Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "auto-routing",
      "accuracy": 0.8430214877843304,
      "accuracy_std_dev": 0.03701709147324906,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.16848753580454928,
      "avg_duration_per_task_ms": 976236.7017691659,
      "avg_output_tokens_per_task": 53637.603201347934,
      "last_run_timestamp": "2026-10-03T16:47:29.532Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "Together",
      "accuracy": 0.8425437868421053,
      "accuracy_std_dev": 0.00573439306339873,
      "run_count": 6,
      "total_questions": 1140,
      "avg_cost_per_task_usd": 0.08756267807017544,
      "avg_duration_per_task_ms": 282086.36140350875,
      "avg_output_tokens_per_task": 21480.616666666665,
      "last_run_timestamp": "2026-10-01T16:06:12.801Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8422559000000002,
      "accuracy_std_dev": 0.012844421749927075,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.13039331515151514,
      "avg_duration_per_task_ms": 426054.8303030303,
      "avg_output_tokens_per_task": 36347.83939393939,
      "last_run_timestamp": "2026-10-03T16:40:33.972Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Together",
      "accuracy": 0.8422381263537906,
      "accuracy_std_dev": 0.013606259020414556,
      "run_count": 7,
      "total_questions": 1385,
      "avg_cost_per_task_usd": 0.1362164772563177,
      "avg_duration_per_task_ms": 189944.0108303249,
      "avg_output_tokens_per_task": 30733.362454873648,
      "last_run_timestamp": "2026-10-03T09:32:18.488Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Inceptron",
      "accuracy": 0.8419940950570343,
      "accuracy_std_dev": 0.005052124553096465,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.02034613561470215,
      "avg_duration_per_task_ms": 621368.3954372624,
      "avg_output_tokens_per_task": 23198.26235741445,
      "last_run_timestamp": "2026-08-15T14:57:08.773Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Friendli",
      "accuracy": 0.8419798924812032,
      "accuracy_std_dev": 0.0161432408803948,
      "run_count": 7,
      "total_questions": 1330,
      "avg_cost_per_task_usd": 0.01795621127819549,
      "avg_duration_per_task_ms": 242875.14962406014,
      "avg_output_tokens_per_task": 35697.367669172934,
      "last_run_timestamp": "2026-10-03T09:24:57.334Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Novita",
      "accuracy": 0.841843140655106,
      "accuracy_std_dev": 0.013637590375502584,
      "run_count": 6,
      "total_questions": 1038,
      "avg_cost_per_task_usd": 0.006975833333333333,
      "avg_duration_per_task_ms": 638052.9354527938,
      "avg_output_tokens_per_task": 17230.314065510596,
      "last_run_timestamp": "2026-10-03T17:20:08.224Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "auto-routing",
      "accuracy": 0.8417508571428571,
      "accuracy_std_dev": 0.025909470769412114,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.06952829942279942,
      "avg_duration_per_task_ms": 590279.7936507936,
      "avg_output_tokens_per_task": 22093.862914862915,
      "last_run_timestamp": "2026-10-03T18:21:53.448Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Google",
      "accuracy": 0.8417268575037897,
      "accuracy_std_dev": 0.03274733184905639,
      "run_count": 30,
      "total_questions": 5937,
      "avg_cost_per_task_usd": 0.16544357267980464,
      "avg_duration_per_task_ms": 149160.101061142,
      "avg_output_tokens_per_task": 15291.950816910898,
      "last_run_timestamp": "2026-10-03T10:45:13.551Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Chutes",
      "accuracy": 0.8414701666666667,
      "accuracy_std_dev": 0.04102371415582561,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.15670842255892256,
      "avg_duration_per_task_ms": 1400165.1641414142,
      "avg_output_tokens_per_task": 51380.03956228956,
      "last_run_timestamp": "2026-10-03T17:37:27.208Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "SambaNova",
      "accuracy": 0.8414065844994617,
      "accuracy_std_dev": 0.024904832897923005,
      "run_count": 7,
      "total_questions": 929,
      "avg_cost_per_task_usd": 0.01384601076426265,
      "avg_duration_per_task_ms": 73903.2508073197,
      "avg_output_tokens_per_task": 11864.46393972013,
      "last_run_timestamp": "2026-10-03T17:24:07.121Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.8413977096774194,
      "accuracy_std_dev": 0.007842951293996439,
      "run_count": 4,
      "total_questions": 744,
      "avg_cost_per_task_usd": 0.03688066532258065,
      "avg_duration_per_task_ms": 47685.283602150535,
      "avg_output_tokens_per_task": 14664.099462365592,
      "last_run_timestamp": "2026-10-01T19:13:29.958Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.8411896666666667,
      "accuracy_std_dev": 0.008729468916008341,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.036658867003367,
      "avg_duration_per_task_ms": 51092.77104377104,
      "avg_output_tokens_per_task": 14574.808080808081,
      "last_run_timestamp": "2026-10-01T19:12:26.032Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Together",
      "accuracy": 0.8409360293968577,
      "accuracy_std_dev": 0.016369283081430274,
      "run_count": 10,
      "total_questions": 1973,
      "avg_cost_per_task_usd": 0.2137119655347187,
      "avg_duration_per_task_ms": 299955.6538266599,
      "avg_output_tokens_per_task": 48480.95995945261,
      "last_run_timestamp": "2026-10-03T16:37:10.188Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Makora",
      "accuracy": 0.840817973713034,
      "accuracy_std_dev": 0.020791979459397293,
      "run_count": 5,
      "total_questions": 913,
      "avg_cost_per_task_usd": 0.1426375947426068,
      "avg_duration_per_task_ms": 229737.46549835705,
      "avg_output_tokens_per_task": 33714.36035049288,
      "last_run_timestamp": "2026-10-03T09:33:59.008Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.84048825,
      "accuracy_std_dev": 0.004667830993887862,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.06197197222222223,
      "avg_duration_per_task_ms": 84839.70707070707,
      "avg_output_tokens_per_task": 13643.420454545454,
      "last_run_timestamp": "2026-10-02T23:24:30.140Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.8402775531250001,
      "accuracy_std_dev": 0.007466846475587956,
      "run_count": 10,
      "total_questions": 1920,
      "avg_cost_per_task_usd": 0.039408469791666664,
      "avg_duration_per_task_ms": 51791.722395833334,
      "avg_output_tokens_per_task": 14757.431770833333,
      "last_run_timestamp": "2026-10-01T19:13:57.987Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "DeepInfra",
      "accuracy": 0.8402425498181818,
      "accuracy_std_dev": 0.00976753775827742,
      "run_count": 8,
      "total_questions": 1375,
      "avg_cost_per_task_usd": 0.0796483170909091,
      "avg_duration_per_task_ms": 509723.63709090906,
      "avg_output_tokens_per_task": 19683.845818181817,
      "last_run_timestamp": "2026-10-08T22:17:54.321Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8399772366412214,
      "accuracy_std_dev": 0.010642150299634022,
      "run_count": 6,
      "total_questions": 1179,
      "avg_cost_per_task_usd": 0.02758528159457167,
      "avg_duration_per_task_ms": 524755.5767599661,
      "avg_output_tokens_per_task": 22800.932994062765,
      "last_run_timestamp": "2026-08-29T18:16:39.840Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Io Net",
      "accuracy": 0.8397870000000001,
      "accuracy_std_dev": 0.06305207943174174,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.013501392255892255,
      "avg_duration_per_task_ms": 269245.1767676768,
      "avg_output_tokens_per_task": 24293.242424242424,
      "last_run_timestamp": "2026-10-01T17:44:47.331Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "GMICloud",
      "accuracy": 0.8392069911894273,
      "accuracy_std_dev": 0.01727950565595804,
      "run_count": 7,
      "total_questions": 1362,
      "avg_cost_per_task_usd": 0.03353528046989722,
      "avg_duration_per_task_ms": 547187.9052863437,
      "avg_output_tokens_per_task": 32760.090308370043,
      "last_run_timestamp": "2026-10-03T16:57:14.717Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Relace",
      "accuracy": 0.8391689163108454,
      "accuracy_std_dev": 0.01637841911378101,
      "run_count": 6,
      "total_questions": 1171,
      "avg_cost_per_task_usd": 0.014026462852263022,
      "avg_duration_per_task_ms": 450960.66951323656,
      "avg_output_tokens_per_task": 38615.76686592656,
      "last_run_timestamp": "2026-10-03T09:28:28.466Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "auto-routing",
      "accuracy": 0.8390219978213507,
      "accuracy_std_dev": 0.019235701159615707,
      "run_count": 7,
      "total_questions": 1377,
      "avg_cost_per_task_usd": 0.047738670297748725,
      "avg_duration_per_task_ms": 642207.128540305,
      "avg_output_tokens_per_task": 37915.77995642702,
      "last_run_timestamp": "2026-10-03T16:58:17.410Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "StreamLake",
      "accuracy": 0.8389739548751007,
      "accuracy_std_dev": 0.023840636777552385,
      "run_count": 7,
      "total_questions": 1241,
      "avg_cost_per_task_usd": 0.07291945044319098,
      "avg_duration_per_task_ms": 368316.54069298954,
      "avg_output_tokens_per_task": 24012.563255439163,
      "last_run_timestamp": "2026-10-03T16:56:41.261Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8389463834196891,
      "accuracy_std_dev": 0.008774616644047783,
      "run_count": 4,
      "total_questions": 772,
      "avg_cost_per_task_usd": 0.02735646632124352,
      "avg_duration_per_task_ms": 654835.9339378239,
      "avg_output_tokens_per_task": 14525.401554404145,
      "last_run_timestamp": "2026-10-08T23:59:53.537Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Fireworks",
      "accuracy": 0.838739,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 185,
      "avg_cost_per_task_usd": 0.0289266,
      "avg_duration_per_task_ms": 182176.34594594594,
      "avg_output_tokens_per_task": 24823.56216216216,
      "last_run_timestamp": "2026-07-18T08:31:24.271Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Reka",
      "accuracy": 0.8385708888888888,
      "accuracy_std_dev": 0.008600437423558384,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.06734114758698093,
      "avg_duration_per_task_ms": 345902.8209876543,
      "avg_output_tokens_per_task": 22364.55892255892,
      "last_run_timestamp": "2026-10-04T12:03:28.619Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Novita",
      "accuracy": 0.8383837142857142,
      "accuracy_std_dev": 0.026235324298762266,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.11822585137085138,
      "avg_duration_per_task_ms": 408512.2987012987,
      "avg_output_tokens_per_task": 35583.771284271286,
      "last_run_timestamp": "2026-10-03T10:08:41.581Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.8381875026968716,
      "accuracy_std_dev": 0.003092760668974499,
      "run_count": 6,
      "total_questions": 927,
      "avg_cost_per_task_usd": 0.13592745091693637,
      "avg_duration_per_task_ms": 121247.60086299892,
      "avg_output_tokens_per_task": 30661.07551240561,
      "last_run_timestamp": "2026-10-03T22:03:15.220Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.837813550776583,
      "accuracy_std_dev": 0.014764985758205111,
      "run_count": 10,
      "total_questions": 1674,
      "avg_cost_per_task_usd": 0.09487115710872161,
      "avg_duration_per_task_ms": 634750.7311827957,
      "avg_output_tokens_per_task": 29572.94982078853,
      "last_run_timestamp": "2026-10-08T23:03:31.199Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Phala",
      "accuracy": 0.837571,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 118,
      "avg_cost_per_task_usd": 0.01632071186440678,
      "avg_duration_per_task_ms": 269506.5254237288,
      "avg_output_tokens_per_task": 11788.906779661016,
      "last_run_timestamp": "2026-07-18T16:47:35.300Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8375589327830189,
      "accuracy_std_dev": 0.020056898198209345,
      "run_count": 9,
      "total_questions": 1696,
      "avg_cost_per_task_usd": 0.054286748820754724,
      "avg_duration_per_task_ms": 282159.9534198113,
      "avg_output_tokens_per_task": 22660.685141509435,
      "last_run_timestamp": "2026-09-26T09:50:39.842Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Novita",
      "accuracy": 0.8375548434650456,
      "accuracy_std_dev": 0.05830789076754535,
      "run_count": 10,
      "total_questions": 1974,
      "avg_cost_per_task_usd": 0.0750043009118541,
      "avg_duration_per_task_ms": 671651.9898682877,
      "avg_output_tokens_per_task": 45629.63880445795,
      "last_run_timestamp": "2026-10-03T17:06:14.822Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Z.AI",
      "accuracy": 0.8374960027752081,
      "accuracy_std_dev": 0.029394078011376546,
      "run_count": 6,
      "total_questions": 1081,
      "avg_cost_per_task_usd": 0.0877097058279371,
      "avg_duration_per_task_ms": 1673535.2590194265,
      "avg_output_tokens_per_task": 40889.8316373728,
      "last_run_timestamp": "2026-08-29T19:12:33.941Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "StreamLake",
      "accuracy": 0.8371595454545454,
      "accuracy_std_dev": 0.01566308613131605,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.07092007805325987,
      "avg_duration_per_task_ms": 680254.4848484849,
      "avg_output_tokens_per_task": 53261.87281910009,
      "last_run_timestamp": "2026-10-03T16:57:33.328Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Fireworks",
      "accuracy": 0.8367169999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 197,
      "avg_cost_per_task_usd": 0.10272846700507614,
      "avg_duration_per_task_ms": 406781.3959390863,
      "avg_output_tokens_per_task": 31764.908629441623,
      "last_run_timestamp": "2026-07-18T08:30:34.727Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "StreamLake",
      "accuracy": 0.8367001428571429,
      "accuracy_std_dev": 0.02855562409668226,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.14463846464646468,
      "avg_duration_per_task_ms": 904627.3023088023,
      "avg_output_tokens_per_task": 54000.30014430014,
      "last_run_timestamp": "2026-10-03T17:37:18.132Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Morph",
      "accuracy": 0.8366436799091941,
      "accuracy_std_dev": 0.015820054069611535,
      "run_count": 9,
      "total_questions": 1762,
      "avg_cost_per_task_usd": 0.08522557775255392,
      "avg_duration_per_task_ms": 546245.1770715096,
      "avg_output_tokens_per_task": 31173.529511918274,
      "last_run_timestamp": "2026-10-03T11:03:00.010Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Friendli",
      "accuracy": 0.8361980881913303,
      "accuracy_std_dev": 0.014593793273938368,
      "run_count": 7,
      "total_questions": 1338,
      "avg_cost_per_task_usd": 0.21788730717488788,
      "avg_duration_per_task_ms": 418506.8639760837,
      "avg_output_tokens_per_task": 49387.9245142003,
      "last_run_timestamp": "2026-10-03T16:59:26.755Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Z.AI",
      "accuracy": 0.8361952000000001,
      "accuracy_std_dev": 0.0746685963545586,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.19211891767676767,
      "avg_duration_per_task_ms": 535463.2151515152,
      "avg_output_tokens_per_task": 43715.205555555556,
      "last_run_timestamp": "2026-10-03T18:53:59.567Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Nebius",
      "accuracy": 0.836139,
      "accuracy_std_dev": 0.05678065343641852,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.005011018518518518,
      "avg_duration_per_task_ms": 214818.47643097644,
      "avg_output_tokens_per_task": 17529.722222222223,
      "last_run_timestamp": "2026-09-26T11:06:41.988Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Reka",
      "accuracy": 0.836139,
      "accuracy_std_dev": 0.0007933738084913017,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.05002690909090909,
      "avg_duration_per_task_ms": 537584.335016835,
      "avg_output_tokens_per_task": 49932.06734006734,
      "last_run_timestamp": "2026-09-30T11:44:08.104Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.8360065448008849,
      "accuracy_std_dev": 0.008779454087962683,
      "run_count": 11,
      "total_questions": 1808,
      "avg_cost_per_task_usd": 0.018033527101769913,
      "avg_duration_per_task_ms": 187582.5735619469,
      "avg_output_tokens_per_task": 35877.24225663717,
      "last_run_timestamp": "2026-10-03T09:31:41.635Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "Novita",
      "accuracy": 0.8358585,
      "accuracy_std_dev": 0.01430950000000003,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0011451691919191918,
      "avg_duration_per_task_ms": 116299.33333333333,
      "avg_output_tokens_per_task": 18355.219696969696,
      "last_run_timestamp": "2026-10-01T08:23:48.135Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Venice",
      "accuracy": 0.8357181666666667,
      "accuracy_std_dev": 0.013680371820198815,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.006888778619528619,
      "avg_duration_per_task_ms": 375845.8905723906,
      "avg_output_tokens_per_task": 26123.255892255893,
      "last_run_timestamp": "2026-10-03T17:23:05.193Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.835648125,
      "accuracy_std_dev": 0.012901937649414363,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.03973862878787879,
      "avg_duration_per_task_ms": 223092.5542929293,
      "avg_output_tokens_per_task": 14589.925505050505,
      "last_run_timestamp": "2026-10-03T09:51:44.286Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "auto-routing",
      "accuracy": 0.8354408285280729,
      "accuracy_std_dev": 0.043474653742611905,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.1658476181082448,
      "avg_duration_per_task_ms": 161596.665149216,
      "avg_output_tokens_per_task": 16364.909458775923,
      "last_run_timestamp": "2026-10-03T09:33:36.803Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Parasail",
      "accuracy": 0.8353909418724279,
      "accuracy_std_dev": 0.02098218819975649,
      "run_count": 10,
      "total_questions": 1944,
      "avg_cost_per_task_usd": 0.21285450617283952,
      "avg_duration_per_task_ms": 291168.728909465,
      "avg_output_tokens_per_task": 48204.40740740741,
      "last_run_timestamp": "2026-10-03T16:57:30.005Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "DeepInfra",
      "accuracy": 0.8351849999999998,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 180,
      "avg_cost_per_task_usd": 0.0030483999999999997,
      "avg_duration_per_task_ms": 288802.22222222225,
      "avg_output_tokens_per_task": 16703.5,
      "last_run_timestamp": "2026-10-01T08:43:11.415Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Novita",
      "accuracy": 0.8350166666666667,
      "accuracy_std_dev": 0.019584493003961645,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.025601432659932657,
      "avg_duration_per_task_ms": 248785.20707070708,
      "avg_output_tokens_per_task": 21309.069865319867,
      "last_run_timestamp": "2026-08-29T18:12:43.943Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.8350166666666667,
      "accuracy_std_dev": 0.005991526757754575,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.06194655555555556,
      "avg_duration_per_task_ms": 84396.87373737374,
      "avg_output_tokens_per_task": 13638.5,
      "last_run_timestamp": "2026-10-01T19:12:44.401Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Modal",
      "accuracy": 0.8348964285714285,
      "accuracy_std_dev": 0.013870860266422272,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.13617697113997115,
      "avg_duration_per_task_ms": 172785.43217893218,
      "avg_output_tokens_per_task": 30709.33621933622,
      "last_run_timestamp": "2026-10-03T09:19:24.458Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DekaLLM",
      "accuracy": 0.834704304347826,
      "accuracy_std_dev": 0.01114896903933274,
      "run_count": 5,
      "total_questions": 851,
      "avg_cost_per_task_usd": 0.04789096122209165,
      "avg_duration_per_task_ms": 91498.8401880141,
      "avg_output_tokens_per_task": 15123.290246768507,
      "last_run_timestamp": "2026-10-09T02:08:01.186Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Mistral",
      "accuracy": 0.8345736842105264,
      "accuracy_std_dev": 0.012051692648327507,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.2574891374269006,
      "avg_duration_per_task_ms": 275645.54146730463,
      "avg_output_tokens_per_task": 56198.80940988836,
      "last_run_timestamp": "2026-10-03T17:33:23.935Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.8344556666666666,
      "accuracy_std_dev": 0.04246379448685927,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1074556616161616,
      "avg_duration_per_task_ms": 1223271.058922559,
      "avg_output_tokens_per_task": 43332.00420875421,
      "last_run_timestamp": "2026-08-29T18:42:20.541Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Phala",
      "accuracy": 0.8341984152249134,
      "accuracy_std_dev": 0.029002188381032377,
      "run_count": 3,
      "total_questions": 578,
      "avg_cost_per_task_usd": 0.005492470588235294,
      "avg_duration_per_task_ms": 956260.6712802768,
      "avg_output_tokens_per_task": 15404.778546712803,
      "last_run_timestamp": "2026-08-29T18:08:44.008Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "WandB",
      "accuracy": 0.834197,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 193,
      "avg_cost_per_task_usd": 0.02599202590673575,
      "avg_duration_per_task_ms": 128468.99481865285,
      "avg_output_tokens_per_task": 20660.492227979274,
      "last_run_timestamp": "2026-07-18T16:00:08.986Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Z.AI",
      "accuracy": 0.8341842341356673,
      "accuracy_std_dev": 0.02342272596676753,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.08859786870897156,
      "avg_duration_per_task_ms": 985152.2742523706,
      "avg_output_tokens_per_task": 40722.371991247266,
      "last_run_timestamp": "2026-10-04T00:10:49.236Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8341792639593909,
      "accuracy_std_dev": 0.020282949394476377,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.07842123350253806,
      "avg_duration_per_task_ms": 728101.9196277496,
      "avg_output_tokens_per_task": 42337.8130287648,
      "last_run_timestamp": "2026-10-04T00:07:08.217Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Crusoe",
      "accuracy": 0.8340809047085203,
      "accuracy_std_dev": 0.01732897775403963,
      "run_count": 5,
      "total_questions": 892,
      "avg_cost_per_task_usd": 0.020895510089686098,
      "avg_duration_per_task_ms": 283776.2948430493,
      "avg_output_tokens_per_task": 41571.40470852018,
      "last_run_timestamp": "2026-10-03T09:31:15.820Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "StreamLake",
      "accuracy": 0.8340712409867174,
      "accuracy_std_dev": 0.026087244621414103,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.15789761669829225,
      "avg_duration_per_task_ms": 841435.927261227,
      "avg_output_tokens_per_task": 51847.42757748261,
      "last_run_timestamp": "2026-10-03T17:11:42.929Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "auto-routing",
      "accuracy": 0.8340364430379746,
      "accuracy_std_dev": 0.01069546220626094,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.0954248565400844,
      "avg_duration_per_task_ms": 293770.05400843883,
      "avg_output_tokens_per_task": 23403.932489451476,
      "last_run_timestamp": "2026-10-01T16:11:05.832Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8338943333333333,
      "accuracy_std_dev": 0.011690806112307073,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03773407239057239,
      "avg_duration_per_task_ms": 276396.1228956229,
      "avg_output_tokens_per_task": 28290.345117845118,
      "last_run_timestamp": "2026-09-12T09:56:44.384Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Cloudflare",
      "accuracy": 0.8338943333333333,
      "accuracy_std_dev": 0.05954210180106922,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0609229090909091,
      "avg_duration_per_task_ms": 629153.6734006734,
      "avg_output_tokens_per_task": 22578.223905723906,
      "last_run_timestamp": "2026-10-03T16:58:19.180Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "CoreWeave",
      "accuracy": 0.8337931059602649,
      "accuracy_std_dev": 0.013355036895202588,
      "run_count": 12,
      "total_questions": 1812,
      "avg_cost_per_task_usd": 0.03549873675496689,
      "avg_duration_per_task_ms": 241451.8725165563,
      "avg_output_tokens_per_task": 28296.954194260485,
      "last_run_timestamp": "2026-10-03T16:12:07.493Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Cloudflare",
      "accuracy": 0.8334570734966593,
      "accuracy_std_dev": 0.0419490749322732,
      "run_count": 7,
      "total_questions": 1347,
      "avg_cost_per_task_usd": 0.02924405196733482,
      "avg_duration_per_task_ms": 281996.3355605048,
      "avg_output_tokens_per_task": 9032.244988864142,
      "last_run_timestamp": "2026-10-03T10:36:42.102Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Parasail",
      "accuracy": 0.8333335829015545,
      "accuracy_std_dev": 0.019006499999999982,
      "run_count": 2,
      "total_questions": 386,
      "avg_cost_per_task_usd": 0.023828624352331605,
      "avg_duration_per_task_ms": 290054.69430051814,
      "avg_output_tokens_per_task": 19802.816062176167,
      "last_run_timestamp": "2026-08-01T18:21:26.176Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-flash-20260826",
      "model_slug": "qwen/qwen3.8-flash",
      "model_name": "Qwen: Qwen3.8 Flash",
      "provider_name": "Makora",
      "accuracy": 0.8333329999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 156,
      "avg_cost_per_task_usd": 0.003982064102564103,
      "avg_duration_per_task_ms": 34178.44871794872,
      "avg_output_tokens_per_task": 8341.692307692309,
      "last_run_timestamp": "2026-09-09T15:03:02.824Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Cloudflare",
      "accuracy": 0.8330725023474179,
      "accuracy_std_dev": 0.010168177287347969,
      "run_count": 7,
      "total_questions": 1278,
      "avg_cost_per_task_usd": 0.029710525039123633,
      "avg_duration_per_task_ms": 655555.5469483568,
      "avg_output_tokens_per_task": 42323.18544600939,
      "last_run_timestamp": "2026-10-03T09:40:53.626Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Parasail",
      "accuracy": 0.8329019637618638,
      "accuracy_std_dev": 0.014650670481433776,
      "run_count": 6,
      "total_questions": 1159,
      "avg_cost_per_task_usd": 0.1486290198446937,
      "avg_duration_per_task_ms": 266639.7955133736,
      "avg_output_tokens_per_task": 33590.909404659185,
      "last_run_timestamp": "2026-10-03T10:00:12.438Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Cloudflare",
      "accuracy": 0.832620296969697,
      "accuracy_std_dev": 0.019550200858985223,
      "run_count": 15,
      "total_questions": 2805,
      "avg_cost_per_task_usd": 0.1869388680926916,
      "avg_duration_per_task_ms": 373798.6777183601,
      "avg_output_tokens_per_task": 38438.90623885918,
      "last_run_timestamp": "2026-10-03T16:59:44.345Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Phala",
      "accuracy": 0.8325081287128712,
      "accuracy_std_dev": 0.06854471399887452,
      "run_count": 4,
      "total_questions": 606,
      "avg_cost_per_task_usd": 0.13443122607260727,
      "avg_duration_per_task_ms": 1513362.3762376239,
      "avg_output_tokens_per_task": 46533.88613861386,
      "last_run_timestamp": "2026-08-29T18:25:52.753Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "InferenceNet",
      "accuracy": 0.8324915000000002,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13435382575757576,
      "avg_duration_per_task_ms": 466692.7878787879,
      "avg_output_tokens_per_task": 61036.83838383838,
      "last_run_timestamp": "2026-10-03T17:03:50.322Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "NextBit",
      "accuracy": 0.8324914999999999,
      "accuracy_std_dev": 0.01902788623310536,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.02279124747474747,
      "avg_duration_per_task_ms": 532516.4444444445,
      "avg_output_tokens_per_task": 41107.4779040404,
      "last_run_timestamp": "2026-10-03T10:07:54.348Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Decart",
      "accuracy": 0.8324914999999999,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.023044303030303033,
      "avg_duration_per_task_ms": 640134.2070707071,
      "avg_output_tokens_per_task": 54011.681818181816,
      "last_run_timestamp": "2026-09-26T08:54:28.265Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Alibaba",
      "accuracy": 0.8323712857142858,
      "accuracy_std_dev": 0.006903653726506745,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.24708544516594516,
      "avg_duration_per_task_ms": 346092.02308802307,
      "avg_output_tokens_per_task": 40400.90404040404,
      "last_run_timestamp": "2026-10-03T10:13:08.858Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8323231,
      "accuracy_std_dev": 0.008045717599443841,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.08128380757575758,
      "avg_duration_per_task_ms": 799166.1005050505,
      "avg_output_tokens_per_task": 48099.10101010101,
      "last_run_timestamp": "2026-10-03T16:25:22.307Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AkashML",
      "accuracy": 0.8321069434069044,
      "accuracy_std_dev": 0.013054586209736332,
      "run_count": 11,
      "total_questions": 1767,
      "avg_cost_per_task_usd": 0.11444472948500282,
      "avg_duration_per_task_ms": 230736.259762309,
      "avg_output_tokens_per_task": 27650.263157894737,
      "last_run_timestamp": "2026-10-09T15:49:59.760Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Sail Research",
      "accuracy": 0.8314979893324156,
      "accuracy_std_dev": 0.10730469177701411,
      "run_count": 15,
      "total_questions": 2906,
      "avg_cost_per_task_usd": 0.007892868203716447,
      "avg_duration_per_task_ms": 643389.9057123193,
      "avg_output_tokens_per_task": 27155.329318651067,
      "last_run_timestamp": "2026-10-03T09:52:45.061Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Io Net",
      "accuracy": 0.8312629347826086,
      "accuracy_std_dev": 0.044536486139877485,
      "run_count": 4,
      "total_questions": 644,
      "avg_cost_per_task_usd": 0.005694647515527951,
      "avg_duration_per_task_ms": 270388.4906832298,
      "avg_output_tokens_per_task": 13521.194099378881,
      "last_run_timestamp": "2026-08-22T09:56:05.550Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "Azure",
      "accuracy": 0.8310885000000001,
      "accuracy_std_dev": 0.012955632388399007,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.06418621717171719,
      "avg_duration_per_task_ms": 117116.24831649831,
      "avg_output_tokens_per_task": 13483.090067340067,
      "last_run_timestamp": "2026-10-01T19:14:05.661Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "ModelRun",
      "accuracy": 0.8309870236486486,
      "accuracy_std_dev": 0.03224080320546207,
      "run_count": 9,
      "total_questions": 1776,
      "avg_cost_per_task_usd": 0.015772429054054054,
      "avg_duration_per_task_ms": 136867.50112612612,
      "avg_output_tokens_per_task": 17821.467905405407,
      "last_run_timestamp": "2026-10-05T18:01:16.100Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Nebius",
      "accuracy": 0.8308079999999999,
      "accuracy_std_dev": 0.019360000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.16246437626262625,
      "avg_duration_per_task_ms": 214673.86111111112,
      "avg_output_tokens_per_task": 36669.38383838384,
      "last_run_timestamp": "2026-10-03T09:42:09.121Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Parasail",
      "accuracy": 0.8307627712082263,
      "accuracy_std_dev": 0.005791790283452965,
      "run_count": 8,
      "total_questions": 1556,
      "avg_cost_per_task_usd": 0.03552349742930591,
      "avg_duration_per_task_ms": 197284.01863753214,
      "avg_output_tokens_per_task": 13956.56233933162,
      "last_run_timestamp": "2026-10-03T09:16:27.631Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "StreamLake",
      "accuracy": 0.8303493580562661,
      "accuracy_std_dev": 0.021497934689561845,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.0706463444160273,
      "avg_duration_per_task_ms": 871151.2676896845,
      "avg_output_tokens_per_task": 40306.14833759591,
      "last_run_timestamp": "2026-08-30T00:37:59.094Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "auto-routing",
      "accuracy": 0.8302722727272728,
      "accuracy_std_dev": 0.03337639966446328,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.05463231588613406,
      "avg_duration_per_task_ms": 268557.7502295684,
      "avg_output_tokens_per_task": 23385.22819100092,
      "last_run_timestamp": "2026-10-03T09:29:36.121Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Cerebras",
      "accuracy": 0.829966,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.022576904040404043,
      "avg_duration_per_task_ms": 10045.939393939394,
      "avg_output_tokens_per_task": 14604.111111111111,
      "last_run_timestamp": "2026-10-03T08:20:05.754Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8298668364249578,
      "accuracy_std_dev": 0.03414970870688992,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.10817975548060708,
      "avg_duration_per_task_ms": 980232.0573355817,
      "avg_output_tokens_per_task": 33808.60314783586,
      "last_run_timestamp": "2026-09-26T10:24:21.190Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Friendli",
      "accuracy": 0.8296294999999999,
      "accuracy_std_dev": 0.023614739190810495,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.22653554494949493,
      "avg_duration_per_task_ms": 333936.53939393937,
      "avg_output_tokens_per_task": 51370.22727272727,
      "last_run_timestamp": "2026-10-03T17:09:33.799Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Morph",
      "accuracy": 0.8292847495662232,
      "accuracy_std_dev": 0.024481615398972544,
      "run_count": 9,
      "total_questions": 1729,
      "avg_cost_per_task_usd": 0.021951503759398496,
      "avg_duration_per_task_ms": 881311.056680162,
      "avg_output_tokens_per_task": 45066.87854251012,
      "last_run_timestamp": "2026-10-03T09:15:02.962Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Phala",
      "accuracy": 0.8289733974683544,
      "accuracy_std_dev": 0.06161366912305707,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.22182355358649788,
      "avg_duration_per_task_ms": 1441908.6481012658,
      "avg_output_tokens_per_task": 52295.71561181435,
      "last_run_timestamp": "2026-10-03T17:35:18.039Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Modal",
      "accuracy": 0.8288054812419146,
      "accuracy_std_dev": 0.014299354744511224,
      "run_count": 8,
      "total_questions": 1546,
      "avg_cost_per_task_usd": 0.03668785575679172,
      "avg_duration_per_task_ms": 429133.35187580856,
      "avg_output_tokens_per_task": 46204.894566623545,
      "last_run_timestamp": "2026-10-03T09:52:10.218Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.828497570984456,
      "accuracy_std_dev": 0.015230029187102692,
      "run_count": 5,
      "total_questions": 965,
      "avg_cost_per_task_usd": 0.11257541554404144,
      "avg_duration_per_task_ms": 438124.0186528497,
      "avg_output_tokens_per_task": 31261.672538860104,
      "last_run_timestamp": "2026-08-29T10:53:09.444Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Baidu",
      "accuracy": 0.8283596285714285,
      "accuracy_std_dev": 0.008957848077942624,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.007575441269841269,
      "avg_duration_per_task_ms": 331346.21904761903,
      "avg_output_tokens_per_task": 19200.022857142856,
      "last_run_timestamp": "2026-10-04T00:00:36.971Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.828283,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.009031212121212122,
      "avg_duration_per_task_ms": 399484.5707070707,
      "avg_output_tokens_per_task": 21544.277777777777,
      "last_run_timestamp": "2026-08-01T20:03:34.768Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8282826999999999,
      "accuracy_std_dev": 0.026967609579085804,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.13044083383838387,
      "avg_duration_per_task_ms": 536030.553030303,
      "avg_output_tokens_per_task": 39996.7904040404,
      "last_run_timestamp": "2026-10-03T16:58:32.612Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "CoreWeave",
      "accuracy": 0.8279570000000002,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 186,
      "avg_cost_per_task_usd": 0.024709930107526885,
      "avg_duration_per_task_ms": 419969.8602150538,
      "avg_output_tokens_per_task": 20901.392473118278,
      "last_run_timestamp": "2026-08-01T18:03:43.857Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.827862,
      "accuracy_std_dev": 0.0497186507736282,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.16292082386363638,
      "avg_duration_per_task_ms": 160402.9981060606,
      "avg_output_tokens_per_task": 16138.333333333334,
      "last_run_timestamp": "2026-10-03T09:35:14.442Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.8275666632911391,
      "accuracy_std_dev": 0.031862730949862604,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.025537988185654007,
      "avg_duration_per_task_ms": 492003.4506329114,
      "avg_output_tokens_per_task": 23813.98734177215,
      "last_run_timestamp": "2026-08-29T18:05:36.579Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Nebius",
      "accuracy": 0.82744125,
      "accuracy_std_dev": 0.11256498383905848,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.10842245328282829,
      "avg_duration_per_task_ms": 1501711.7184343433,
      "avg_output_tokens_per_task": 26982.195707070707,
      "last_run_timestamp": "2026-10-03T16:52:29.704Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Wafer",
      "accuracy": 0.8274407499999998,
      "accuracy_std_dev": 0.002791768917281665,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.06294326136363637,
      "avg_duration_per_task_ms": 238417.01893939395,
      "avg_output_tokens_per_task": 15108.617424242424,
      "last_run_timestamp": "2026-10-03T10:04:25.365Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Chutes",
      "accuracy": 0.8274200658895084,
      "accuracy_std_dev": 0.03473866469065846,
      "run_count": 10,
      "total_questions": 1973,
      "avg_cost_per_task_usd": 0.06533462037506337,
      "avg_duration_per_task_ms": 1047959.904713634,
      "avg_output_tokens_per_task": 32930.106943740495,
      "last_run_timestamp": "2026-10-03T17:39:05.807Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "BaseTen",
      "accuracy": 0.8269358,
      "accuracy_std_dev": 0.006767606146932625,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.04578402121212121,
      "avg_duration_per_task_ms": 60291.652525252524,
      "avg_output_tokens_per_task": 11285.754545454545,
      "last_run_timestamp": "2026-09-26T09:19:03.901Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Venice",
      "accuracy": 0.826839857142857,
      "accuracy_std_dev": 0.052812585974039414,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.09028968326118327,
      "avg_duration_per_task_ms": 468556.38167388167,
      "avg_output_tokens_per_task": 20487.358585858587,
      "last_run_timestamp": "2026-10-03T17:40:21.237Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Phala",
      "accuracy": 0.8267676,
      "accuracy_std_dev": 0.014373097433747551,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.06195891818181818,
      "avg_duration_per_task_ms": 501765.85909090907,
      "avg_output_tokens_per_task": 48813.566666666666,
      "last_run_timestamp": "2026-10-03T17:23:58.689Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8265992857142858,
      "accuracy_std_dev": 0.05408198806498543,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.09954717027417027,
      "avg_duration_per_task_ms": 788800.0598845598,
      "avg_output_tokens_per_task": 39778.419913419915,
      "last_run_timestamp": "2026-10-03T17:58:22.995Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "StreamLake",
      "accuracy": 0.8265991666666667,
      "accuracy_std_dev": 0.013432985736817995,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.024518459595959594,
      "avg_duration_per_task_ms": 477551.41245791246,
      "avg_output_tokens_per_task": 22795.994107744107,
      "last_run_timestamp": "2026-08-29T18:09:03.379Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "WandB",
      "accuracy": 0.8265989999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02378419696969697,
      "avg_duration_per_task_ms": 421976.08585858584,
      "avg_output_tokens_per_task": 20387.959595959597,
      "last_run_timestamp": "2026-07-18T16:18:27.151Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Crusoe",
      "accuracy": 0.8264543613323679,
      "accuracy_std_dev": 0.018188360323715465,
      "run_count": 7,
      "total_questions": 1381,
      "avg_cost_per_task_usd": 0.20616153874004342,
      "avg_duration_per_task_ms": 315117.1817523534,
      "avg_output_tokens_per_task": 46875.9225199131,
      "last_run_timestamp": "2026-09-12T08:53:53.147Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Friendli",
      "accuracy": 0.8263587142857144,
      "accuracy_std_dev": 0.007278841366086593,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.1363799206349206,
      "avg_duration_per_task_ms": 152940.04473304472,
      "avg_output_tokens_per_task": 33517.63347763348,
      "last_run_timestamp": "2026-10-03T09:28:23.102Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Xiaomi",
      "accuracy": 0.8262626,
      "accuracy_std_dev": 0.023626710774883566,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.01075091717171717,
      "avg_duration_per_task_ms": 644954.0035353536,
      "avg_output_tokens_per_task": 38156.42727272727,
      "last_run_timestamp": "2026-10-03T09:23:44.336Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Novita",
      "accuracy": 0.8261182857142857,
      "accuracy_std_dev": 0.010787762507506711,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.043446235209235205,
      "avg_duration_per_task_ms": 226902.49855699856,
      "avg_output_tokens_per_task": 14365.386002886004,
      "last_run_timestamp": "2026-10-03T09:30:12.452Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Crusoe",
      "accuracy": 0.8261094014447884,
      "accuracy_std_dev": 0.026551028658038853,
      "run_count": 5,
      "total_questions": 969,
      "avg_cost_per_task_usd": 0.2076026150670795,
      "avg_duration_per_task_ms": 857894.439628483,
      "avg_output_tokens_per_task": 47560.28689370485,
      "last_run_timestamp": "2026-08-29T18:53:23.077Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AkashML",
      "accuracy": 0.8259840411061286,
      "accuracy_std_dev": 0.022038029406052623,
      "run_count": 14,
      "total_questions": 2676,
      "avg_cost_per_task_usd": 0.04315680829596412,
      "avg_duration_per_task_ms": 371792.15134529147,
      "avg_output_tokens_per_task": 46418.25635276532,
      "last_run_timestamp": "2026-10-09T16:15:58.733Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "CoreWeave",
      "accuracy": 0.8258873707224335,
      "accuracy_std_dev": 0.009676826643527784,
      "run_count": 6,
      "total_questions": 1052,
      "avg_cost_per_task_usd": 0.025350776615969583,
      "avg_duration_per_task_ms": 120122.08840304182,
      "avg_output_tokens_per_task": 20151.38783269962,
      "last_run_timestamp": "2026-08-29T18:29:13.946Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.8254038475770924,
      "accuracy_std_dev": 0.017871570248271126,
      "run_count": 6,
      "total_questions": 1135,
      "avg_cost_per_task_usd": 0.024938470484581497,
      "avg_duration_per_task_ms": 321173.1885462555,
      "avg_output_tokens_per_task": 49662.51806167401,
      "last_run_timestamp": "2026-10-03T09:30:56.352Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.8253399469387755,
      "accuracy_std_dev": 0.06202361079266511,
      "run_count": 5,
      "total_questions": 980,
      "avg_cost_per_task_usd": 0.048117852040816325,
      "avg_duration_per_task_ms": 1077180.0714285714,
      "avg_output_tokens_per_task": 26155.976530612246,
      "last_run_timestamp": "2026-08-29T19:26:41.546Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.824916,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.022821656565656566,
      "avg_duration_per_task_ms": 516406.73232323234,
      "avg_output_tokens_per_task": 45403.40909090909,
      "last_run_timestamp": "2026-10-04T03:04:38.670Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "CoreWeave",
      "accuracy": 0.8249157142857142,
      "accuracy_std_dev": 0.019837994431424387,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.1225871738816739,
      "avg_duration_per_task_ms": 339042.2626262626,
      "avg_output_tokens_per_task": 50612.39971139971,
      "last_run_timestamp": "2026-09-26T08:56:23.149Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.824777114967462,
      "accuracy_std_dev": 0.009549509751280208,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.19052757122198116,
      "avg_duration_per_task_ms": 927003.5198843095,
      "avg_output_tokens_per_task": 50705.37816341287,
      "last_run_timestamp": "2026-10-03T16:39:56.568Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Ionstream",
      "accuracy": 0.8243546666666666,
      "accuracy_std_dev": 0.012396478769751095,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.038657538720538724,
      "avg_duration_per_task_ms": 268369.8787878788,
      "avg_output_tokens_per_task": 15453.594276094276,
      "last_run_timestamp": "2026-10-03T09:38:33.682Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "BaseTen",
      "accuracy": 0.8243470488301118,
      "accuracy_std_dev": 0.01897362841419636,
      "run_count": 5,
      "total_questions": 983,
      "avg_cost_per_task_usd": 0.09452765412004069,
      "avg_duration_per_task_ms": 201477.20142421158,
      "avg_output_tokens_per_task": 27045.996948118005,
      "last_run_timestamp": "2026-08-29T18:58:14.477Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Chutes",
      "accuracy": 0.8238636249999999,
      "accuracy_std_dev": 0.022648157226016746,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.03663502967171717,
      "avg_duration_per_task_ms": 270780.9375,
      "avg_output_tokens_per_task": 14336.900252525253,
      "last_run_timestamp": "2026-10-03T09:40:25.832Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Mancer 2",
      "accuracy": 0.823569,
      "accuracy_std_dev": 0.003264448253533828,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.039843759595959594,
      "avg_duration_per_task_ms": 211450.69292929294,
      "avg_output_tokens_per_task": 14991.045454545454,
      "last_run_timestamp": "2026-10-03T09:53:48.526Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8234848000000001,
      "accuracy_std_dev": 0.06057572578979139,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.05154826161616162,
      "avg_duration_per_task_ms": 588991.1358585858,
      "avg_output_tokens_per_task": 24679.851515151517,
      "last_run_timestamp": "2026-10-03T08:20:09.017Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8233863549206349,
      "accuracy_std_dev": 0.011000693398599945,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.025066622222222223,
      "avg_duration_per_task_ms": 470143.7276190476,
      "avg_output_tokens_per_task": 25031.77904761905,
      "last_run_timestamp": "2026-10-08T23:31:01.240Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Crusoe",
      "accuracy": 0.8233255173210161,
      "accuracy_std_dev": 0.0467612856991764,
      "run_count": 5,
      "total_questions": 866,
      "avg_cost_per_task_usd": 0.16841330484988454,
      "avg_duration_per_task_ms": 1763547.682448037,
      "avg_output_tokens_per_task": 49858.13856812933,
      "last_run_timestamp": "2026-10-03T17:19:48.286Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8232324285714286,
      "accuracy_std_dev": 0.007793090691799505,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.048816260461760456,
      "avg_duration_per_task_ms": 449944.26046176045,
      "avg_output_tokens_per_task": 27045.45238095238,
      "last_run_timestamp": "2026-10-03T16:27:31.571Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "GLM 5.3",
      "provider_name": "digitalocean",
      "accuracy": 0.823232,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.025912994949494946,
      "avg_duration_per_task_ms": 162312.898989899,
      "avg_output_tokens_per_task": 9036.257575757576,
      "last_run_timestamp": "2026-10-01T23:50:45.024Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Cloudflare",
      "accuracy": 0.8231693085966202,
      "accuracy_std_dev": 0.01720174744208347,
      "run_count": 7,
      "total_questions": 1361,
      "avg_cost_per_task_usd": 0.13977533284349744,
      "avg_duration_per_task_ms": 264582.22189566493,
      "avg_output_tokens_per_task": 31542.92799412197,
      "last_run_timestamp": "2026-10-03T09:52:37.682Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Together",
      "accuracy": 0.823129,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.032144397959183675,
      "avg_duration_per_task_ms": 345053.9591836735,
      "avg_output_tokens_per_task": 27884.80612244898,
      "last_run_timestamp": "2026-07-18T08:29:12.575Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Google",
      "accuracy": 0.8231118571428572,
      "accuracy_std_dev": 0.010745952705601317,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.09818569913419915,
      "avg_duration_per_task_ms": 370933.02958152955,
      "avg_output_tokens_per_task": 45322.332611832615,
      "last_run_timestamp": "2026-10-04T00:05:42.411Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.8227512857142858,
      "accuracy_std_dev": 0.013638696342332996,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.03510758802308803,
      "avg_duration_per_task_ms": 250326.75108225108,
      "avg_output_tokens_per_task": 13920.564213564214,
      "last_run_timestamp": "2026-10-03T09:31:25.768Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Venice",
      "accuracy": 0.8223905999999999,
      "accuracy_std_dev": 0.023475537008554266,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.03811036666666667,
      "avg_duration_per_task_ms": 320984.0297979798,
      "avg_output_tokens_per_task": 39454.10808080808,
      "last_run_timestamp": "2026-10-03T16:50:17.392Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8223905000000001,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.007053464646464647,
      "avg_duration_per_task_ms": 622803.648989899,
      "avg_output_tokens_per_task": 16887.560606060608,
      "last_run_timestamp": "2026-08-01T18:25:37.046Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-pro",
      "model_slug": "google/gemini-2.5-pro",
      "model_name": "Google: Gemini 2.5 Pro",
      "provider_name": "Google AI Studio",
      "accuracy": 0.8223905000000001,
      "accuracy_std_dev": 0.0244105,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.2415952904040404,
      "avg_duration_per_task_ms": 181953.0101010101,
      "avg_output_tokens_per_task": 24066.49494949495,
      "last_run_timestamp": "2026-10-01T19:12:31.045Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "StreamLake",
      "accuracy": 0.8223905,
      "accuracy_std_dev": 0.021663157431839608,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.10283124074074075,
      "avg_duration_per_task_ms": 597427.9688552188,
      "avg_output_tokens_per_task": 38154.76262626263,
      "last_run_timestamp": "2026-08-29T18:26:14.942Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Friendli",
      "accuracy": 0.8215404398976982,
      "accuracy_std_dev": 0.032891408388902345,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.02587831713554987,
      "avg_duration_per_task_ms": 460054.41943734017,
      "avg_output_tokens_per_task": 21496.38874680307,
      "last_run_timestamp": "2026-08-29T19:19:06.625Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Makora",
      "accuracy": 0.8213506405228759,
      "accuracy_std_dev": 0.02248216924587128,
      "run_count": 3,
      "total_questions": 459,
      "avg_cost_per_task_usd": 0.08758356862745097,
      "avg_duration_per_task_ms": 304618.9760348584,
      "avg_output_tokens_per_task": 34720.03050108933,
      "last_run_timestamp": "2026-08-29T10:57:24.797Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Parasail",
      "accuracy": 0.8212070710659898,
      "accuracy_std_dev": 0.011946281848340899,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.01807649915397631,
      "avg_duration_per_task_ms": 302940.0592216582,
      "avg_output_tokens_per_task": 16413.382402707277,
      "last_run_timestamp": "2026-08-29T09:59:35.603Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Alibaba",
      "accuracy": 0.8211085620915033,
      "accuracy_std_dev": 0.010055056516550273,
      "run_count": 7,
      "total_questions": 1377,
      "avg_cost_per_task_usd": 0.022989802469135803,
      "avg_duration_per_task_ms": 339486.9368191721,
      "avg_output_tokens_per_task": 20491.059549745823,
      "last_run_timestamp": "2026-10-04T00:17:30.035Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Io Net",
      "accuracy": 0.8210525603238865,
      "accuracy_std_dev": 0.07599126782917703,
      "run_count": 7,
      "total_questions": 1235,
      "avg_cost_per_task_usd": 0.007044106882591094,
      "avg_duration_per_task_ms": 525854.9595141701,
      "avg_output_tokens_per_task": 14201.662348178137,
      "last_run_timestamp": "2026-10-03T10:22:02.091Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.8210436999999998,
      "accuracy_std_dev": 0.014005426348740687,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.3147447828282829,
      "avg_duration_per_task_ms": 289184.82020202023,
      "avg_output_tokens_per_task": 20793.45303030303,
      "last_run_timestamp": "2026-10-03T09:11:51.622Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.8208440232558138,
      "accuracy_std_dev": 0.005892000000000008,
      "run_count": 2,
      "total_questions": 387,
      "avg_cost_per_task_usd": 0.007978992248062014,
      "avg_duration_per_task_ms": 646108.8320413437,
      "avg_output_tokens_per_task": 19074.485788113696,
      "last_run_timestamp": "2026-08-01T18:07:34.003Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "AkashML",
      "accuracy": 0.820801884564352,
      "accuracy_std_dev": 0.01789318941878136,
      "run_count": 12,
      "total_questions": 2261,
      "avg_cost_per_task_usd": 0.03227863688633348,
      "avg_duration_per_task_ms": 262227.01946041576,
      "avg_output_tokens_per_task": 14197.505528527201,
      "last_run_timestamp": "2026-10-09T15:42:38.051Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Venice",
      "accuracy": 0.82070725,
      "accuracy_std_dev": 0.0036691031857253717,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.1439596148989899,
      "avg_duration_per_task_ms": 393442.0984848485,
      "avg_output_tokens_per_task": 28794.420454545456,
      "last_run_timestamp": "2026-10-03T09:54:41.255Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8205311347826086,
      "accuracy_std_dev": 0.003366096503885395,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.007773560869565217,
      "avg_duration_per_task_ms": 433269.1550724638,
      "avg_output_tokens_per_task": 19994.68913043478,
      "last_run_timestamp": "2026-10-04T00:12:29.471Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "Novita",
      "accuracy": 0.8202530101265821,
      "accuracy_std_dev": 0.008020329102765456,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.018872242194092826,
      "avg_duration_per_task_ms": 516686.08607594937,
      "avg_output_tokens_per_task": 18743.98734177215,
      "last_run_timestamp": "2026-08-29T18:35:57.732Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "SambaNova",
      "accuracy": 0.820012649122807,
      "accuracy_std_dev": 0.022660477062222247,
      "run_count": 4,
      "total_questions": 513,
      "avg_cost_per_task_usd": 0.08905661598440547,
      "avg_duration_per_task_ms": 310577.5692007797,
      "avg_output_tokens_per_task": 36897.34307992203,
      "last_run_timestamp": "2026-10-03T17:40:40.949Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.819865,
      "accuracy_std_dev": 0.006733999999999962,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.008803729797979798,
      "avg_duration_per_task_ms": 1022564.7878787878,
      "avg_output_tokens_per_task": 21115.48232323232,
      "last_run_timestamp": "2026-08-01T22:57:01.346Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "StreamLake",
      "accuracy": 0.819865,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.13854771717171718,
      "avg_duration_per_task_ms": 540824.1212121212,
      "avg_output_tokens_per_task": 31240.651515151516,
      "last_run_timestamp": "2026-09-07T06:00:57.184Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8193234782608696,
      "accuracy_std_dev": 0.007347505135369806,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.007821787681159422,
      "avg_duration_per_task_ms": 636163.7028985508,
      "avg_output_tokens_per_task": 18195.539130434783,
      "last_run_timestamp": "2026-10-04T01:46:36.499Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Chutes",
      "accuracy": 0.8193039999999999,
      "accuracy_std_dev": 0.01926018039029404,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0067461641414141415,
      "avg_duration_per_task_ms": 634323.356902357,
      "avg_output_tokens_per_task": 18174.79292929293,
      "last_run_timestamp": "2026-10-03T17:09:07.230Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8192329035532995,
      "accuracy_std_dev": 0.01830389154129678,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.022038197123519458,
      "avg_duration_per_task_ms": 499412.81133671745,
      "avg_output_tokens_per_task": 22395.89847715736,
      "last_run_timestamp": "2026-08-29T18:16:26.161Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Novita",
      "accuracy": 0.8191919000000001,
      "accuracy_std_dev": 0.0203168642878275,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.012244186363636363,
      "avg_duration_per_task_ms": 659276.7813131313,
      "avg_output_tokens_per_task": 36210.856060606064,
      "last_run_timestamp": "2026-10-03T10:16:01.896Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8181816666666667,
      "accuracy_std_dev": 0.013255807590972678,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.17135338047138046,
      "avg_duration_per_task_ms": 45814.95117845118,
      "avg_output_tokens_per_task": 3234.355218855219,
      "last_run_timestamp": "2026-08-15T09:15:35.213Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "BaseTen",
      "accuracy": 0.818035862576065,
      "accuracy_std_dev": 0.01278056208505712,
      "run_count": 10,
      "total_questions": 1972,
      "avg_cost_per_task_usd": 0.0909173985801217,
      "avg_duration_per_task_ms": 165913.36257606492,
      "avg_output_tokens_per_task": 22308.299188640973,
      "last_run_timestamp": "2026-09-19T10:14:07.966Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.8179946666666665,
      "accuracy_std_dev": 0.06050420910059803,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.027902544332211,
      "avg_duration_per_task_ms": 645703.9062850729,
      "avg_output_tokens_per_task": 28951.189674523008,
      "last_run_timestamp": "2026-10-03T10:05:26.294Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Novita",
      "accuracy": 0.8179010000000001,
      "accuracy_std_dev": 0.032213432208733475,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.09231441077441079,
      "avg_duration_per_task_ms": 1504600.9158249157,
      "avg_output_tokens_per_task": 43006.99410774411,
      "last_run_timestamp": "2026-10-03T17:44:28.471Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Phala",
      "accuracy": 0.817901,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 54,
      "avg_cost_per_task_usd": 0.008988981481481481,
      "avg_duration_per_task_ms": 114091.12962962964,
      "avg_output_tokens_per_task": 16251.87037037037,
      "last_run_timestamp": "2026-07-18T09:01:08.846Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Together",
      "accuracy": 0.8178077777777777,
      "accuracy_std_dev": 0.009999885885768644,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.024910318742985414,
      "avg_duration_per_task_ms": 190389.1571268238,
      "avg_output_tokens_per_task": 16514.1734006734,
      "last_run_timestamp": "2026-10-03T10:10:35.343Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.8178076666666667,
      "accuracy_std_dev": 0.020010550661532987,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.04244091638608306,
      "avg_duration_per_task_ms": 255376.0824915825,
      "avg_output_tokens_per_task": 14176.158810325476,
      "last_run_timestamp": "2026-10-03T10:33:26.506Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8169791428571429,
      "accuracy_std_dev": 0.022383364227087234,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.04123865007215007,
      "avg_duration_per_task_ms": 175510.44805194804,
      "avg_output_tokens_per_task": 22808.474025974025,
      "last_run_timestamp": "2026-10-03T17:30:27.921Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Phala",
      "accuracy": 0.816967559819413,
      "accuracy_std_dev": 0.022532849378432182,
      "run_count": 9,
      "total_questions": 1772,
      "avg_cost_per_task_usd": 0.021389000564334085,
      "avg_duration_per_task_ms": 297135.54796839727,
      "avg_output_tokens_per_task": 17946.336343115123,
      "last_run_timestamp": "2026-10-03T09:25:02.390Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.8168942867981791,
      "accuracy_std_dev": 0.04478955350480735,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.08493498533131005,
      "avg_duration_per_task_ms": 815555.1937278705,
      "avg_output_tokens_per_task": 31553.02023267577,
      "last_run_timestamp": "2026-10-03T19:00:29.739Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8165714782608695,
      "accuracy_std_dev": 0.012063277850998734,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.3338848906455862,
      "avg_duration_per_task_ms": 299201.53820816864,
      "avg_output_tokens_per_task": 21149.658322353975,
      "last_run_timestamp": "2026-10-03T09:43:11.426Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Novita",
      "accuracy": 0.8162174999999999,
      "accuracy_std_dev": 0.009642570572725913,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.019066546296296293,
      "avg_duration_per_task_ms": 525999.5555555555,
      "avg_output_tokens_per_task": 18868.49074074074,
      "last_run_timestamp": "2026-08-30T00:16:38.394Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.815937,
      "accuracy_std_dev": 0.021042110493009022,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04137441245791246,
      "avg_duration_per_task_ms": 82874.20875420875,
      "avg_output_tokens_per_task": 27465.196969696968,
      "last_run_timestamp": "2026-10-01T19:23:01.327Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "WandB",
      "accuracy": 0.815838,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 181,
      "avg_cost_per_task_usd": 0.03376955248618784,
      "avg_duration_per_task_ms": 178211.5359116022,
      "avg_output_tokens_per_task": 26893.3591160221,
      "last_run_timestamp": "2026-07-18T08:04:57.082Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DeepInfra",
      "accuracy": 0.8156999235181644,
      "accuracy_std_dev": 0.04951986238986445,
      "run_count": 9,
      "total_questions": 1569,
      "avg_cost_per_task_usd": 0.04647088910133843,
      "avg_duration_per_task_ms": 572516.0726577438,
      "avg_output_tokens_per_task": 26962.156787762906,
      "last_run_timestamp": "2026-09-26T09:56:00.145Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Google",
      "accuracy": 0.8156565,
      "accuracy_std_dev": 0.01038452301023016,
      "run_count": 20,
      "total_questions": 3960,
      "avg_cost_per_task_usd": 0.36446038207070713,
      "avg_duration_per_task_ms": 359197.3244949495,
      "avg_output_tokens_per_task": 22956.0851010101,
      "last_run_timestamp": "2026-10-03T09:59:43.843Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "auto-routing",
      "accuracy": 0.8154587608695651,
      "accuracy_std_dev": 0.035429792247399325,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.07787585652173913,
      "avg_duration_per_task_ms": 892615.7195652174,
      "avg_output_tokens_per_task": 39764.602173913045,
      "last_run_timestamp": "2026-10-04T00:10:24.974Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8152592970976255,
      "accuracy_std_dev": 0.06424712746051377,
      "run_count": 20,
      "total_questions": 3790,
      "avg_cost_per_task_usd": 0.007318012664907651,
      "avg_duration_per_task_ms": 480925.70474934037,
      "avg_output_tokens_per_task": 17239.7055408971,
      "last_run_timestamp": "2026-10-08T22:38:34.721Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Fireworks",
      "accuracy": 0.8143338571428571,
      "accuracy_std_dev": 0.014248444761532711,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.026212479076479073,
      "avg_duration_per_task_ms": 329590.34992784995,
      "avg_output_tokens_per_task": 17511.52380952381,
      "last_run_timestamp": "2026-09-19T09:05:52.633Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Relace",
      "accuracy": 0.814024,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 164,
      "avg_cost_per_task_usd": 0.010100225609756097,
      "avg_duration_per_task_ms": 515065.23780487804,
      "avg_output_tokens_per_task": 33245.84146341463,
      "last_run_timestamp": "2026-09-28T09:46:11.365Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "StreamLake",
      "accuracy": 0.8139730000000001,
      "accuracy_std_dev": 0.007059363734312956,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.006141271885521885,
      "avg_duration_per_task_ms": 521938.39983164985,
      "avg_output_tokens_per_task": 18689.60521885522,
      "last_run_timestamp": "2026-08-29T19:09:30.030Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Morph",
      "accuracy": 0.8136272825651302,
      "accuracy_std_dev": 0.008773473238499465,
      "run_count": 3,
      "total_questions": 499,
      "avg_cost_per_task_usd": 0.0068851422845691386,
      "avg_duration_per_task_ms": 1090136,
      "avg_output_tokens_per_task": 17404.21643286573,
      "last_run_timestamp": "2026-08-15T10:05:32.709Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Novita",
      "accuracy": 0.813518124726477,
      "accuracy_std_dev": 0.03129658994855769,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.2371703887673231,
      "avg_duration_per_task_ms": 955825.8431801605,
      "avg_output_tokens_per_task": 54088.58570386579,
      "last_run_timestamp": "2026-10-03T17:05:40.300Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Venice",
      "accuracy": 0.8133717142857142,
      "accuracy_std_dev": 0.009499021449064938,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.029931976190476193,
      "avg_duration_per_task_ms": 201570.87085137086,
      "avg_output_tokens_per_task": 23744.84992784993,
      "last_run_timestamp": "2026-10-03T17:45:01.711Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-pro",
      "model_slug": "google/gemini-2.5-pro",
      "model_name": "Google: Gemini 2.5 Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.8131313333333333,
      "accuracy_std_dev": 0.02212166805545088,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.2550713585858586,
      "avg_duration_per_task_ms": 227479.37878787878,
      "avg_output_tokens_per_task": 25491.23232323232,
      "last_run_timestamp": "2026-10-01T19:12:42.150Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.2-chat-20251211",
      "model_slug": "openai/gpt-5.2-chat",
      "model_name": "OpenAI: GPT-5.2 Chat",
      "provider_name": "OpenAI",
      "accuracy": 0.813131,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.037260742424242425,
      "avg_duration_per_task_ms": 35156.99494949495,
      "avg_output_tokens_per_task": 2566.8030303030305,
      "last_run_timestamp": "2026-08-01T08:32:23.484Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Darkbloom",
      "accuracy": 0.8128504999999999,
      "accuracy_std_dev": 0.008607050515904558,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.03321095707070707,
      "avg_duration_per_task_ms": 706759.8703703703,
      "avg_output_tokens_per_task": 15781.767676767677,
      "last_run_timestamp": "2026-10-03T09:29:30.693Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "CoreWeave",
      "accuracy": 0.8122894999999999,
      "accuracy_std_dev": 0.020078789204448232,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.006274926767676768,
      "avg_duration_per_task_ms": 514195.62205387204,
      "avg_output_tokens_per_task": 18652.397306397306,
      "last_run_timestamp": "2026-10-03T16:58:24.968Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "Together",
      "accuracy": 0.812016,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 172,
      "avg_cost_per_task_usd": 0.14135672674418603,
      "avg_duration_per_task_ms": 305120.88372093026,
      "avg_output_tokens_per_task": 39843.74418604651,
      "last_run_timestamp": "2026-08-04T03:11:59.879Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.8118160109409189,
      "accuracy_std_dev": 0.026061913159839602,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.2012101502552881,
      "avg_duration_per_task_ms": 871685.5309992706,
      "avg_output_tokens_per_task": 54661.03574033552,
      "last_run_timestamp": "2026-10-03T17:03:42.533Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AkashML",
      "accuracy": 0.811448,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.004932363636363637,
      "avg_duration_per_task_ms": 1604879.5959595959,
      "avg_output_tokens_per_task": 24145.00505050505,
      "last_run_timestamp": "2026-07-18T08:42:46.001Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "StreamLake",
      "accuracy": 0.8112072857142857,
      "accuracy_std_dev": 0.030080678790376323,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.014970598124098124,
      "avg_duration_per_task_ms": 908622.7445887446,
      "avg_output_tokens_per_task": 44238.631313131315,
      "last_run_timestamp": "2026-10-03T08:04:44.082Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "auto-routing",
      "accuracy": 0.8108866666666668,
      "accuracy_std_dev": 0.013699776466140677,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.006567723063973063,
      "avg_duration_per_task_ms": 517378.0252525252,
      "avg_output_tokens_per_task": 17537.287037037036,
      "last_run_timestamp": "2026-10-03T17:03:40.743Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.8103253333333333,
      "accuracy_std_dev": 0.011527714209195522,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04107354882154882,
      "avg_duration_per_task_ms": 75169.15151515152,
      "avg_output_tokens_per_task": 27258.996632996634,
      "last_run_timestamp": "2026-10-01T19:16:58.595Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.8099964482758623,
      "accuracy_std_dev": 0.03423948248791152,
      "run_count": 5,
      "total_questions": 957,
      "avg_cost_per_task_usd": 0.07441870846394985,
      "avg_duration_per_task_ms": 1390345.897596656,
      "avg_output_tokens_per_task": 42629.52142110763,
      "last_run_timestamp": "2026-08-22T17:50:04.003Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "GMICloud",
      "accuracy": 0.8095960000000001,
      "accuracy_std_dev": 0.036971651099186785,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.11996981868686869,
      "avg_duration_per_task_ms": 468196.27575757576,
      "avg_output_tokens_per_task": 33838.74393939394,
      "last_run_timestamp": "2026-10-03T16:58:37.335Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Relace",
      "accuracy": 0.8091166153846153,
      "accuracy_std_dev": 0.03183613126622014,
      "run_count": 3,
      "total_questions": 585,
      "avg_cost_per_task_usd": 0.19830280854700855,
      "avg_duration_per_task_ms": 1060541.5008547008,
      "avg_output_tokens_per_task": 55592.3264957265,
      "last_run_timestamp": "2026-10-03T16:58:49.489Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Z.AI",
      "accuracy": 0.8087628910675383,
      "accuracy_std_dev": 0.05175650112216013,
      "run_count": 7,
      "total_questions": 1377,
      "avg_cost_per_task_usd": 0.23057392955700798,
      "avg_duration_per_task_ms": 1023108.186637618,
      "avg_output_tokens_per_task": 52917.13652868555,
      "last_run_timestamp": "2026-10-03T17:12:36.980Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "OpenInference",
      "accuracy": 0.8087552731239093,
      "accuracy_std_dev": 0.0546199990705989,
      "run_count": 6,
      "total_questions": 1146,
      "avg_cost_per_task_usd": 0.024335609075043632,
      "avg_duration_per_task_ms": 1150538.0907504363,
      "avg_output_tokens_per_task": 49299.98254799302,
      "last_run_timestamp": "2026-10-03T10:11:21.294Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.808728230769231,
      "accuracy_std_dev": 0.016181671721919724,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.34003047086247085,
      "avg_duration_per_task_ms": 354646.60295260296,
      "avg_output_tokens_per_task": 22500.726495726496,
      "last_run_timestamp": "2026-10-03T09:08:27.457Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "StreamLake",
      "accuracy": 0.8085015000000001,
      "accuracy_std_dev": 0.026787617671230127,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.1100901268939394,
      "avg_duration_per_task_ms": 771715.6508838384,
      "avg_output_tokens_per_task": 57264.88383838384,
      "last_run_timestamp": "2026-10-03T17:13:30.899Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.808081,
      "accuracy_std_dev": 0.018027342695657245,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.044340436026936034,
      "avg_duration_per_task_ms": 217888.1734006734,
      "avg_output_tokens_per_task": 22075.030303030304,
      "last_run_timestamp": "2026-10-01T19:19:28.473Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Mancer 2",
      "accuracy": 0.8072389999999999,
      "accuracy_std_dev": 0.0870155432092451,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.010984919191919193,
      "avg_duration_per_task_ms": 659527.0782828283,
      "avg_output_tokens_per_task": 25386.00420875421,
      "last_run_timestamp": "2026-10-03T17:34:53.587Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Google",
      "accuracy": 0.8066155190839694,
      "accuracy_std_dev": 0.009958499999999981,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.07586704071246819,
      "avg_duration_per_task_ms": 385930.98473282444,
      "avg_output_tokens_per_task": 32444.25445292621,
      "last_run_timestamp": "2026-08-01T17:52:55.742Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Azure",
      "accuracy": 0.8065657,
      "accuracy_std_dev": 0.022453683925137975,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.3406123287878787,
      "avg_duration_per_task_ms": 355782.31666666665,
      "avg_output_tokens_per_task": 22552.313636363637,
      "last_run_timestamp": "2026-10-03T09:49:05.353Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8063972857142858,
      "accuracy_std_dev": 0.04410277557094591,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.04038211976911977,
      "avg_duration_per_task_ms": 957664.380952381,
      "avg_output_tokens_per_task": 25231.96176046176,
      "last_run_timestamp": "2026-09-26T09:34:13.954Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-pro",
      "model_slug": "google/gemini-2.5-pro",
      "model_name": "Google: Gemini 2.5 Pro",
      "provider_name": "Google",
      "accuracy": 0.806210222222222,
      "accuracy_std_dev": 0.03697273211969953,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.25474633557800225,
      "avg_duration_per_task_ms": 231596.29349046017,
      "avg_output_tokens_per_task": 25466.63243546577,
      "last_run_timestamp": "2026-10-01T19:17:25.275Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "auto-routing",
      "accuracy": 0.8059763749999999,
      "accuracy_std_dev": 0.01367028747811745,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.022531703914141412,
      "avg_duration_per_task_ms": 251092.985479798,
      "avg_output_tokens_per_task": 17458.415404040403,
      "last_run_timestamp": "2026-10-03T09:13:36.277Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8058417989690723,
      "accuracy_std_dev": 0.019860476718151096,
      "run_count": 6,
      "total_questions": 1164,
      "avg_cost_per_task_usd": 0.1695892783505155,
      "avg_duration_per_task_ms": 882585.4982817869,
      "avg_output_tokens_per_task": 53988.948453608245,
      "last_run_timestamp": "2026-08-29T19:16:33.097Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Chutes",
      "accuracy": 0.805556,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 192,
      "avg_cost_per_task_usd": 0.16585933333333333,
      "avg_duration_per_task_ms": 688743.8177083334,
      "avg_output_tokens_per_task": 42426.447916666664,
      "last_run_timestamp": "2026-07-31T00:21:35.641Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DekaLLM",
      "accuracy": 0.805555546875,
      "accuracy_std_dev": 0.01288499999999998,
      "run_count": 2,
      "total_questions": 384,
      "avg_cost_per_task_usd": 0.0362872109375,
      "avg_duration_per_task_ms": 729621.7864583334,
      "avg_output_tokens_per_task": 36271.864583333336,
      "last_run_timestamp": "2026-09-12T16:45:54.034Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Io Net",
      "accuracy": 0.805068,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 171,
      "avg_cost_per_task_usd": 0.07262640935672515,
      "avg_duration_per_task_ms": 352658,
      "avg_output_tokens_per_task": 15572.333333333334,
      "last_run_timestamp": "2026-07-18T08:30:38.228Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8047647874762808,
      "accuracy_std_dev": 0.032856312422424996,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.016785316255534474,
      "avg_duration_per_task_ms": 574047.4073371284,
      "avg_output_tokens_per_task": 17963.88172043011,
      "last_run_timestamp": "2026-10-05T20:59:39.536Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.3-chat-20260303",
      "model_slug": "openai/gpt-5.3-chat",
      "model_name": "OpenAI: GPT-5.3 Chat",
      "provider_name": "auto-routing",
      "accuracy": 0.804714,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.024081237373737373,
      "avg_duration_per_task_ms": 24354.015151515152,
      "avg_output_tokens_per_task": 1625.5555555555557,
      "last_run_timestamp": "2026-08-01T08:38:11.900Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "auto-routing",
      "accuracy": 0.8047135714285714,
      "accuracy_std_dev": 0.022568642992670425,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.007838231601731601,
      "avg_duration_per_task_ms": 497728.7373737374,
      "avg_output_tokens_per_task": 18010.08152958153,
      "last_run_timestamp": "2026-10-04T00:40:00.465Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Cloudflare",
      "accuracy": 0.8037036,
      "accuracy_std_dev": 0.10533181249480138,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.16579600808080808,
      "avg_duration_per_task_ms": 1110510.5636363637,
      "avg_output_tokens_per_task": 45531.156565656565,
      "last_run_timestamp": "2026-10-03T17:43:01.258Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Mara",
      "accuracy": 0.8034189230769232,
      "accuracy_std_dev": 0.024712500000000026,
      "run_count": 2,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.08983526153846154,
      "avg_duration_per_task_ms": 329694.73846153845,
      "avg_output_tokens_per_task": 37199.13846153846,
      "last_run_timestamp": "2026-10-03T16:53:50.090Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8032406250000002,
      "accuracy_std_dev": 0.023560394356512266,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.3350150732323232,
      "avg_duration_per_task_ms": 352876.84217171714,
      "avg_output_tokens_per_task": 22221.229166666668,
      "last_run_timestamp": "2026-10-03T09:34:21.327Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "Alibaba",
      "accuracy": 0.8030233054141601,
      "accuracy_std_dev": 0.11015684719346006,
      "run_count": 11,
      "total_questions": 2161,
      "avg_cost_per_task_usd": 0.046368811198519204,
      "avg_duration_per_task_ms": 182261.37390097178,
      "avg_output_tokens_per_task": 22191.346136048127,
      "last_run_timestamp": "2026-10-03T10:14:18.163Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.8020951111111111,
      "accuracy_std_dev": 0.027492606765554845,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.04329651571268237,
      "avg_duration_per_task_ms": 84651.98765432098,
      "avg_output_tokens_per_task": 26957.542087542086,
      "last_run_timestamp": "2026-10-01T19:25:58.919Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "DeepInfra",
      "accuracy": 0.8016821288180611,
      "accuracy_std_dev": 0.026730709555864766,
      "run_count": 5,
      "total_questions": 753,
      "avg_cost_per_task_usd": 0.050311730411686593,
      "avg_duration_per_task_ms": 763057.3041168659,
      "avg_output_tokens_per_task": 21804.07835325365,
      "last_run_timestamp": "2026-08-22T16:06:14.894Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Io Net",
      "accuracy": 0.8016385562742562,
      "accuracy_std_dev": 0.007745278218856984,
      "run_count": 4,
      "total_questions": 773,
      "avg_cost_per_task_usd": 0.0594386532988357,
      "avg_duration_per_task_ms": 220979.64165588617,
      "avg_output_tokens_per_task": 17329.77360931436,
      "last_run_timestamp": "2026-09-26T08:53:08.262Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8013615579868709,
      "accuracy_std_dev": 0.030272935016787066,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.2393555025528811,
      "avg_duration_per_task_ms": 897241.6353026987,
      "avg_output_tokens_per_task": 60323.876002917576,
      "last_run_timestamp": "2026-10-03T16:51:04.342Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Mara",
      "accuracy": 0.8011870934718102,
      "accuracy_std_dev": 0.030695975976321205,
      "run_count": 6,
      "total_questions": 674,
      "avg_cost_per_task_usd": 0.030323706231454008,
      "avg_duration_per_task_ms": 103892.11572700297,
      "avg_output_tokens_per_task": 31484.501483679524,
      "last_run_timestamp": "2026-08-29T19:44:46.543Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8008437316455697,
      "accuracy_std_dev": 0.022930653266364256,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.028163265822784812,
      "avg_duration_per_task_ms": 479213.82784810127,
      "avg_output_tokens_per_task": 23504.80253164557,
      "last_run_timestamp": "2026-08-29T18:09:57.600Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.8007853333333332,
      "accuracy_std_dev": 0.008398773614178595,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0439676531986532,
      "avg_duration_per_task_ms": 201662.1632996633,
      "avg_output_tokens_per_task": 21888.12962962963,
      "last_run_timestamp": "2026-10-01T19:26:25.385Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Phala",
      "accuracy": 0.8006493194805195,
      "accuracy_std_dev": 0.04527692613517397,
      "run_count": 4,
      "total_questions": 770,
      "avg_cost_per_task_usd": 0.00850762987012987,
      "avg_duration_per_task_ms": 622509.1441558441,
      "avg_output_tokens_per_task": 19295.080519480518,
      "last_run_timestamp": "2026-08-29T19:08:46.030Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "ModelRun",
      "accuracy": 0.7996635,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.046308881313131316,
      "avg_duration_per_task_ms": 53641.583333333336,
      "avg_output_tokens_per_task": 15349.742424242424,
      "last_run_timestamp": "2026-10-05T17:49:50.156Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "Azure",
      "accuracy": 0.7996631666666666,
      "accuracy_std_dev": 0.024569953550469384,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.045980551346801346,
      "avg_duration_per_task_ms": 194150.27525252526,
      "avg_output_tokens_per_task": 21808.128787878788,
      "last_run_timestamp": "2026-10-01T19:34:33.810Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Friendli",
      "accuracy": 0.7987012857142858,
      "accuracy_std_dev": 0.02392447384532228,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.028126163059163056,
      "avg_duration_per_task_ms": 291404.7186147186,
      "avg_output_tokens_per_task": 18604.992784992784,
      "last_run_timestamp": "2026-10-04T00:12:29.472Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.7987011428571429,
      "accuracy_std_dev": 0.04484802990553949,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.02699170274170274,
      "avg_duration_per_task_ms": 370961.9646464646,
      "avg_output_tokens_per_task": 25531.39466089466,
      "last_run_timestamp": "2026-10-03T19:56:06.052Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.2-chat-20251211",
      "model_slug": "openai/gpt-5.2-chat",
      "model_name": "OpenAI: GPT-5.2 Chat",
      "provider_name": "auto-routing",
      "accuracy": 0.79798,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.03602157070707071,
      "avg_duration_per_task_ms": 32858.843434343435,
      "avg_output_tokens_per_task": 2479.818181818182,
      "last_run_timestamp": "2026-08-01T08:32:36.352Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7979798571428571,
      "accuracy_std_dev": 0.006036422413236218,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.05221803823953824,
      "avg_duration_per_task_ms": 1137486.8282828282,
      "avg_output_tokens_per_task": 23798.989898989897,
      "last_run_timestamp": "2026-10-03T18:46:58.291Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7979796666666668,
      "accuracy_std_dev": 0.009816464683830371,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.01347027946127946,
      "avg_duration_per_task_ms": 287878.5387205387,
      "avg_output_tokens_per_task": 14030.314814814816,
      "last_run_timestamp": "2026-08-29T19:12:57.906Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Io Net",
      "accuracy": 0.7976190000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.012265704081632654,
      "avg_duration_per_task_ms": 285854.5612244898,
      "avg_output_tokens_per_task": 10545.989795918367,
      "last_run_timestamp": "2026-10-03T17:16:45.102Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "WandB",
      "accuracy": 0.797481,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 172,
      "avg_cost_per_task_usd": 0.07670138953488372,
      "avg_duration_per_task_ms": 240811.48837209304,
      "avg_output_tokens_per_task": 21222.406976744187,
      "last_run_timestamp": "2026-07-18T08:37:00.147Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "Novita",
      "accuracy": 0.7972860381679391,
      "accuracy_std_dev": 0.010877999999999999,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.01688720610687023,
      "avg_duration_per_task_ms": 167801.82442748093,
      "avg_output_tokens_per_task": 6563.4376590330785,
      "last_run_timestamp": "2026-08-01T17:44:11.144Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7972582857142857,
      "accuracy_std_dev": 0.005891180301683102,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.014391279220779221,
      "avg_duration_per_task_ms": 318122.0873015873,
      "avg_output_tokens_per_task": 14823.277056277057,
      "last_run_timestamp": "2026-10-04T00:00:36.942Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Fireworks",
      "accuracy": 0.797043,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 124,
      "avg_cost_per_task_usd": 0.11967770967741935,
      "avg_duration_per_task_ms": 375576.0080645161,
      "avg_output_tokens_per_task": 29097.58064516129,
      "last_run_timestamp": "2026-07-18T08:32:17.535Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "StreamLake",
      "accuracy": 0.7968575,
      "accuracy_std_dev": 0.009506569460992069,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.015575840067340066,
      "avg_duration_per_task_ms": 306607.4048821549,
      "avg_output_tokens_per_task": 14907.355218855218,
      "last_run_timestamp": "2026-08-29T16:10:33.070Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Makora",
      "accuracy": 0.7964375923664123,
      "accuracy_std_dev": 0.11378331267012533,
      "run_count": 4,
      "total_questions": 655,
      "avg_cost_per_task_usd": 0.011028116030534352,
      "avg_duration_per_task_ms": 270893.28854961833,
      "avg_output_tokens_per_task": 23218.567938931297,
      "last_run_timestamp": "2026-09-12T08:53:48.637Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.7962963,
      "accuracy_std_dev": 0.10523964572541092,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.0923363292929293,
      "avg_duration_per_task_ms": 515255.396969697,
      "avg_output_tokens_per_task": 34061.135353535356,
      "last_run_timestamp": "2026-10-03T17:43:17.222Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Novita",
      "accuracy": 0.7956228,
      "accuracy_std_dev": 0.021522407276138952,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.006844225252525253,
      "avg_duration_per_task_ms": 363476.91515151516,
      "avg_output_tokens_per_task": 16652.98888888889,
      "last_run_timestamp": "2026-08-29T19:20:50.265Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.7956228,
      "accuracy_std_dev": 0.009072301348610534,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.051206081818181814,
      "avg_duration_per_task_ms": 443200.41616161616,
      "avg_output_tokens_per_task": 24093.476767676766,
      "last_run_timestamp": "2026-08-22T17:18:01.115Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Io Net",
      "accuracy": 0.7950336666666666,
      "accuracy_std_dev": 0.03307482054205114,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.03994933501683502,
      "avg_duration_per_task_ms": 251252.56818181818,
      "avg_output_tokens_per_task": 12806.098484848484,
      "last_run_timestamp": "2026-09-19T09:05:41.015Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Alibaba",
      "accuracy": 0.7947205871946083,
      "accuracy_std_dev": 0.04278072577075129,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.1992195315922494,
      "avg_duration_per_task_ms": 633124.2401010952,
      "avg_output_tokens_per_task": 47420.48188711036,
      "last_run_timestamp": "2026-10-03T17:09:08.099Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "SambaNova",
      "accuracy": 0.794007441011236,
      "accuracy_std_dev": 0.03030512423976348,
      "run_count": 4,
      "total_questions": 356,
      "avg_cost_per_task_usd": 0.07293162640449438,
      "avg_duration_per_task_ms": 115799.61235955056,
      "avg_output_tokens_per_task": 31686.629213483146,
      "last_run_timestamp": "2026-08-15T14:54:40.470Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Venice",
      "accuracy": 0.7932827976391231,
      "accuracy_std_dev": 0.12137721453581074,
      "run_count": 6,
      "total_questions": 1186,
      "avg_cost_per_task_usd": 0.12359699156829679,
      "avg_duration_per_task_ms": 670423.452782462,
      "avg_output_tokens_per_task": 39297.81534569983,
      "last_run_timestamp": "2026-08-29T18:23:09.250Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7929293333333333,
      "accuracy_std_dev": 0.008361060791284552,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.01913193771043771,
      "avg_duration_per_task_ms": 454912.569023569,
      "avg_output_tokens_per_task": 25421.24579124579,
      "last_run_timestamp": "2026-10-03T18:54:19.540Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7924050379746834,
      "accuracy_std_dev": 0.007365480183486586,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.1897013772151899,
      "avg_duration_per_task_ms": 842584.7670886077,
      "avg_output_tokens_per_task": 47454.72573839663,
      "last_run_timestamp": "2026-10-03T17:06:15.877Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7923849784482757,
      "accuracy_std_dev": 0.02939258755007609,
      "run_count": 3,
      "total_questions": 464,
      "avg_cost_per_task_usd": 0.010662515086206897,
      "avg_duration_per_task_ms": 1267747.5711206896,
      "avg_output_tokens_per_task": 37807.1400862069,
      "last_run_timestamp": "2026-10-08T22:56:38.634Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7920900579268293,
      "accuracy_std_dev": 0.011512971285033247,
      "run_count": 10,
      "total_questions": 1968,
      "avg_cost_per_task_usd": 0.01824498475609756,
      "avg_duration_per_task_ms": 181543.2469512195,
      "avg_output_tokens_per_task": 15051.83943089431,
      "last_run_timestamp": "2026-10-08T22:29:11.504Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "GMICloud",
      "accuracy": 0.7918084721906924,
      "accuracy_std_dev": 0.14984325869964388,
      "run_count": 9,
      "total_questions": 1762,
      "avg_cost_per_task_usd": 0.016061496027241773,
      "avg_duration_per_task_ms": 528116.8059023836,
      "avg_output_tokens_per_task": 25582.743473325765,
      "last_run_timestamp": "2026-10-03T08:06:43.537Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "DeepInfra",
      "accuracy": 0.7914971899592945,
      "accuracy_std_dev": 0.018448252406647506,
      "run_count": 8,
      "total_questions": 1474,
      "avg_cost_per_task_usd": 0.04180698643147897,
      "avg_duration_per_task_ms": 703517.4267299864,
      "avg_output_tokens_per_task": 19318.71234735414,
      "last_run_timestamp": "2026-10-08T23:28:00.201Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7912458333333334,
      "accuracy_std_dev": 0.01576267340287031,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.01776989141414141,
      "avg_duration_per_task_ms": 61590.219696969696,
      "avg_output_tokens_per_task": 5770.594276094276,
      "last_run_timestamp": "2026-10-03T16:56:50.423Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Io Net",
      "accuracy": 0.790816,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.013702010204081632,
      "avg_duration_per_task_ms": 552474.6071428572,
      "avg_output_tokens_per_task": 38202.7193877551,
      "last_run_timestamp": "2026-10-05T09:56:55.335Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Phala",
      "accuracy": 0.7906953916083916,
      "accuracy_std_dev": 0.10313122487825346,
      "run_count": 9,
      "total_questions": 1716,
      "avg_cost_per_task_usd": 0.07981278321678323,
      "avg_duration_per_task_ms": 812868.9388111888,
      "avg_output_tokens_per_task": 29607.183566433567,
      "last_run_timestamp": "2026-10-03T19:06:55.619Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Chutes",
      "accuracy": 0.7903146482300885,
      "accuracy_std_dev": 0.04010989916950373,
      "run_count": 7,
      "total_questions": 1356,
      "avg_cost_per_task_usd": 0.14060938569321535,
      "avg_duration_per_task_ms": 1540325.2824483777,
      "avg_output_tokens_per_task": 46576.46017699115,
      "last_run_timestamp": "2026-10-03T16:58:27.160Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7900582903225808,
      "accuracy_std_dev": 0.015780706633417893,
      "run_count": 10,
      "total_questions": 1891,
      "avg_cost_per_task_usd": 0.004934497091485986,
      "avg_duration_per_task_ms": 1252981.0317292437,
      "avg_output_tokens_per_task": 32456.054997355895,
      "last_run_timestamp": "2026-10-09T00:11:24.697Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Venice",
      "accuracy": 0.7891350392405064,
      "accuracy_std_dev": 0.020945155322890284,
      "run_count": 8,
      "total_questions": 1580,
      "avg_cost_per_task_usd": 0.06166090379746835,
      "avg_duration_per_task_ms": 176165.68037974683,
      "avg_output_tokens_per_task": 19378.685443037975,
      "last_run_timestamp": "2026-10-03T10:40:10.028Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.7890571,
      "accuracy_std_dev": 0.03611534435790416,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.04233968282828283,
      "avg_duration_per_task_ms": 479457.0171717172,
      "avg_output_tokens_per_task": 42468.52626262626,
      "last_run_timestamp": "2026-10-03T17:31:01.734Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Venice",
      "accuracy": 0.7880317272727273,
      "accuracy_std_dev": 0.018930562586323512,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.004806210284664829,
      "avg_duration_per_task_ms": 261371.71671258035,
      "avg_output_tokens_per_task": 31496.861340679523,
      "last_run_timestamp": "2026-10-03T17:02:47.316Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Together",
      "accuracy": 0.7865109761019649,
      "accuracy_std_dev": 0.015645091156333998,
      "run_count": 10,
      "total_questions": 1883,
      "avg_cost_per_task_usd": 0.008648618693574084,
      "avg_duration_per_task_ms": 393137.65427509294,
      "avg_output_tokens_per_task": 34261.50026553372,
      "last_run_timestamp": "2026-10-03T18:48:40.173Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Phala",
      "accuracy": 0.7861111000000001,
      "accuracy_std_dev": 0.045978898789444694,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.17852181111111112,
      "avg_duration_per_task_ms": 793247.394949495,
      "avg_output_tokens_per_task": 59712.75959595959,
      "last_run_timestamp": "2026-10-03T17:12:34.213Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "DeepInfra",
      "accuracy": 0.7853823536121673,
      "accuracy_std_dev": 0.0034548367906458246,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.013724529784537389,
      "avg_duration_per_task_ms": 462082.7122940431,
      "avg_output_tokens_per_task": 14299.747782002534,
      "last_run_timestamp": "2026-08-15T14:56:55.269Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Phala",
      "accuracy": 0.7835498571428571,
      "accuracy_std_dev": 0.056925857417292156,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.17942975324675325,
      "avg_duration_per_task_ms": 1414227.1096681096,
      "avg_output_tokens_per_task": 45704.106782106785,
      "last_run_timestamp": "2026-10-03T17:00:45.414Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Wafer",
      "accuracy": 0.7832240326797386,
      "accuracy_std_dev": 0.07801749999999996,
      "run_count": 2,
      "total_questions": 153,
      "avg_cost_per_task_usd": 0.07358333333333333,
      "avg_duration_per_task_ms": 500220.6274509804,
      "avg_output_tokens_per_task": 27733.385620915033,
      "last_run_timestamp": "2026-08-15T15:24:07.556Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "StreamLake",
      "accuracy": 0.7820105485714286,
      "accuracy_std_dev": 0.025091739481102143,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.055241464761904766,
      "avg_duration_per_task_ms": 515409.27365079365,
      "avg_output_tokens_per_task": 24015.726984126984,
      "last_run_timestamp": "2026-10-03T16:18:18.642Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "WandB",
      "accuracy": 0.781145,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.020832949494949494,
      "avg_duration_per_task_ms": 266171.0505050505,
      "avg_output_tokens_per_task": 12380.515151515152,
      "last_run_timestamp": "2026-07-18T16:16:42.667Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "CoreWeave",
      "accuracy": 0.7804232857142855,
      "accuracy_std_dev": 0.00964416607835736,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.021076865079365077,
      "avg_duration_per_task_ms": 281668.99494949495,
      "avg_output_tokens_per_task": 12536.775613275613,
      "last_run_timestamp": "2026-10-04T00:50:16.196Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7788999999999999,
      "accuracy_std_dev": 0.011690582106978247,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.018513430134680136,
      "avg_duration_per_task_ms": 636352.8888888889,
      "avg_output_tokens_per_task": 46067.3063973064,
      "last_run_timestamp": "2026-10-03T16:23:37.615Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Baidu",
      "accuracy": 0.7785069474835887,
      "accuracy_std_dev": 0.0632509601166138,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.21840516265499638,
      "avg_duration_per_task_ms": 918904.7702407002,
      "avg_output_tokens_per_task": 66679.34062727935,
      "last_run_timestamp": "2026-10-03T17:10:01.034Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.7784511999999999,
      "accuracy_std_dev": 0.0206074889047647,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.05598715454545455,
      "avg_duration_per_task_ms": 643404.0878787879,
      "avg_output_tokens_per_task": 23899.834343434344,
      "last_run_timestamp": "2026-08-22T17:09:41.940Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "auto-routing",
      "accuracy": 0.7774409,
      "accuracy_std_dev": 0.02011474071644972,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.005779197474747475,
      "avg_duration_per_task_ms": 733463.898989899,
      "avg_output_tokens_per_task": 36574.07828282828,
      "last_run_timestamp": "2026-10-03T19:04:58.692Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "auto-routing",
      "accuracy": 0.7772166666666667,
      "accuracy_std_dev": 0.003459479954495376,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.013080006734006736,
      "avg_duration_per_task_ms": 67697.39225589225,
      "avg_output_tokens_per_task": 10341.513468013469,
      "last_run_timestamp": "2026-10-01T19:12:45.924Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Parasail",
      "accuracy": 0.7769442558139535,
      "accuracy_std_dev": 0.14033063886698557,
      "run_count": 7,
      "total_questions": 1333,
      "avg_cost_per_task_usd": 0.022806729182295572,
      "avg_duration_per_task_ms": 351644.4351087772,
      "avg_output_tokens_per_task": 23363.498124531132,
      "last_run_timestamp": "2026-10-03T19:19:20.712Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7763605976676385,
      "accuracy_std_dev": 0.09509307709474409,
      "run_count": 7,
      "total_questions": 1372,
      "avg_cost_per_task_usd": 0.05297574416909621,
      "avg_duration_per_task_ms": 1717526.1807580176,
      "avg_output_tokens_per_task": 18596.132653061224,
      "last_run_timestamp": "2026-10-05T23:17:57.198Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.77567325,
      "accuracy_std_dev": 0.010934638822910419,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.013401992424242422,
      "avg_duration_per_task_ms": 67621.48106060606,
      "avg_output_tokens_per_task": 10598.125,
      "last_run_timestamp": "2026-10-02T23:20:38.343Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.7756622557947019,
      "accuracy_std_dev": 0.011948399639177628,
      "run_count": 8,
      "total_questions": 1208,
      "avg_cost_per_task_usd": 0.052177753311258276,
      "avg_duration_per_task_ms": 995012.5024834437,
      "avg_output_tokens_per_task": 26633.495860927153,
      "last_run_timestamp": "2026-10-09T00:04:10.638Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Google",
      "accuracy": 0.7751544671163576,
      "accuracy_std_dev": 0.021435175618724375,
      "run_count": 6,
      "total_questions": 1186,
      "avg_cost_per_task_usd": 0.022097016020236088,
      "avg_duration_per_task_ms": 821841.6846543002,
      "avg_output_tokens_per_task": 38407.37015177066,
      "last_run_timestamp": "2026-10-03T16:57:26.262Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Venice",
      "accuracy": 0.7747440605802048,
      "accuracy_std_dev": 0.08568368177148243,
      "run_count": 6,
      "total_questions": 1172,
      "avg_cost_per_task_usd": 0.1293695656996587,
      "avg_duration_per_task_ms": 995226.8643344709,
      "avg_output_tokens_per_task": 37181.283276450515,
      "last_run_timestamp": "2026-10-03T16:47:49.659Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Crusoe",
      "accuracy": 0.774411,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.007553676767676767,
      "avg_duration_per_task_ms": 145590.66666666666,
      "avg_output_tokens_per_task": 30019.929292929293,
      "last_run_timestamp": "2026-10-03T16:43:00.637Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Parasail",
      "accuracy": 0.7742798061728395,
      "accuracy_std_dev": 0.09894578471411179,
      "run_count": 9,
      "total_questions": 810,
      "avg_cost_per_task_usd": 0.003759838271604938,
      "avg_duration_per_task_ms": 204604.40740740742,
      "avg_output_tokens_per_task": 15966.761728395062,
      "last_run_timestamp": "2026-10-03T16:12:33.499Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7741139703903096,
      "accuracy_std_dev": 0.09718478875928833,
      "run_count": 4,
      "total_questions": 743,
      "avg_cost_per_task_usd": 0.004475604306864065,
      "avg_duration_per_task_ms": 312870.95827725437,
      "avg_output_tokens_per_task": 22926.07267833109,
      "last_run_timestamp": "2026-10-08T22:27:01.555Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.7739002377332412,
      "accuracy_std_dev": 0.03919556245291041,
      "run_count": 9,
      "total_questions": 1447,
      "avg_cost_per_task_usd": 0.05904209744298548,
      "avg_duration_per_task_ms": 844538.2218382861,
      "avg_output_tokens_per_task": 50153.10711817553,
      "last_run_timestamp": "2026-08-29T19:07:57.500Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "auto-routing",
      "accuracy": 0.7730078333333333,
      "accuracy_std_dev": 0.02074231892706846,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.015851826599326598,
      "avg_duration_per_task_ms": 424890.48232323234,
      "avg_output_tokens_per_task": 15979.028619528619,
      "last_run_timestamp": "2026-08-29T18:21:38.666Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.7728757843137255,
      "accuracy_std_dev": 0.0424837728239854,
      "run_count": 5,
      "total_questions": 714,
      "avg_cost_per_task_usd": 0.13942839355742298,
      "avg_duration_per_task_ms": 1052327.7044817926,
      "avg_output_tokens_per_task": 40065.01820728291,
      "last_run_timestamp": "2026-08-22T17:30:52.646Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "BaseTen",
      "accuracy": 0.7716751249999998,
      "accuracy_std_dev": 0.010940746814517491,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0148653023989899,
      "avg_duration_per_task_ms": 144716.75757575757,
      "avg_output_tokens_per_task": 29722.781565656565,
      "last_run_timestamp": "2026-10-03T17:01:22.843Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Baidu",
      "accuracy": 0.7715112928416485,
      "accuracy_std_dev": 0.04364907130227424,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.13478754808387564,
      "avg_duration_per_task_ms": 1017266.7816341287,
      "avg_output_tokens_per_task": 61731.43745480839,
      "last_run_timestamp": "2026-10-03T17:38:07.604Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7714368913043479,
      "accuracy_std_dev": 0.014261246953428234,
      "run_count": 3,
      "total_questions": 552,
      "avg_cost_per_task_usd": 0.23206160869565215,
      "avg_duration_per_task_ms": 164099.60869565216,
      "avg_output_tokens_per_task": 15307.03079710145,
      "last_run_timestamp": "2026-09-01T13:33:46.786Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Mara",
      "accuracy": 0.7713930343283582,
      "accuracy_std_dev": 0.03931269277500872,
      "run_count": 6,
      "total_questions": 670,
      "avg_cost_per_task_usd": 0.012700616417910445,
      "avg_duration_per_task_ms": 53684.10149253732,
      "avg_output_tokens_per_task": 16854.058208955223,
      "last_run_timestamp": "2026-10-03T17:06:37.556Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Parasail",
      "accuracy": 0.7711041709956711,
      "accuracy_std_dev": 0.019798063555812727,
      "run_count": 5,
      "total_questions": 924,
      "avg_cost_per_task_usd": 0.020921933982683983,
      "avg_duration_per_task_ms": 243233.94264069264,
      "avg_output_tokens_per_task": 27811.309523809523,
      "last_run_timestamp": "2026-10-03T16:58:21.850Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "Azure",
      "accuracy": 0.770202,
      "accuracy_std_dev": 0.008625232557251217,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.013833271043771043,
      "avg_duration_per_task_ms": 119359.86616161616,
      "avg_output_tokens_per_task": 10422.487373737375,
      "last_run_timestamp": "2026-10-01T19:20:04.986Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "auto-routing",
      "accuracy": 0.770081857142857,
      "accuracy_std_dev": 0.03628441834581173,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.052005661616161616,
      "avg_duration_per_task_ms": 855189.3722943723,
      "avg_output_tokens_per_task": 23489.02380952381,
      "last_run_timestamp": "2026-10-03T18:36:41.004Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.7700529885714287,
      "accuracy_std_dev": 0.007227262196667206,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.011466961904761905,
      "avg_duration_per_task_ms": 1005165.3326984127,
      "avg_output_tokens_per_task": 11934.448888888888,
      "last_run_timestamp": "2026-10-09T00:38:37.035Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Novita",
      "accuracy": 0.7697014052863436,
      "accuracy_std_dev": 0.07503161624057866,
      "run_count": 7,
      "total_questions": 1362,
      "avg_cost_per_task_usd": 0.22283792731277535,
      "avg_duration_per_task_ms": 1174636.9618208518,
      "avg_output_tokens_per_task": 71500.95080763583,
      "last_run_timestamp": "2026-10-03T16:59:22.739Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7693388734177216,
      "accuracy_std_dev": 0.018315649062573065,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.058541679324894506,
      "avg_duration_per_task_ms": 602157.2911392405,
      "avg_output_tokens_per_task": 26894.77299578059,
      "last_run_timestamp": "2026-08-29T18:26:43.144Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "AkashML",
      "accuracy": 0.7689819826353422,
      "accuracy_std_dev": 0.01778415708685683,
      "run_count": 10,
      "total_questions": 1958,
      "avg_cost_per_task_usd": 0.008032920837589377,
      "avg_duration_per_task_ms": 455602.9090909091,
      "avg_output_tokens_per_task": 27245.283452502554,
      "last_run_timestamp": "2026-10-09T16:20:32.526Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Reka",
      "accuracy": 0.7688279969604862,
      "accuracy_std_dev": 0.009398481531077219,
      "run_count": 10,
      "total_questions": 1974,
      "avg_cost_per_task_usd": 0.019866108409321178,
      "avg_duration_per_task_ms": 104728.69604863222,
      "avg_output_tokens_per_task": 5941.306484295846,
      "last_run_timestamp": "2026-10-03T13:44:15.277Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "CoreWeave",
      "accuracy": 0.7685897384615384,
      "accuracy_std_dev": 0.017564000000000024,
      "run_count": 2,
      "total_questions": 260,
      "avg_cost_per_task_usd": 0.008261646153846154,
      "avg_duration_per_task_ms": 560681.2384615385,
      "avg_output_tokens_per_task": 27419.23076923077,
      "last_run_timestamp": "2026-10-03T16:10:46.858Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Chutes",
      "accuracy": 0.768254,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 105,
      "avg_cost_per_task_usd": 0.04709281904761905,
      "avg_duration_per_task_ms": 1085275.6761904762,
      "avg_output_tokens_per_task": 21896.628571428573,
      "last_run_timestamp": "2026-07-18T16:43:10.588Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Venice",
      "accuracy": 0.767797,
      "accuracy_std_dev": 0.059792957643139584,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.10487614862914862,
      "avg_duration_per_task_ms": 545212.5101010101,
      "avg_output_tokens_per_task": 31481.984126984127,
      "last_run_timestamp": "2026-10-03T16:59:42.233Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7676768,
      "accuracy_std_dev": 0.04513522126410817,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.002215911111111111,
      "avg_duration_per_task_ms": 785689.8707070707,
      "avg_output_tokens_per_task": 15243.155555555555,
      "last_run_timestamp": "2026-08-22T17:47:50.063Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "GMICloud",
      "accuracy": 0.7674760463576159,
      "accuracy_std_dev": 0.05316627521008299,
      "run_count": 7,
      "total_questions": 1359,
      "avg_cost_per_task_usd": 0.22231543340691684,
      "avg_duration_per_task_ms": 898904.8646063282,
      "avg_output_tokens_per_task": 67989.31420161884,
      "last_run_timestamp": "2026-10-03T16:12:13.512Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DekaLLM",
      "accuracy": 0.7673963333333335,
      "accuracy_std_dev": 0.012107287621740685,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.005097089225589225,
      "avg_duration_per_task_ms": 1089299.569023569,
      "avg_output_tokens_per_task": 28258.040404040403,
      "last_run_timestamp": "2026-10-09T03:53:21.415Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "Google",
      "accuracy": 0.7672559999999999,
      "accuracy_std_dev": 0.008469894233892938,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.27037940151515155,
      "avg_duration_per_task_ms": 263755.6675084175,
      "avg_output_tokens_per_task": 17837.74579124579,
      "last_run_timestamp": "2026-09-01T13:23:02.059Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "auto-routing",
      "accuracy": 0.7668349999999999,
      "accuracy_std_dev": 0.01935999999999999,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.2670539015151515,
      "avg_duration_per_task_ms": 177095.6616161616,
      "avg_output_tokens_per_task": 17616.689393939392,
      "last_run_timestamp": "2026-09-01T13:19:59.945Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Crusoe",
      "accuracy": 0.76638602247191,
      "accuracy_std_dev": 0.0119351260850762,
      "run_count": 3,
      "total_questions": 356,
      "avg_cost_per_task_usd": 0.002961258426966292,
      "avg_duration_per_task_ms": 69864.89325842696,
      "avg_output_tokens_per_task": 3595.182584269663,
      "last_run_timestamp": "2026-08-15T14:54:48.654Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.765993,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.08016306060606061,
      "avg_duration_per_task_ms": 425226.22222222225,
      "avg_output_tokens_per_task": 69822.58585858585,
      "last_run_timestamp": "2026-07-18T08:29:23.765Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Venice",
      "accuracy": 0.7647376591639872,
      "accuracy_std_dev": 0.13379493307180026,
      "run_count": 4,
      "total_questions": 622,
      "avg_cost_per_task_usd": 0.01636310610932476,
      "avg_duration_per_task_ms": 580282.652733119,
      "avg_output_tokens_per_task": 22094.519292604502,
      "last_run_timestamp": "2026-09-26T10:43:59.946Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "SambaNova",
      "accuracy": 0.7644980960665658,
      "accuracy_std_dev": 0.006467699256572579,
      "run_count": 7,
      "total_questions": 1322,
      "avg_cost_per_task_usd": 0.024349456127080185,
      "avg_duration_per_task_ms": 70318.21331316188,
      "avg_output_tokens_per_task": 25499.71180030257,
      "last_run_timestamp": "2026-10-03T17:11:16.743Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Parasail",
      "accuracy": 0.7643096666666667,
      "accuracy_std_dev": 0.01620599943093779,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.018970599326599325,
      "avg_duration_per_task_ms": 988034.0404040404,
      "avg_output_tokens_per_task": 47417.47222222222,
      "last_run_timestamp": "2026-10-03T16:16:35.793Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7639796666666666,
      "accuracy_std_dev": 0.03504930162627777,
      "run_count": 7,
      "total_questions": 1377,
      "avg_cost_per_task_usd": 0.11775099782135079,
      "avg_duration_per_task_ms": 799196.2178649238,
      "avg_output_tokens_per_task": 45973.8779956427,
      "last_run_timestamp": "2026-10-03T16:58:32.422Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "DigitalOcean",
      "accuracy": 0.763468,
      "accuracy_std_dev": 0.12590177777802292,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1458534191919192,
      "avg_duration_per_task_ms": 1775663.6776094276,
      "avg_output_tokens_per_task": 49845.639730639734,
      "last_run_timestamp": "2026-10-03T17:21:47.311Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Reka",
      "accuracy": 0.762821,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 78,
      "avg_cost_per_task_usd": 0.02737005128205128,
      "avg_duration_per_task_ms": 183098.3205128205,
      "avg_output_tokens_per_task": 13986.846153846154,
      "last_run_timestamp": "2026-10-08T18:00:04.415Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Io Net",
      "accuracy": 0.7627866948853615,
      "accuracy_std_dev": 0.026695964643035898,
      "run_count": 3,
      "total_questions": 567,
      "avg_cost_per_task_usd": 0.011795488536155203,
      "avg_duration_per_task_ms": 255692.54497354498,
      "avg_output_tokens_per_task": 43172.62081128748,
      "last_run_timestamp": "2026-10-05T10:43:46.241Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.7626260000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.022027883838383836,
      "avg_duration_per_task_ms": 90904.42929292929,
      "avg_output_tokens_per_task": 14522.767676767677,
      "last_run_timestamp": "2026-10-04T03:02:47.748Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.7626259999999999,
      "accuracy_std_dev": 0.01346799999999998,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.01717197474747475,
      "avg_duration_per_task_ms": 622307.4974747475,
      "avg_output_tokens_per_task": 61038.34090909091,
      "last_run_timestamp": "2026-10-03T10:05:09.589Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DeepInfra",
      "accuracy": 0.7613937607313196,
      "accuracy_std_dev": 0.03469481572741633,
      "run_count": 8,
      "total_questions": 1258,
      "avg_cost_per_task_usd": 0.04567820906200318,
      "avg_duration_per_task_ms": 263290.3680445151,
      "avg_output_tokens_per_task": 17206.77265500795,
      "last_run_timestamp": "2026-10-08T22:23:04.239Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "StepFun",
      "accuracy": 0.760943,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.08179280303030302,
      "avg_duration_per_task_ms": 426340.89898989897,
      "avg_output_tokens_per_task": 70994.31818181818,
      "last_run_timestamp": "2026-07-18T08:26:55.806Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7605754720812183,
      "accuracy_std_dev": 0.010447569992586773,
      "run_count": 4,
      "total_questions": 788,
      "avg_cost_per_task_usd": 0.0025078527918781725,
      "avg_duration_per_task_ms": 100160.06345177664,
      "avg_output_tokens_per_task": 31388.34771573604,
      "last_run_timestamp": "2026-10-01T16:04:09.046Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "OpenInference",
      "accuracy": 0.7605443265306122,
      "accuracy_std_dev": 0.0343317834601635,
      "run_count": 3,
      "total_questions": 490,
      "avg_cost_per_task_usd": 0.0037374408163265305,
      "avg_duration_per_task_ms": 668153.0571428571,
      "avg_output_tokens_per_task": 14206.583673469388,
      "last_run_timestamp": "2026-08-08T11:42:26.676Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Groq",
      "accuracy": 0.7597402857142859,
      "accuracy_std_dev": 0.006953716574903084,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.01760951298701299,
      "avg_duration_per_task_ms": 82322.40548340548,
      "avg_output_tokens_per_task": 29108.308802308802,
      "last_run_timestamp": "2026-10-03T16:55:04.036Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "DekaLLM",
      "accuracy": 0.7594230424028269,
      "accuracy_std_dev": 0.004131696745890251,
      "run_count": 3,
      "total_questions": 566,
      "avg_cost_per_task_usd": 0.017543240282685515,
      "avg_duration_per_task_ms": 1026275.9469964664,
      "avg_output_tokens_per_task": 53058.5777385159,
      "last_run_timestamp": "2026-10-09T04:51:43.235Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-small-2603",
      "model_slug": "mistralai/mistral-small-2603",
      "model_name": "Mistral: Mistral Small 4",
      "provider_name": "Venice",
      "accuracy": 0.759259,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.018210444444444442,
      "avg_duration_per_task_ms": 254627.46717171717,
      "avg_output_tokens_per_task": 24079.401515151516,
      "last_run_timestamp": "2026-09-01T12:59:55.618Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7590187142857142,
      "accuracy_std_dev": 0.03157783272538906,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.168772468975469,
      "avg_duration_per_task_ms": 557901.2611832612,
      "avg_output_tokens_per_task": 52553.58585858586,
      "last_run_timestamp": "2026-10-03T17:03:01.289Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Cerebras",
      "accuracy": 0.7590111668383109,
      "accuracy_std_dev": 0.016611473377157127,
      "run_count": 5,
      "total_questions": 971,
      "avg_cost_per_task_usd": 0.004071515962924819,
      "avg_duration_per_task_ms": 4054.4541709577757,
      "avg_output_tokens_per_task": 2252.576725025747,
      "last_run_timestamp": "2026-08-29T18:58:30.919Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.758698,
      "accuracy_std_dev": 0.012009662443216293,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.056970961279461274,
      "avg_duration_per_task_ms": 99388.35016835017,
      "avg_output_tokens_per_task": 22699.07239057239,
      "last_run_timestamp": "2026-10-01T19:11:42.323Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "NextBit",
      "accuracy": 0.7586280000000001,
      "accuracy_std_dev": 0.019486833125215618,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.01705406755050505,
      "avg_duration_per_task_ms": 1145479.2203282828,
      "avg_output_tokens_per_task": 49997.517676767675,
      "last_run_timestamp": "2026-10-03T17:13:23.633Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-small-2603",
      "model_slug": "mistralai/mistral-small-2603",
      "model_name": "Mistral: Mistral Small 4",
      "provider_name": "auto-routing",
      "accuracy": 0.7575759999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.014556641414141414,
      "avg_duration_per_task_ms": 173179.91919191918,
      "avg_output_tokens_per_task": 23980.964646464647,
      "last_run_timestamp": "2026-09-01T13:30:26.223Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Nebius",
      "accuracy": 0.7568542857142856,
      "accuracy_std_dev": 0.010527309976767046,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.018623878066378066,
      "avg_duration_per_task_ms": 169039.05050505052,
      "avg_output_tokens_per_task": 30853.982683982686,
      "last_run_timestamp": "2026-10-03T16:34:22.379Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7565450000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 191,
      "avg_cost_per_task_usd": 0.020916916230366494,
      "avg_duration_per_task_ms": 124793.11518324607,
      "avg_output_tokens_per_task": 18115.47120418848,
      "last_run_timestamp": "2026-07-18T08:26:49.504Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7558922857142859,
      "accuracy_std_dev": 0.012501460717101212,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.018682439754689754,
      "avg_duration_per_task_ms": 169052.6507936508,
      "avg_output_tokens_per_task": 30918.866883116883,
      "last_run_timestamp": "2026-10-03T16:20:04.323Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Near AI",
      "accuracy": 0.755892,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.035218818181818184,
      "avg_duration_per_task_ms": 133646.34848484848,
      "avg_output_tokens_per_task": 10567.11616161616,
      "last_run_timestamp": "2026-10-09T15:00:04.485Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Alibaba",
      "accuracy": 0.7556517142857143,
      "accuracy_std_dev": 0.007605047378725084,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.022500825396825398,
      "avg_duration_per_task_ms": 97229.08225108225,
      "avg_output_tokens_per_task": 14839.808080808081,
      "last_run_timestamp": "2026-10-03T16:56:57.458Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Venice",
      "accuracy": 0.7547599848156182,
      "accuracy_std_dev": 0.05601727479072497,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.0070590672451193055,
      "avg_duration_per_task_ms": 989833.9660159075,
      "avg_output_tokens_per_task": 18416.717281272595,
      "last_run_timestamp": "2026-10-04T00:33:38.079Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Together",
      "accuracy": 0.7547307336244543,
      "accuracy_std_dev": 0.013966941432525313,
      "run_count": 7,
      "total_questions": 1374,
      "avg_cost_per_task_usd": 0.016579247452692868,
      "avg_duration_per_task_ms": 299714.7489082969,
      "avg_output_tokens_per_task": 28149.186317321688,
      "last_run_timestamp": "2026-10-03T17:05:00.424Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7537275714285715,
      "accuracy_std_dev": 0.04841524516353187,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.012184289321789322,
      "avg_duration_per_task_ms": 804903.6428571428,
      "avg_output_tokens_per_task": 36700.15007215007,
      "last_run_timestamp": "2026-10-03T10:19:08.905Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-small-2603",
      "model_slug": "mistralai/mistral-small-2603",
      "model_name": "Mistral: Mistral Small 4",
      "provider_name": "Mistral",
      "accuracy": 0.7528056666666667,
      "accuracy_std_dev": 0.004601995207395269,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.015185508417508417,
      "avg_duration_per_task_ms": 152622.4537037037,
      "avg_output_tokens_per_task": 24359.508417508416,
      "last_run_timestamp": "2026-09-17T02:17:49.284Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.7525250000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008515843434343435,
      "avg_duration_per_task_ms": 110884.25757575757,
      "avg_output_tokens_per_task": 8312.732323232323,
      "last_run_timestamp": "2026-10-04T03:06:11.710Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "DeepInfra",
      "accuracy": 0.7524220592034444,
      "accuracy_std_dev": 0.028639192749568526,
      "run_count": 7,
      "total_questions": 929,
      "avg_cost_per_task_usd": 0.05453179763186222,
      "avg_duration_per_task_ms": 423386.32615715824,
      "avg_output_tokens_per_task": 15836.034445640473,
      "last_run_timestamp": "2026-09-19T08:53:38.031Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Z.AI",
      "accuracy": 0.7509868781725888,
      "accuracy_std_dev": 0.06676355715258636,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.2357046700507614,
      "avg_duration_per_task_ms": 1233772.4551607445,
      "avg_output_tokens_per_task": 75729.32741116751,
      "last_run_timestamp": "2026-10-03T16:45:32.765Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.7503429835390947,
      "accuracy_std_dev": 0.031096298525708804,
      "run_count": 7,
      "total_questions": 1215,
      "avg_cost_per_task_usd": 0.007341672427983538,
      "avg_duration_per_task_ms": 1064868.6510288066,
      "avg_output_tokens_per_task": 18926.89547325103,
      "last_run_timestamp": "2026-10-09T00:38:01.437Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "StreamLake",
      "accuracy": 0.749368625,
      "accuracy_std_dev": 0.008643231874384451,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.013557867424242424,
      "avg_duration_per_task_ms": 96888.43560606061,
      "avg_output_tokens_per_task": 14901.809343434343,
      "last_run_timestamp": "2026-10-03T16:54:39.591Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Xiaomi",
      "accuracy": 0.749158,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.016407893939393937,
      "avg_duration_per_task_ms": 1120280.1313131314,
      "avg_output_tokens_per_task": 58254.84848484849,
      "last_run_timestamp": "2026-10-03T11:59:33.633Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7487311624365482,
      "accuracy_std_dev": 0.019042500000000018,
      "run_count": 2,
      "total_questions": 394,
      "avg_cost_per_task_usd": 0.00260698730964467,
      "avg_duration_per_task_ms": 166191.14467005077,
      "avg_output_tokens_per_task": 30980.81725888325,
      "last_run_timestamp": "2026-10-01T16:04:24.420Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Google",
      "accuracy": 0.7481442884615385,
      "accuracy_std_dev": 0.0227081467548543,
      "run_count": 5,
      "total_questions": 988,
      "avg_cost_per_task_usd": 0.020536520242914982,
      "avg_duration_per_task_ms": 151472.34615384616,
      "avg_output_tokens_per_task": 11878.324898785426,
      "last_run_timestamp": "2026-08-22T17:12:08.919Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Venice",
      "accuracy": 0.7478116,
      "accuracy_std_dev": 0.02151720496347049,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.08178134747474748,
      "avg_duration_per_task_ms": 885057.2202020203,
      "avg_output_tokens_per_task": 23371.92828282828,
      "last_run_timestamp": "2026-08-22T17:38:59.682Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7474749999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.08105034848484848,
      "avg_duration_per_task_ms": 422601,
      "avg_output_tokens_per_task": 70367.68181818182,
      "last_run_timestamp": "2026-07-18T08:27:39.521Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7474749999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.016666974747474746,
      "avg_duration_per_task_ms": 1024552.0252525252,
      "avg_output_tokens_per_task": 59525.38383838384,
      "last_run_timestamp": "2026-10-03T09:35:36.415Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7470487959527824,
      "accuracy_std_dev": 0.08033451694532995,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.005070625070264193,
      "avg_duration_per_task_ms": 656330.5952782462,
      "avg_output_tokens_per_task": 33326.38673412029,
      "last_run_timestamp": "2026-09-26T18:04:35.621Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "auto-routing",
      "accuracy": 0.746633,
      "accuracy_std_dev": 0.017677,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017912724747474747,
      "avg_duration_per_task_ms": 70698.09595959596,
      "avg_output_tokens_per_task": 6119.75,
      "last_run_timestamp": "2026-08-01T17:45:27.258Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Io Net",
      "accuracy": 0.7465941264367815,
      "accuracy_std_dev": 0.10532480903970348,
      "run_count": 8,
      "total_questions": 1566,
      "avg_cost_per_task_usd": 0.04816395977011494,
      "avg_duration_per_task_ms": 468882.4540229885,
      "avg_output_tokens_per_task": 36470.09642401022,
      "last_run_timestamp": "2026-09-19T17:18:00.087Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Morph",
      "accuracy": 0.7465302921840761,
      "accuracy_std_dev": 0.11934042810107118,
      "run_count": 7,
      "total_questions": 1369,
      "avg_cost_per_task_usd": 0.16837692549306063,
      "avg_duration_per_task_ms": 1144637.61577794,
      "avg_output_tokens_per_task": 50455.26296566837,
      "last_run_timestamp": "2026-10-03T18:23:36.655Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "auto-routing",
      "accuracy": 0.74642275,
      "accuracy_std_dev": 0.008945525344411051,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.01488374558080808,
      "avg_duration_per_task_ms": 107961.75883838384,
      "avg_output_tokens_per_task": 14400.517676767677,
      "last_run_timestamp": "2026-10-03T16:43:33.673Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "CoreWeave",
      "accuracy": 0.7461931421319796,
      "accuracy_std_dev": 0.031200414988782142,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.004617404399323182,
      "avg_duration_per_task_ms": 792703.1945854484,
      "avg_output_tokens_per_task": 27906.86040609137,
      "last_run_timestamp": "2026-10-03T17:02:01.121Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "StreamLake",
      "accuracy": 0.7459782222222222,
      "accuracy_std_dev": 0.009438398095931553,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.006886203142536476,
      "avg_duration_per_task_ms": 163678.69079685747,
      "avg_output_tokens_per_task": 8003.4517396184065,
      "last_run_timestamp": "2026-10-03T16:58:32.590Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.7457915,
      "accuracy_std_dev": 0.008417500000000022,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.014859042929292928,
      "avg_duration_per_task_ms": 987926.6818181818,
      "avg_output_tokens_per_task": 52717.79797979798,
      "last_run_timestamp": "2026-10-03T10:50:00.676Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.7457912857142857,
      "accuracy_std_dev": 0.038516164256857485,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.04952885064935065,
      "avg_duration_per_task_ms": 952735.3326118327,
      "avg_output_tokens_per_task": 26072.224386724385,
      "last_run_timestamp": "2026-10-03T17:12:39.572Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Groq",
      "accuracy": 0.7443558079681274,
      "accuracy_std_dev": 0.10340629541961245,
      "run_count": 7,
      "total_questions": 1255,
      "avg_cost_per_task_usd": 0.033485443027888445,
      "avg_duration_per_task_ms": 150958.26772908366,
      "avg_output_tokens_per_task": 29080.37370517928,
      "last_run_timestamp": "2026-10-03T17:57:43.271Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Wafer",
      "accuracy": 0.744108,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.012609747474747475,
      "avg_duration_per_task_ms": 233838.9696969697,
      "avg_output_tokens_per_task": 62878.5101010101,
      "last_run_timestamp": "2026-10-09T00:00:06.206Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7436087853982302,
      "accuracy_std_dev": 0.07436948975876863,
      "run_count": 7,
      "total_questions": 1356,
      "avg_cost_per_task_usd": 0.13460593289085546,
      "avg_duration_per_task_ms": 1037566.0044247787,
      "avg_output_tokens_per_task": 57351.7883480826,
      "last_run_timestamp": "2026-10-03T16:56:39.926Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Parasail",
      "accuracy": 0.7434143503039515,
      "accuracy_std_dev": 0.004376274946591358,
      "run_count": 7,
      "total_questions": 1316,
      "avg_cost_per_task_usd": 0.009106649696048632,
      "avg_duration_per_task_ms": 140201.952887538,
      "avg_output_tokens_per_task": 11539.308510638299,
      "last_run_timestamp": "2026-10-03T17:03:40.665Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Nebius",
      "accuracy": 0.7433862857142857,
      "accuracy_std_dev": 0.006785370391285875,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.005382690476190476,
      "avg_duration_per_task_ms": 149045.53823953823,
      "avg_output_tokens_per_task": 8700.336940836942,
      "last_run_timestamp": "2026-10-03T17:05:50.656Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DigitalOcean",
      "accuracy": 0.743266,
      "accuracy_std_dev": 0.09234344800932008,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.029721953914141414,
      "avg_duration_per_task_ms": 1313239.2840909092,
      "avg_output_tokens_per_task": 18789.802398989897,
      "last_run_timestamp": "2026-10-05T21:43:30.018Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Novita",
      "accuracy": 0.743145857142857,
      "accuracy_std_dev": 0.011088924595914524,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.023026453102453104,
      "avg_duration_per_task_ms": 99585.75685425685,
      "avg_output_tokens_per_task": 15242.642857142857,
      "last_run_timestamp": "2026-10-03T16:58:08.462Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Phala",
      "accuracy": 0.74242425,
      "accuracy_std_dev": 0.08095717220967824,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.017491213383838383,
      "avg_duration_per_task_ms": 2247201.621212121,
      "avg_output_tokens_per_task": 17091.73232323232,
      "last_run_timestamp": "2026-10-04T01:00:47.859Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "Groq",
      "accuracy": 0.7415824999999999,
      "accuracy_std_dev": 0.005892500000000023,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.019016893939393938,
      "avg_duration_per_task_ms": 33403.05808080808,
      "avg_output_tokens_per_task": 6109.055555555556,
      "last_run_timestamp": "2026-08-01T17:45:49.025Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Morph",
      "accuracy": 0.740741,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02792859090909091,
      "avg_duration_per_task_ms": 682232.3585858586,
      "avg_output_tokens_per_task": 23364.934343434343,
      "last_run_timestamp": "2026-07-18T09:15:42.531Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "GMICloud",
      "accuracy": 0.7406781050847459,
      "accuracy_std_dev": 0.10340659429589846,
      "run_count": 9,
      "total_questions": 1770,
      "avg_cost_per_task_usd": 0.006412678531073445,
      "avg_duration_per_task_ms": 534746.5333333333,
      "avg_output_tokens_per_task": 27485.332768361583,
      "last_run_timestamp": "2026-10-03T10:03:14.593Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.739899,
      "accuracy_std_dev": 0.010812416138403098,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0048791994949494944,
      "avg_duration_per_task_ms": 108739.66098484848,
      "avg_output_tokens_per_task": 7956.729797979798,
      "last_run_timestamp": "2026-10-03T16:59:36.480Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Friendli",
      "accuracy": 0.7395381428571429,
      "accuracy_std_dev": 0.07707797586939116,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0066781443001443,
      "avg_duration_per_task_ms": 95940.86147186147,
      "avg_output_tokens_per_task": 16425.329725829724,
      "last_run_timestamp": "2026-10-03T16:54:56.858Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "NextBit",
      "accuracy": 0.7392292244897959,
      "accuracy_std_dev": 0.110097,
      "run_count": 2,
      "total_questions": 294,
      "avg_cost_per_task_usd": 0.03822905782312925,
      "avg_duration_per_task_ms": 1403982.7517006802,
      "avg_output_tokens_per_task": 27800.656462585033,
      "last_run_timestamp": "2026-08-01T23:27:37.975Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Mara",
      "accuracy": 0.7387387252252252,
      "accuracy_std_dev": 0.018499983466208795,
      "run_count": 4,
      "total_questions": 444,
      "avg_cost_per_task_usd": 0.01493145045045045,
      "avg_duration_per_task_ms": 79529.70945945945,
      "avg_output_tokens_per_task": 8581.265765765766,
      "last_run_timestamp": "2026-10-04T00:00:36.943Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Cloudflare",
      "accuracy": 0.7387387027027028,
      "accuracy_std_dev": 0.06584221235841033,
      "run_count": 6,
      "total_questions": 1184,
      "avg_cost_per_task_usd": 0.010049336993243244,
      "avg_duration_per_task_ms": 495836.71199324325,
      "avg_output_tokens_per_task": 34244.69847972973,
      "last_run_timestamp": "2026-10-03T17:30:20.154Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.7385762857142857,
      "accuracy_std_dev": 0.1483657159518949,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.030072527417027418,
      "avg_duration_per_task_ms": 138129.5901875902,
      "avg_output_tokens_per_task": 23037.04256854257,
      "last_run_timestamp": "2026-10-03T16:48:25.700Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "Google",
      "accuracy": 0.738309,
      "accuracy_std_dev": 0.008136215650479735,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.059655349046015715,
      "avg_duration_per_task_ms": 125601.6632996633,
      "avg_output_tokens_per_task": 23853.966891133558,
      "last_run_timestamp": "2026-10-01T19:15:31.968Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Parasail",
      "accuracy": 0.7378293623304071,
      "accuracy_std_dev": 0.0050744218168684305,
      "run_count": 7,
      "total_questions": 1253,
      "avg_cost_per_task_usd": 0.0064119712689545095,
      "avg_duration_per_task_ms": 226069.1827613727,
      "avg_output_tokens_per_task": 7932.774142059058,
      "last_run_timestamp": "2026-10-03T16:52:01.284Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "auto-routing",
      "accuracy": 0.736652142857143,
      "accuracy_std_dev": 0.05013373986342806,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.016218962481962484,
      "avg_duration_per_task_ms": 475113.9826839827,
      "avg_output_tokens_per_task": 15698.901154401154,
      "last_run_timestamp": "2026-10-04T00:29:23.413Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "Ambient",
      "accuracy": 0.736111,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 168,
      "avg_cost_per_task_usd": 0.05987852380952381,
      "avg_duration_per_task_ms": 1235915.3511904762,
      "avg_output_tokens_per_task": 62102.07738095238,
      "last_run_timestamp": "2026-07-18T08:29:05.636Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Phala",
      "accuracy": 0.7344877142857142,
      "accuracy_std_dev": 0.06178190798217846,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.011764064935064936,
      "avg_duration_per_task_ms": 95375.22294372294,
      "avg_output_tokens_per_task": 5009.694083694084,
      "last_run_timestamp": "2026-10-03T10:46:59.690Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Novita",
      "accuracy": 0.7322156158536586,
      "accuracy_std_dev": 0.0628101300320259,
      "run_count": 5,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.015606842479674796,
      "avg_duration_per_task_ms": 1212509.0548780488,
      "avg_output_tokens_per_task": 39797.57621951219,
      "last_run_timestamp": "2026-10-03T17:20:21.640Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Novita",
      "accuracy": 0.7316240615384616,
      "accuracy_std_dev": 0.012678999999999996,
      "run_count": 2,
      "total_questions": 390,
      "avg_cost_per_task_usd": 0.014848610256410256,
      "avg_duration_per_task_ms": 978212.5923076923,
      "avg_output_tokens_per_task": 52658.97692307692,
      "last_run_timestamp": "2026-10-03T10:14:05.783Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7313943448275863,
      "accuracy_std_dev": 0.013515629737139094,
      "run_count": 29,
      "total_questions": 5742,
      "avg_cost_per_task_usd": 0.2044269764890282,
      "avg_duration_per_task_ms": 243437.65552072448,
      "avg_output_tokens_per_task": 38184.69731800766,
      "last_run_timestamp": "2026-10-03T09:50:25.399Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Google",
      "accuracy": 0.7300516810519218,
      "accuracy_std_dev": 0.016762033534018325,
      "run_count": 15,
      "total_questions": 2966,
      "avg_cost_per_task_usd": 0.0067854190829399865,
      "avg_duration_per_task_ms": 104234.09912339851,
      "avg_output_tokens_per_task": 7076.775792312879,
      "last_run_timestamp": "2026-10-03T16:40:17.848Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7297978999999999,
      "accuracy_std_dev": 0.016326424363282986,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.18559751616161616,
      "avg_duration_per_task_ms": 291435.0106060606,
      "avg_output_tokens_per_task": 36955.371717171714,
      "last_run_timestamp": "2026-10-03T09:40:02.418Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7286196,
      "accuracy_std_dev": 0.010123398502479267,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.19021829797979797,
      "avg_duration_per_task_ms": 241955.48484848486,
      "avg_output_tokens_per_task": 37852.89646464647,
      "last_run_timestamp": "2026-10-03T09:25:15.662Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.7277538571428572,
      "accuracy_std_dev": 0.020296230068157783,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.005383534632034633,
      "avg_duration_per_task_ms": 215362.52525252526,
      "avg_output_tokens_per_task": 7827.647907647907,
      "last_run_timestamp": "2026-10-03T17:25:00.727Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Novita",
      "accuracy": 0.7272725714285714,
      "accuracy_std_dev": 0.015455781912624552,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.004581955266955267,
      "avg_duration_per_task_ms": 160113.02813852814,
      "avg_output_tokens_per_task": 7774.143578643579,
      "last_run_timestamp": "2026-10-03T16:57:10.991Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Venice",
      "accuracy": 0.726431,
      "accuracy_std_dev": 0.012012832596852407,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.00610699074074074,
      "avg_duration_per_task_ms": 365363.45791245793,
      "avg_output_tokens_per_task": 8100.830808080808,
      "last_run_timestamp": "2026-10-03T17:45:28.847Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Tenstorrent",
      "accuracy": 0.7264309999999999,
      "accuracy_std_dev": 0.021044000000000007,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.1303771111111111,
      "avg_duration_per_task_ms": 318087.7601010101,
      "avg_output_tokens_per_task": 43267.416666666664,
      "last_run_timestamp": "2026-10-04T05:57:29.940Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7261503333333333,
      "accuracy_std_dev": 0.004418546090690373,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.059926168350168355,
      "avg_duration_per_task_ms": 123042.70033670033,
      "avg_output_tokens_per_task": 23940.483164983165,
      "last_run_timestamp": "2026-10-01T19:12:41.089Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "DeepInfra",
      "accuracy": 0.725799625,
      "accuracy_std_dev": 0.008125736488735932,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.004432963383838384,
      "avg_duration_per_task_ms": 640133.6079545454,
      "avg_output_tokens_per_task": 7937.186237373738,
      "last_run_timestamp": "2026-10-09T00:02:28.846Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "auto-routing",
      "accuracy": 0.7255892857142857,
      "accuracy_std_dev": 0.030568975597183873,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.007296898268398268,
      "avg_duration_per_task_ms": 250965.08297258298,
      "avg_output_tokens_per_task": 16490.68903318903,
      "last_run_timestamp": "2026-10-03T17:08:10.486Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Azure",
      "accuracy": 0.7234887554404145,
      "accuracy_std_dev": 0.009381891562472875,
      "run_count": 10,
      "total_questions": 1930,
      "avg_cost_per_task_usd": 0.18696787461139894,
      "avg_duration_per_task_ms": 349087.42279792746,
      "avg_output_tokens_per_task": 37203.45544041451,
      "last_run_timestamp": "2026-10-03T09:27:16.866Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Krea",
      "accuracy": 0.7234044255319149,
      "accuracy_std_dev": 0.1917994784299478,
      "run_count": 3,
      "total_questions": 517,
      "avg_cost_per_task_usd": 0.033565591876208896,
      "avg_duration_per_task_ms": 242772.43907156674,
      "avg_output_tokens_per_task": 37102.30560928433,
      "last_run_timestamp": "2026-09-29T20:57:20.673Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Google",
      "accuracy": 0.7225028666666667,
      "accuracy_std_dev": 0.012643758973589382,
      "run_count": 30,
      "total_questions": 5940,
      "avg_cost_per_task_usd": 0.1973207321548822,
      "avg_duration_per_task_ms": 263893.9067340067,
      "avg_output_tokens_per_task": 36800.280303030304,
      "last_run_timestamp": "2026-10-03T09:27:45.066Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Novita",
      "accuracy": 0.7222221428571429,
      "accuracy_std_dev": 0.020500479544416028,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.006640360028860029,
      "avg_duration_per_task_ms": 334191.47041847045,
      "avg_output_tokens_per_task": 26367.450937950936,
      "last_run_timestamp": "2026-10-03T17:40:21.277Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Nebius",
      "accuracy": 0.7220491744548286,
      "accuracy_std_dev": 0.030848909598882115,
      "run_count": 5,
      "total_questions": 963,
      "avg_cost_per_task_usd": 0.16807563966770508,
      "avg_duration_per_task_ms": 1503192.1879543094,
      "avg_output_tokens_per_task": 41124.30737279335,
      "last_run_timestamp": "2026-08-22T16:49:30.177Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "ModelRun",
      "accuracy": 0.7219408075949366,
      "accuracy_std_dev": 0.025598999999999983,
      "run_count": 2,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.27939689620253166,
      "avg_duration_per_task_ms": 450249.90379746835,
      "avg_output_tokens_per_task": 45941.00253164557,
      "last_run_timestamp": "2026-08-15T08:16:15.061Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "auto-routing",
      "accuracy": 0.7211000000000001,
      "accuracy_std_dev": 0.05129065401220774,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.015817102693602692,
      "avg_duration_per_task_ms": 896061.7971380472,
      "avg_output_tokens_per_task": 43246.910774410775,
      "last_run_timestamp": "2026-10-03T18:22:40.169Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Phala",
      "accuracy": 0.71923592494929,
      "accuracy_std_dev": 0.031713523289599956,
      "run_count": 5,
      "total_questions": 986,
      "avg_cost_per_task_usd": 0.01125353448275862,
      "avg_duration_per_task_ms": 528851.3123732251,
      "avg_output_tokens_per_task": 56006.78093306288,
      "last_run_timestamp": "2026-10-03T09:18:14.437Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7184345000000002,
      "accuracy_std_dev": 0.00575544305418794,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.006676434343434343,
      "avg_duration_per_task_ms": 1198975.5896464647,
      "avg_output_tokens_per_task": 54074.54671717172,
      "last_run_timestamp": "2026-09-26T17:35:27.801Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Friendli",
      "accuracy": 0.7184345000000001,
      "accuracy_std_dev": 0.002104500000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.005931222222222222,
      "avg_duration_per_task_ms": 277051.3181818182,
      "avg_output_tokens_per_task": 7213.616161616162,
      "last_run_timestamp": "2026-08-01T17:52:16.799Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Venice",
      "accuracy": 0.7171302234809476,
      "accuracy_std_dev": 0.05915089253318348,
      "run_count": 6,
      "total_questions": 971,
      "avg_cost_per_task_usd": 0.035917489186405756,
      "avg_duration_per_task_ms": 751381.8146240988,
      "avg_output_tokens_per_task": 38973.5623069001,
      "last_run_timestamp": "2026-08-29T18:51:07.939Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "GMICloud",
      "accuracy": 0.7170402190889371,
      "accuracy_std_dev": 0.1172294876991454,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.004009827187274042,
      "avg_duration_per_task_ms": 280047.9363702097,
      "avg_output_tokens_per_task": 10071.284887924801,
      "last_run_timestamp": "2026-10-04T00:01:24.489Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "AtlasCloud",
      "accuracy": 0.715769,
      "accuracy_std_dev": 0.047798870551509894,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.00680639898989899,
      "avg_duration_per_task_ms": 127099.18602693603,
      "avg_output_tokens_per_task": 7554.406565656565,
      "last_run_timestamp": "2026-08-29T18:01:46.940Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "WandB",
      "accuracy": 0.715488,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0037085252525252527,
      "avg_duration_per_task_ms": 839996.601010101,
      "avg_output_tokens_per_task": 28561.196969696968,
      "last_run_timestamp": "2026-07-18T16:28:34.215Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Makora",
      "accuracy": 0.715488,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.018400479797979798,
      "avg_duration_per_task_ms": 232169.87878787878,
      "avg_output_tokens_per_task": 65366.41414141414,
      "last_run_timestamp": "2026-10-07T21:00:03.225Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Io Net",
      "accuracy": 0.7133716737044147,
      "accuracy_std_dev": 0.07918700388461619,
      "run_count": 3,
      "total_questions": 521,
      "avg_cost_per_task_usd": 0.07439642610364683,
      "avg_duration_per_task_ms": 652167.8637236084,
      "avg_output_tokens_per_task": 28734.940499040305,
      "last_run_timestamp": "2026-08-15T09:29:41.209Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Darkbloom",
      "accuracy": 0.713213,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 111,
      "avg_cost_per_task_usd": 0.007665369369369369,
      "avg_duration_per_task_ms": 1401051.936936937,
      "avg_output_tokens_per_task": 27183.8018018018,
      "last_run_timestamp": "2026-10-03T09:34:00.571Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.7118807142857142,
      "accuracy_std_dev": 0.008414141170235586,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.011084775613275614,
      "avg_duration_per_task_ms": 94339.78354978355,
      "avg_output_tokens_per_task": 10485.185425685426,
      "last_run_timestamp": "2026-10-03T16:59:33.949Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Novita",
      "accuracy": 0.71170025,
      "accuracy_std_dev": 0.018456257615982147,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0649322058080808,
      "avg_duration_per_task_ms": 1066951.6603535353,
      "avg_output_tokens_per_task": 21770.531565656565,
      "last_run_timestamp": "2026-08-22T17:45:06.645Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.7113995714285716,
      "accuracy_std_dev": 0.010488730194930244,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.007985493506493507,
      "avg_duration_per_task_ms": 77746.15223665224,
      "avg_output_tokens_per_task": 10141.235930735931,
      "last_run_timestamp": "2026-10-03T16:40:12.993Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.7109188571428572,
      "accuracy_std_dev": 0.009120000930898651,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.008814419191919191,
      "avg_duration_per_task_ms": 134404.79509379508,
      "avg_output_tokens_per_task": 5713.798701298701,
      "last_run_timestamp": "2026-10-03T16:45:33.688Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.7106782857142858,
      "accuracy_std_dev": 0.008509694097478088,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.015819831890331888,
      "avg_duration_per_task_ms": 94996.398989899,
      "avg_output_tokens_per_task": 10465.311688311689,
      "last_run_timestamp": "2026-10-03T16:46:22.205Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "CoreWeave",
      "accuracy": 0.70883299486008,
      "accuracy_std_dev": 0.036241367349453366,
      "run_count": 9,
      "total_questions": 1751,
      "avg_cost_per_task_usd": 0.010877087378640777,
      "avg_duration_per_task_ms": 281941.7778412336,
      "avg_output_tokens_per_task": 47824.729868646486,
      "last_run_timestamp": "2026-10-03T09:36:23.802Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Mancer 2",
      "accuracy": 0.7083335000000002,
      "accuracy_std_dev": 0.053758401224831845,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.05601492550505051,
      "avg_duration_per_task_ms": 1569531.130050505,
      "avg_output_tokens_per_task": 33285.93813131313,
      "last_run_timestamp": "2026-10-04T00:00:36.972Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.7080899875730994,
      "accuracy_std_dev": 0.01572159476124274,
      "run_count": 7,
      "total_questions": 1368,
      "avg_cost_per_task_usd": 0.010683734649122806,
      "avg_duration_per_task_ms": 102665.67105263157,
      "avg_output_tokens_per_task": 9845.541666666666,
      "last_run_timestamp": "2026-10-03T16:43:17.105Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Io Net",
      "accuracy": 0.7078187849794239,
      "accuracy_std_dev": 0.11990455329935001,
      "run_count": 6,
      "total_questions": 972,
      "avg_cost_per_task_usd": 0.004627639917695473,
      "avg_duration_per_task_ms": 565545.3467078189,
      "avg_output_tokens_per_task": 25055.12037037037,
      "last_run_timestamp": "2026-10-05T10:41:53.582Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.707071,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008867828282828283,
      "avg_duration_per_task_ms": 116045.65151515152,
      "avg_output_tokens_per_task": 5749.39898989899,
      "last_run_timestamp": "2026-10-04T03:05:59.018Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.7051465714285715,
      "accuracy_std_dev": 0.008927717479163138,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.006021121212121212,
      "avg_duration_per_task_ms": 109220.25974025975,
      "avg_output_tokens_per_task": 5588.1832611832615,
      "last_run_timestamp": "2026-10-03T17:02:24.058Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "auto-routing",
      "accuracy": 0.7045455,
      "accuracy_std_dev": 0.004208500000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.01336774494949495,
      "avg_duration_per_task_ms": 231791.09848484848,
      "avg_output_tokens_per_task": 33323.15909090909,
      "last_run_timestamp": "2026-10-01T19:27:22.032Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "OpenInference",
      "accuracy": 0.70328275,
      "accuracy_std_dev": 0.02153875190644758,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0019388005050505052,
      "avg_duration_per_task_ms": 636092.8409090909,
      "avg_output_tokens_per_task": 14000.35101010101,
      "last_run_timestamp": "2026-09-05T08:28:07.290Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "GMICloud",
      "accuracy": 0.7028622499999999,
      "accuracy_std_dev": 0.08200476941122087,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0070032777777777775,
      "avg_duration_per_task_ms": 172461.12373737374,
      "avg_output_tokens_per_task": 7212.60101010101,
      "last_run_timestamp": "2026-10-03T16:56:36.098Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.7024411666666667,
      "accuracy_std_dev": 0.14218853509503576,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.013173934343434343,
      "avg_duration_per_task_ms": 786438.8409090909,
      "avg_output_tokens_per_task": 27569.127104377105,
      "last_run_timestamp": "2026-09-22T02:20:03.900Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DeepInfra",
      "accuracy": 0.7021902115384616,
      "accuracy_std_dev": 0.04770844742400843,
      "run_count": 16,
      "total_questions": 3120,
      "avg_cost_per_task_usd": 0.00835266217948718,
      "avg_duration_per_task_ms": 486695.7522435897,
      "avg_output_tokens_per_task": 25245.021153846155,
      "last_run_timestamp": "2026-10-09T00:36:36.941Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.701832040429564,
      "accuracy_std_dev": 0.009254626092359155,
      "run_count": 8,
      "total_questions": 1583,
      "avg_cost_per_task_usd": 0.010471590018951359,
      "avg_duration_per_task_ms": 144460.25710675932,
      "avg_output_tokens_per_task": 9453.23373341756,
      "last_run_timestamp": "2026-10-08T22:40:37.311Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.7017798571428572,
      "accuracy_std_dev": 0.013719199460490942,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.011180141414141413,
      "avg_duration_per_task_ms": 199426.63852813852,
      "avg_output_tokens_per_task": 5820.556277056277,
      "last_run_timestamp": "2026-10-03T16:59:57.847Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Google",
      "accuracy": 0.7010584285714286,
      "accuracy_std_dev": 0.024870525026666448,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.011166472582972582,
      "avg_duration_per_task_ms": 61418.856421356424,
      "avg_output_tokens_per_task": 9244.735930735931,
      "last_run_timestamp": "2026-10-03T16:36:16.895Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "SambaNova",
      "accuracy": 0.6952191613545817,
      "accuracy_std_dev": 0.030707686936010004,
      "run_count": 5,
      "total_questions": 502,
      "avg_cost_per_task_usd": 0.009013709163346614,
      "avg_duration_per_task_ms": 48627.035856573704,
      "avg_output_tokens_per_task": 5832.633466135459,
      "last_run_timestamp": "2026-10-04T00:12:48.620Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6948820649606299,
      "accuracy_std_dev": 0.027729042227588305,
      "run_count": 8,
      "total_questions": 1524,
      "avg_cost_per_task_usd": 0.015564356299212597,
      "avg_duration_per_task_ms": 765563.3963254594,
      "avg_output_tokens_per_task": 77634.76837270342,
      "last_run_timestamp": "2026-09-26T16:34:21.219Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.6928812857142858,
      "accuracy_std_dev": 0.0184605601038251,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.009158278499278501,
      "avg_duration_per_task_ms": 267700.48124098126,
      "avg_output_tokens_per_task": 6402.472582972583,
      "last_run_timestamp": "2026-10-03T17:02:48.284Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "auto-routing",
      "accuracy": 0.6926927837837837,
      "accuracy_std_dev": 0.02593753677273029,
      "run_count": 7,
      "total_questions": 1332,
      "avg_cost_per_task_usd": 0.009898264264264264,
      "avg_duration_per_task_ms": 376272.99399399396,
      "avg_output_tokens_per_task": 45270.37087087087,
      "last_run_timestamp": "2026-10-03T09:53:13.324Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.692129625,
      "accuracy_std_dev": 0.010010686751385976,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.005500493055555555,
      "avg_duration_per_task_ms": 380445.4021464646,
      "avg_output_tokens_per_task": 6066.023358585859,
      "last_run_timestamp": "2026-10-08T23:09:15.796Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Wafer",
      "accuracy": 0.691919,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01248349494949495,
      "avg_duration_per_task_ms": 400806.7474747475,
      "avg_output_tokens_per_task": 31001.929292929293,
      "last_run_timestamp": "2026-07-18T08:04:42.524Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Google",
      "accuracy": 0.6917878385615913,
      "accuracy_std_dev": 0.2742550167259424,
      "run_count": 7,
      "total_questions": 1307,
      "avg_cost_per_task_usd": 0.02420585692425402,
      "avg_duration_per_task_ms": 312930.753634277,
      "avg_output_tokens_per_task": 14561.572302983932,
      "last_run_timestamp": "2026-10-04T00:21:14.629Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "Azure",
      "accuracy": 0.689815,
      "accuracy_std_dev": 0.0034448118235979193,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.013852137626262627,
      "avg_duration_per_task_ms": 234308.42045454544,
      "avg_output_tokens_per_task": 32893.11994949495,
      "last_run_timestamp": "2026-10-01T19:27:10.142Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Io Net",
      "accuracy": 0.685185,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.007298616161616161,
      "avg_duration_per_task_ms": 578052.9191919192,
      "avg_output_tokens_per_task": 17923.358585858587,
      "last_run_timestamp": "2026-07-18T08:44:31.047Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Phala",
      "accuracy": 0.685185,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.06544819696969698,
      "avg_duration_per_task_ms": 945099.904040404,
      "avg_output_tokens_per_task": 33006.14646464647,
      "last_run_timestamp": "2026-07-18T16:15:33.587Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403:free",
      "model_slug": null,
      "model_name": "Gemma 4 26B A4B IT (free)",
      "provider_name": "Darkbloom",
      "accuracy": 0.684722,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 120,
      "avg_cost_per_task_usd": null,
      "avg_duration_per_task_ms": 382535.98333333334,
      "avg_output_tokens_per_task": 6994.291666666667,
      "last_run_timestamp": "2026-08-04T05:36:23.551Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.6843435000000001,
      "accuracy_std_dev": 0.005892499999999967,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.01356670707070707,
      "avg_duration_per_task_ms": 229761.702020202,
      "avg_output_tokens_per_task": 33820.56313131313,
      "last_run_timestamp": "2026-10-01T19:28:34.261Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "DeepInfra",
      "accuracy": 0.68308075,
      "accuracy_std_dev": 0.050431518420403505,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.007981098484848485,
      "avg_duration_per_task_ms": 281140.702020202,
      "avg_output_tokens_per_task": 41483.12815656565,
      "last_run_timestamp": "2026-10-08T22:19:02.584Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Venice",
      "accuracy": 0.6809765714285714,
      "accuracy_std_dev": 0.0445115533536989,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.014742932178932178,
      "avg_duration_per_task_ms": 268442.60533910536,
      "avg_output_tokens_per_task": 7991.235930735931,
      "last_run_timestamp": "2026-10-03T17:36:48.437Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Phala",
      "accuracy": 0.6782486254237288,
      "accuracy_std_dev": 0.021428719439237503,
      "run_count": 6,
      "total_questions": 590,
      "avg_cost_per_task_usd": 0.004103340677966102,
      "avg_duration_per_task_ms": 494419.3559322034,
      "avg_output_tokens_per_task": 27366.218644067798,
      "last_run_timestamp": "2026-08-29T17:24:11.557Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Parasail",
      "accuracy": 0.6779694919852034,
      "accuracy_std_dev": 0.0848101181666433,
      "run_count": 5,
      "total_questions": 811,
      "avg_cost_per_task_usd": 0.007035279901356351,
      "avg_duration_per_task_ms": 476700.88409371144,
      "avg_output_tokens_per_task": 34252.66584463625,
      "last_run_timestamp": "2026-08-29T12:30:09.617Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Morph",
      "accuracy": 0.676784615514334,
      "accuracy_std_dev": 0.11674700703943836,
      "run_count": 3,
      "total_questions": 593,
      "avg_cost_per_task_usd": 0.05877558516020236,
      "avg_duration_per_task_ms": 524798.5919055649,
      "avg_output_tokens_per_task": 24486.29679595278,
      "last_run_timestamp": "2026-08-15T10:07:37.662Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "OpenInference",
      "accuracy": 0.6750842,
      "accuracy_std_dev": 0.023641210809939493,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.005652539393939394,
      "avg_duration_per_task_ms": 500294.62828282826,
      "avg_output_tokens_per_task": 16033.422222222222,
      "last_run_timestamp": "2026-08-29T18:40:21.422Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "GMICloud",
      "accuracy": 0.6737955840707965,
      "accuracy_std_dev": 0.07392353888102138,
      "run_count": 7,
      "total_questions": 1356,
      "avg_cost_per_task_usd": 0.1320932337758112,
      "avg_duration_per_task_ms": 1052059.2595870206,
      "avg_output_tokens_per_task": 71927.51327433628,
      "last_run_timestamp": "2026-10-03T16:56:56.694Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Venice",
      "accuracy": 0.6728395,
      "accuracy_std_dev": 0.030761988442177545,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.005704560606060605,
      "avg_duration_per_task_ms": 413488.24074074073,
      "avg_output_tokens_per_task": 15636.112794612794,
      "last_run_timestamp": "2026-10-03T17:31:07.224Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "StreamLake",
      "accuracy": 0.6689113333333334,
      "accuracy_std_dev": 0.0067804647005611455,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0026409175084175086,
      "avg_duration_per_task_ms": 47935.90572390572,
      "avg_output_tokens_per_task": 2383.6902356902356,
      "last_run_timestamp": "2026-10-02T15:53:31.869Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6-20251208",
      "model_slug": "z-ai/glm-4.6v",
      "model_name": "Z.ai: GLM 4.6V",
      "provider_name": "auto-routing",
      "accuracy": 0.66835,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.013946772727272728,
      "avg_duration_per_task_ms": 324889.0353535354,
      "avg_output_tokens_per_task": 15315.242424242424,
      "last_run_timestamp": "2026-10-01T08:41:30.639Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "AkashML",
      "accuracy": 0.6679221148775893,
      "accuracy_std_dev": 0.015946572790414858,
      "run_count": 5,
      "total_questions": 531,
      "avg_cost_per_task_usd": 0.004098755178907722,
      "avg_duration_per_task_ms": 736779.5291902071,
      "avg_output_tokens_per_task": 40851.82109227872,
      "last_run_timestamp": "2026-10-09T16:25:16.151Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.6678159586206897,
      "accuracy_std_dev": 0.07976717100897782,
      "run_count": 7,
      "total_questions": 1015,
      "avg_cost_per_task_usd": 0.04199740492610838,
      "avg_duration_per_task_ms": 264794.87290640397,
      "avg_output_tokens_per_task": 15915.662068965517,
      "last_run_timestamp": "2026-10-08T22:52:37.590Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6673423259988418,
      "accuracy_std_dev": 0.09700477765553617,
      "run_count": 10,
      "total_questions": 1727,
      "avg_cost_per_task_usd": 0.030584045165026053,
      "avg_duration_per_task_ms": 798364.2964678634,
      "avg_output_tokens_per_task": 32108.440069484655,
      "last_run_timestamp": "2026-10-09T00:06:52.220Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "WandB",
      "accuracy": 0.666667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.009070929292929292,
      "avg_duration_per_task_ms": 492518.0909090909,
      "avg_output_tokens_per_task": 69555.08585858585,
      "last_run_timestamp": "2026-07-18T17:04:25.343Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.666667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.002231131313131313,
      "avg_duration_per_task_ms": 1386615.121212121,
      "avg_output_tokens_per_task": 14198.39898989899,
      "last_run_timestamp": "2026-07-18T09:58:41.490Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DeepInfra",
      "accuracy": 0.6654557258916777,
      "accuracy_std_dev": 0.015299762836703567,
      "run_count": 8,
      "total_questions": 1514,
      "avg_cost_per_task_usd": 0.007279266182298546,
      "avg_duration_per_task_ms": 504110.9504623514,
      "avg_output_tokens_per_task": 51807.61756935271,
      "last_run_timestamp": "2026-10-08T23:53:57.221Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6654041250000001,
      "accuracy_std_dev": 0.0318059948022283,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0025243055555555557,
      "avg_duration_per_task_ms": 211186.44002525252,
      "avg_output_tokens_per_task": 5302.534090909091,
      "last_run_timestamp": "2026-10-03T16:12:07.496Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Mancer 2",
      "accuracy": 0.6647026666666667,
      "accuracy_std_dev": 0.07359682177392055,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.006663367845117845,
      "avg_duration_per_task_ms": 681258.2592592592,
      "avg_output_tokens_per_task": 22537.15824915825,
      "last_run_timestamp": "2026-10-03T17:57:14.269Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6-20251208",
      "model_slug": "z-ai/glm-4.6v",
      "model_name": "Z.ai: GLM 4.6V",
      "provider_name": "Novita",
      "accuracy": 0.6633000000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.013461267676767677,
      "avg_duration_per_task_ms": 327894.71717171714,
      "avg_output_tokens_per_task": 14738.742424242424,
      "last_run_timestamp": "2026-10-01T08:27:45.371Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6632997142857142,
      "accuracy_std_dev": 0.009351528808768375,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0021804473304473304,
      "avg_duration_per_task_ms": 240867.02741702742,
      "avg_output_tokens_per_task": 3011.5793650793653,
      "last_run_timestamp": "2026-10-05T20:44:48.611Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "auto-routing",
      "accuracy": 0.663019,
      "accuracy_std_dev": 0.007921062218330754,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0027949452861952863,
      "avg_duration_per_task_ms": 83787.81902356903,
      "avg_output_tokens_per_task": 2948.4612794612794,
      "last_run_timestamp": "2026-08-29T19:49:56.943Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.6616161818181817,
      "accuracy_std_dev": 0.013544295262167041,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.005160062442607897,
      "avg_duration_per_task_ms": 106935.65518824609,
      "avg_output_tokens_per_task": 9720.085399449035,
      "last_run_timestamp": "2026-10-03T09:23:08.256Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "Novita",
      "accuracy": 0.661616,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.003362974747474748,
      "avg_duration_per_task_ms": 65871.15656565657,
      "avg_output_tokens_per_task": 2362.121212121212,
      "last_run_timestamp": "2026-08-01T09:30:53.884Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "OpenAI",
      "accuracy": 0.6607744999999999,
      "accuracy_std_dev": 0.011572094721786573,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.015565636363636365,
      "avg_duration_per_task_ms": 14159.460858585859,
      "avg_output_tokens_per_task": 1751.3977272727273,
      "last_run_timestamp": "2026-10-02T23:17:48.072Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6602454107648725,
      "accuracy_std_dev": 0.21094196359873227,
      "run_count": 9,
      "total_questions": 1765,
      "avg_cost_per_task_usd": 0.014101193201133146,
      "avg_duration_per_task_ms": 357051.99943342776,
      "avg_output_tokens_per_task": 17061.87252124646,
      "last_run_timestamp": "2026-09-26T08:53:03.343Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6-20251208",
      "model_slug": "z-ai/glm-4.6v",
      "model_name": "Z.ai: GLM 4.6V",
      "provider_name": "Z.AI",
      "accuracy": 0.659933,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.014984520202020202,
      "avg_duration_per_task_ms": 348987.1767676768,
      "avg_output_tokens_per_task": 16461.863636363636,
      "last_run_timestamp": "2026-10-01T08:28:22.621Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Darkbloom",
      "accuracy": 0.6596062028241334,
      "accuracy_std_dev": 0.016648273731831805,
      "run_count": 4,
      "total_questions": 779,
      "avg_cost_per_task_usd": 0.011759379974326058,
      "avg_duration_per_task_ms": 1526360.2888318356,
      "avg_output_tokens_per_task": 50819.251604621306,
      "last_run_timestamp": "2026-10-03T17:41:49.078Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6595962,
      "accuracy_std_dev": 0.011187342962473291,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0024944898989898985,
      "avg_duration_per_task_ms": 89627.26666666666,
      "avg_output_tokens_per_task": 2311.8535353535353,
      "last_run_timestamp": "2026-08-29T20:54:16.286Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Venice",
      "accuracy": 0.6594517142857144,
      "accuracy_std_dev": 0.023871300387789566,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.027211155122655123,
      "avg_duration_per_task_ms": 890187.8643578644,
      "avg_output_tokens_per_task": 18800.659451659452,
      "last_run_timestamp": "2026-10-03T16:57:14.637Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.658249,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0066776414141414145,
      "avg_duration_per_task_ms": 101359.9898989899,
      "avg_output_tokens_per_task": 9307.358585858587,
      "last_run_timestamp": "2026-10-04T03:07:35.168Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "ibm-granite/granite-4.2-8b-20260831",
      "model_slug": "ibm-granite/granite-4.2-8b",
      "model_name": "IBM: Granite 4.2 8B",
      "provider_name": "auto-routing",
      "accuracy": 0.65736,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 197,
      "avg_cost_per_task_usd": 0.014864365482233502,
      "avg_duration_per_task_ms": 1717906.416243655,
      "avg_output_tokens_per_task": 64415.350253807104,
      "last_run_timestamp": "2026-10-01T08:48:38.304Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.656205,
      "accuracy_std_dev": 0.016336828302947918,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.008925360028860031,
      "avg_duration_per_task_ms": 319523.6370851371,
      "avg_output_tokens_per_task": 59199.669191919194,
      "last_run_timestamp": "2026-10-03T17:56:49.434Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "DeepInfra",
      "accuracy": 0.6549602541636494,
      "accuracy_std_dev": 0.09036662854873644,
      "run_count": 7,
      "total_questions": 1381,
      "avg_cost_per_task_usd": 0.011515717595944969,
      "avg_duration_per_task_ms": 1066625.0246198408,
      "avg_output_tokens_per_task": 33708.20130340333,
      "last_run_timestamp": "2026-10-08T22:36:32.095Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "ibm-granite/granite-4.2-8b-20260831",
      "model_slug": "ibm-granite/granite-4.2-8b",
      "model_name": "IBM: Granite 4.2 8B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6523569999999999,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.01652235606060606,
      "avg_duration_per_task_ms": 1917916.494949495,
      "avg_output_tokens_per_task": 65966.15909090909,
      "last_run_timestamp": "2026-10-09T00:15:37.020Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6512783352272727,
      "accuracy_std_dev": 0.07301920016707937,
      "run_count": 5,
      "total_questions": 704,
      "avg_cost_per_task_usd": 0.06126106249999999,
      "avg_duration_per_task_ms": 1570715.0482954546,
      "avg_output_tokens_per_task": 22881.873579545456,
      "last_run_timestamp": "2026-08-22T17:09:42.218Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "Crusoe",
      "accuracy": 0.650363823580786,
      "accuracy_std_dev": 0.007103203700131033,
      "run_count": 6,
      "total_questions": 1145,
      "avg_cost_per_task_usd": 0.0036315545851528385,
      "avg_duration_per_task_ms": 58503.03493449782,
      "avg_output_tokens_per_task": 2202.661135371179,
      "last_run_timestamp": "2026-08-29T16:10:45.295Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "Parasail",
      "accuracy": 0.6495510000000001,
      "accuracy_std_dev": 0.00792134559276387,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0032256414141414147,
      "avg_duration_per_task_ms": 40985.67760942761,
      "avg_output_tokens_per_task": 3060.60101010101,
      "last_run_timestamp": "2026-08-29T19:51:51.198Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "Novita",
      "accuracy": 0.649551,
      "accuracy_std_dev": 0.009444634067377443,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.002492604377104377,
      "avg_duration_per_task_ms": 61525.246632996634,
      "avg_output_tokens_per_task": 2691.0799663299663,
      "last_run_timestamp": "2026-08-29T17:03:33.297Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.6492706666666667,
      "accuracy_std_dev": 0.011195005920300148,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.004075954545454546,
      "avg_duration_per_task_ms": 28668.510101010103,
      "avg_output_tokens_per_task": 2354.649831649832,
      "last_run_timestamp": "2026-10-01T19:12:31.553Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.6487794999999998,
      "accuracy_std_dev": 0.12844600199597497,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.6039836092171716,
      "avg_duration_per_task_ms": 625325.3693181818,
      "avg_output_tokens_per_task": 40216.142676767675,
      "last_run_timestamp": "2026-10-03T10:37:00.593Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "CoreWeave",
      "accuracy": 0.6476671428571429,
      "accuracy_std_dev": 0.01801337100543905,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.007287767676767677,
      "avg_duration_per_task_ms": 701126.4617604618,
      "avg_output_tokens_per_task": 56538.52525252525,
      "last_run_timestamp": "2026-10-03T18:53:13.795Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "ibm-granite/granite-4.2-8b-20260831",
      "model_slug": "ibm-granite/granite-4.2-8b",
      "model_name": "IBM: Granite 4.2 8B",
      "provider_name": "CoreWeave",
      "accuracy": 0.647493,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 113,
      "avg_cost_per_task_usd": 0.005189893805309735,
      "avg_duration_per_task_ms": 619234.2654867256,
      "avg_output_tokens_per_task": 34365.4778761062,
      "last_run_timestamp": "2026-10-01T08:28:32.911Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.6473065000000001,
      "accuracy_std_dev": 0.009707729665065862,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.00413114393939394,
      "avg_duration_per_task_ms": 27377.72601010101,
      "avg_output_tokens_per_task": 2388.9305555555557,
      "last_run_timestamp": "2026-10-02T23:15:53.868Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "GMICloud",
      "accuracy": 0.646465,
      "accuracy_std_dev": 0.004761657064510171,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0029047154882154884,
      "avg_duration_per_task_ms": 53981.06734006734,
      "avg_output_tokens_per_task": 2363.939393939394,
      "last_run_timestamp": "2026-08-29T16:33:11.860Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Parasail",
      "accuracy": 0.64646475,
      "accuracy_std_dev": 0.02185317805692116,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.007503073232323232,
      "avg_duration_per_task_ms": 814987.6553030303,
      "avg_output_tokens_per_task": 52529.47348484849,
      "last_run_timestamp": "2026-10-03T20:08:16.264Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "auto-routing",
      "accuracy": 0.6464646666666667,
      "accuracy_std_dev": 0.013537798302858865,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.015368191919191919,
      "avg_duration_per_task_ms": 14060.520202020201,
      "avg_output_tokens_per_task": 1727.3535353535353,
      "last_run_timestamp": "2026-10-01T19:12:45.033Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.6457914,
      "accuracy_std_dev": 0.017547304854022453,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.007042358080808081,
      "avg_duration_per_task_ms": 238513.37575757576,
      "avg_output_tokens_per_task": 11322.903535353535,
      "last_run_timestamp": "2026-10-03T10:20:18.247Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "DekaLLM",
      "accuracy": 0.6447479677419355,
      "accuracy_std_dev": 0.02407747186802543,
      "run_count": 3,
      "total_questions": 403,
      "avg_cost_per_task_usd": 0.003165789081885856,
      "avg_duration_per_task_ms": 173618.99503722086,
      "avg_output_tokens_per_task": 10324.468982630273,
      "last_run_timestamp": "2026-10-09T03:20:25.108Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "DeepInfra",
      "accuracy": 0.6445006666666666,
      "accuracy_std_dev": 0.013998443861928202,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.002421077441077441,
      "avg_duration_per_task_ms": 66682.89730639731,
      "avg_output_tokens_per_task": 2836.0850168350166,
      "last_run_timestamp": "2026-08-29T16:11:23.388Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "Azure",
      "accuracy": 0.643659,
      "accuracy_std_dev": 0.015752669318774726,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.015761657407407408,
      "avg_duration_per_task_ms": 18961.11531986532,
      "avg_output_tokens_per_task": 1683.8872053872053,
      "last_run_timestamp": "2026-10-01T19:14:40.644Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.643098,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.018197469696969695,
      "avg_duration_per_task_ms": 1435371.2424242424,
      "avg_output_tokens_per_task": 32217.19191919192,
      "last_run_timestamp": "2026-07-18T17:46:02.370Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Parasail",
      "accuracy": 0.6428570571428571,
      "accuracy_std_dev": 0.0682565,
      "run_count": 2,
      "total_questions": 175,
      "avg_cost_per_task_usd": 0.09669735428571428,
      "avg_duration_per_task_ms": 1267820.5257142857,
      "avg_output_tokens_per_task": 34805.99428571429,
      "last_run_timestamp": "2026-08-01T18:16:39.158Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.6428055813449023,
      "accuracy_std_dev": 0.011146034629373067,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.0026296478669558935,
      "avg_duration_per_task_ms": 140713.84381778742,
      "avg_output_tokens_per_task": 9405.384671005062,
      "last_run_timestamp": "2026-10-03T17:16:11.892Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Ambient",
      "accuracy": 0.641414,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00284089898989899,
      "avg_duration_per_task_ms": 209352.64646464647,
      "avg_output_tokens_per_task": 9266.040404040405,
      "last_run_timestamp": "2026-10-09T15:00:06.492Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6413729024390243,
      "accuracy_std_dev": 0.03637250000000003,
      "run_count": 2,
      "total_questions": 369,
      "avg_cost_per_task_usd": 0.027402341463414633,
      "avg_duration_per_task_ms": 991300.8970189702,
      "avg_output_tokens_per_task": 32627.39024390244,
      "last_run_timestamp": "2026-08-01T17:59:38.872Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6406363579175706,
      "accuracy_std_dev": 0.016263383886158854,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.002550553145336226,
      "avg_duration_per_task_ms": 225461.96384671004,
      "avg_output_tokens_per_task": 8275.081706435285,
      "last_run_timestamp": "2026-10-03T17:03:50.406Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "Azure",
      "accuracy": 0.6405723333333334,
      "accuracy_std_dev": 0.0030349496792460334,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.004210166666666667,
      "avg_duration_per_task_ms": 30431.28367003367,
      "avg_output_tokens_per_task": 2308.9638047138046,
      "last_run_timestamp": "2026-10-01T19:13:29.560Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6402916666666668,
      "accuracy_std_dev": 0.11880589124562244,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0025993900112233448,
      "avg_duration_per_task_ms": 225334.6874298541,
      "avg_output_tokens_per_task": 6052.3058361391695,
      "last_run_timestamp": "2026-10-05T20:06:05.844Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.6402118571428571,
      "accuracy_std_dev": 0.01348951470511788,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.004506121212121212,
      "avg_duration_per_task_ms": 72990.16883116883,
      "avg_output_tokens_per_task": 8472.02886002886,
      "last_run_timestamp": "2026-10-03T16:51:25.541Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "WandB",
      "accuracy": 0.639316,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.0024375846153846152,
      "avg_duration_per_task_ms": 97724.68205128206,
      "avg_output_tokens_per_task": 7853.553846153846,
      "last_run_timestamp": "2026-07-18T08:33:31.052Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6384837407407408,
      "accuracy_std_dev": 0.1301056437821581,
      "run_count": 27,
      "total_questions": 5346,
      "avg_cost_per_task_usd": 0.6548667364384586,
      "avg_duration_per_task_ms": 493199.0927796483,
      "avg_output_tokens_per_task": 40750.34081556304,
      "last_run_timestamp": "2026-10-03T10:04:02.304Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "auto-routing",
      "accuracy": 0.6380471428571428,
      "accuracy_std_dev": 0.01709744716473838,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.009828660894660893,
      "avg_duration_per_task_ms": 659565.3961038961,
      "avg_output_tokens_per_task": 50620.7886002886,
      "last_run_timestamp": "2026-10-03T20:22:45.900Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Nebius",
      "accuracy": 0.6371981347826087,
      "accuracy_std_dev": 0.014429669830892581,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.0032497253623188405,
      "avg_duration_per_task_ms": 218812.71086956523,
      "avg_output_tokens_per_task": 10563.355072463768,
      "last_run_timestamp": "2026-10-03T16:41:22.866Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Google",
      "accuracy": 0.6368643521865888,
      "accuracy_std_dev": 0.13454240742991402,
      "run_count": 26,
      "total_questions": 5145,
      "avg_cost_per_task_usd": 0.6621504369290573,
      "avg_duration_per_task_ms": 598729.5551020408,
      "avg_output_tokens_per_task": 41288.28415937804,
      "last_run_timestamp": "2026-10-03T10:34:21.171Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.636738,
      "accuracy_std_dev": 0.1325038277837034,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.6152635606060606,
      "avg_duration_per_task_ms": 631219.6868686868,
      "avg_output_tokens_per_task": 40865.88888888889,
      "last_run_timestamp": "2026-10-03T09:11:52.463Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Groq",
      "accuracy": 0.6361772567678425,
      "accuracy_std_dev": 0.013327375155318935,
      "run_count": 7,
      "total_questions": 1219,
      "avg_cost_per_task_usd": 0.007504426579163248,
      "avg_duration_per_task_ms": 35368.206726825265,
      "avg_output_tokens_per_task": 12428.807219031993,
      "last_run_timestamp": "2026-08-29T16:23:45.123Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Groq",
      "accuracy": 0.6353917965162311,
      "accuracy_std_dev": 0.014982452781401947,
      "run_count": 7,
      "total_questions": 1263,
      "avg_cost_per_task_usd": 0.016417965162311955,
      "avg_duration_per_task_ms": 97650.78147268409,
      "avg_output_tokens_per_task": 54500.377672209026,
      "last_run_timestamp": "2026-10-03T17:21:59.201Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Darkbloom",
      "accuracy": 0.6352981489141675,
      "accuracy_std_dev": 0.02103570346434842,
      "run_count": 5,
      "total_questions": 967,
      "avg_cost_per_task_usd": 0.006571570837642193,
      "avg_duration_per_task_ms": 890942.791106515,
      "avg_output_tokens_per_task": 36274.74457083764,
      "last_run_timestamp": "2026-10-03T08:12:46.446Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Fireworks",
      "accuracy": 0.6352257765567765,
      "accuracy_std_dev": 0.022072905759475025,
      "run_count": 3,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.013415796703296703,
      "avg_duration_per_task_ms": 548616.1373626373,
      "avg_output_tokens_per_task": 46655.3021978022,
      "last_run_timestamp": "2026-08-22T17:46:40.285Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "StreamLake",
      "accuracy": 0.6340491250000001,
      "accuracy_std_dev": 0.012397945832248788,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0016212228535353534,
      "avg_duration_per_task_ms": 88876.8813131313,
      "avg_output_tokens_per_task": 8195.30366161616,
      "last_run_timestamp": "2026-10-03T16:47:32.559Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Azure",
      "accuracy": 0.633745,
      "accuracy_std_dev": 0.1400560587510904,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.6191078804713805,
      "avg_duration_per_task_ms": 634861.303030303,
      "avg_output_tokens_per_task": 41204.60381593715,
      "last_run_timestamp": "2026-10-03T10:40:39.138Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Venice",
      "accuracy": 0.632856543563069,
      "accuracy_std_dev": 0.17446938590265632,
      "run_count": 4,
      "total_questions": 769,
      "avg_cost_per_task_usd": 0.06782110143042913,
      "avg_duration_per_task_ms": 1537627.533159948,
      "avg_output_tokens_per_task": 34138.841352405725,
      "last_run_timestamp": "2026-08-22T17:02:13.071Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.6328437272727273,
      "accuracy_std_dev": 0.03335048624993999,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.006873024334251608,
      "avg_duration_per_task_ms": 112649.99265381084,
      "avg_output_tokens_per_task": 9586.507805325988,
      "last_run_timestamp": "2026-10-03T10:32:36.729Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "Crusoe",
      "accuracy": 0.6325426993295513,
      "accuracy_std_dev": 0.010534965676735725,
      "run_count": 10,
      "total_questions": 1939,
      "avg_cost_per_task_usd": 0.01766013512119649,
      "avg_duration_per_task_ms": 478451.96183599794,
      "avg_output_tokens_per_task": 88282.73388344508,
      "last_run_timestamp": "2026-10-03T17:08:02.072Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-fin-20260827",
      "model_slug": "inclusionai/ling-3.0-flash-fin",
      "model_name": "inclusionAI: Ling 3.0 Flash Fin",
      "provider_name": "Novita",
      "accuracy": 0.632155,
      "accuracy_std_dev": 0.005892000000000008,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.005966434343434343,
      "avg_duration_per_task_ms": 235473.01515151514,
      "avg_output_tokens_per_task": 48249.11363636364,
      "last_run_timestamp": "2026-10-01T08:24:52.614Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "SiliconFlow",
      "accuracy": 0.632154875,
      "accuracy_std_dev": 0.01720972334494007,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0017519513888888888,
      "avg_duration_per_task_ms": 214138.922979798,
      "avg_output_tokens_per_task": 9606.703282828283,
      "last_run_timestamp": "2026-10-03T16:12:29.903Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Venice",
      "accuracy": 0.6319606293706294,
      "accuracy_std_dev": 0.11679624770601185,
      "run_count": 7,
      "total_questions": 1287,
      "avg_cost_per_task_usd": 0.09059485703185705,
      "avg_duration_per_task_ms": 737977.4195804195,
      "avg_output_tokens_per_task": 28229.06216006216,
      "last_run_timestamp": "2026-10-03T16:55:14.697Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Ionstream",
      "accuracy": 0.63118499609375,
      "accuracy_std_dev": 0.12107794090401254,
      "run_count": 3,
      "total_questions": 512,
      "avg_cost_per_task_usd": 0.00688423046875,
      "avg_duration_per_task_ms": 160958.9140625,
      "avg_output_tokens_per_task": 8535.359375,
      "last_run_timestamp": "2026-08-08T16:46:11.805Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "CoreWeave",
      "accuracy": 0.6308680607902736,
      "accuracy_std_dev": 0.011456259937693471,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.0034108257345491387,
      "avg_duration_per_task_ms": 140057.4062816616,
      "avg_output_tokens_per_task": 11472.941236068895,
      "last_run_timestamp": "2026-08-29T18:20:19.448Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "Nebius",
      "accuracy": 0.6297015953307392,
      "accuracy_std_dev": 0.01592646788319292,
      "run_count": 8,
      "total_questions": 1542,
      "avg_cost_per_task_usd": 0.023244770428015562,
      "avg_duration_per_task_ms": 1042943.9273670558,
      "avg_output_tokens_per_task": 96652.03437094683,
      "last_run_timestamp": "2026-10-03T17:27:37.016Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Darkbloom",
      "accuracy": 0.6293489999999999,
      "accuracy_std_dev": 0.008199388066597833,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0011419814814814814,
      "avg_duration_per_task_ms": 334863.3063973064,
      "avg_output_tokens_per_task": 11634.42760942761,
      "last_run_timestamp": "2026-10-03T17:38:50.485Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Venice",
      "accuracy": 0.6292411856287424,
      "accuracy_std_dev": 0.0117815,
      "run_count": 2,
      "total_questions": 334,
      "avg_cost_per_task_usd": 0.0071094161676646716,
      "avg_duration_per_task_ms": 112258.81736526947,
      "avg_output_tokens_per_task": 28232.077844311378,
      "last_run_timestamp": "2026-08-15T08:02:44.624Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.6285795064570466,
      "accuracy_std_dev": 0.020919173358428876,
      "run_count": 9,
      "total_questions": 1781,
      "avg_cost_per_task_usd": 0.01718929421673217,
      "avg_duration_per_task_ms": 631165.4817518248,
      "avg_output_tokens_per_task": 83912.1010668164,
      "last_run_timestamp": "2026-10-03T17:40:51.448Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Alibaba",
      "accuracy": 0.626263,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.012936333333333333,
      "avg_duration_per_task_ms": 385474.66161616164,
      "avg_output_tokens_per_task": 30893.848484848484,
      "last_run_timestamp": "2026-07-18T16:20:00.500Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.6261091711026616,
      "accuracy_std_dev": 0.05551308764606775,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.0414238669201521,
      "avg_duration_per_task_ms": 392853.54372623574,
      "avg_output_tokens_per_task": 20615.842839036755,
      "last_run_timestamp": "2026-08-22T10:29:25.897Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "auto-routing",
      "accuracy": 0.623263890625,
      "accuracy_std_dev": 0.046540499999999985,
      "run_count": 2,
      "total_questions": 384,
      "avg_cost_per_task_usd": 0.03095198958333333,
      "avg_duration_per_task_ms": 834783.7890625,
      "avg_output_tokens_per_task": 38538.341145833336,
      "last_run_timestamp": "2026-08-01T18:25:56.196Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "SiliconFlow",
      "accuracy": 0.623188452685422,
      "accuracy_std_dev": 0.0187657309831867,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.014858657289002557,
      "avg_duration_per_task_ms": 724190.4782608695,
      "avg_output_tokens_per_task": 15645.849104859335,
      "last_run_timestamp": "2026-10-03T11:06:38.068Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "auto-routing",
      "accuracy": 0.6212125000000001,
      "accuracy_std_dev": 0.011784500000000031,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0023548611111111114,
      "avg_duration_per_task_ms": 64938.90151515151,
      "avg_output_tokens_per_task": 2165.318181818182,
      "last_run_timestamp": "2026-10-01T19:21:21.945Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.6212125000000001,
      "accuracy_std_dev": 0.011784500000000031,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.008206156565656566,
      "avg_duration_per_task_ms": 339475.95454545453,
      "avg_output_tokens_per_task": 15670.888888888889,
      "last_run_timestamp": "2026-08-01T18:08:49.019Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.621212,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00552190909090909,
      "avg_duration_per_task_ms": 174055.64141414143,
      "avg_output_tokens_per_task": 10415.944444444445,
      "last_run_timestamp": "2026-10-04T03:09:40.101Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Venice",
      "accuracy": 0.619708208029197,
      "accuracy_std_dev": 0.029586940351224878,
      "run_count": 7,
      "total_questions": 1370,
      "avg_cost_per_task_usd": 0.06579531313868613,
      "avg_duration_per_task_ms": 1037486.9379562044,
      "avg_output_tokens_per_task": 26177.450364963504,
      "last_run_timestamp": "2026-10-04T00:28:53.518Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Phala",
      "accuracy": 0.6148083751080381,
      "accuracy_std_dev": 0.23157299340777357,
      "run_count": 7,
      "total_questions": 1157,
      "avg_cost_per_task_usd": 0.005702518582541054,
      "avg_duration_per_task_ms": 156483.37078651684,
      "avg_output_tokens_per_task": 9454.971477960242,
      "last_run_timestamp": "2026-10-03T16:57:17.907Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "NextBit",
      "accuracy": 0.614478,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.009234040404040404,
      "avg_duration_per_task_ms": 990972.4696969697,
      "avg_output_tokens_per_task": 16689.560606060608,
      "last_run_timestamp": "2026-07-18T08:03:56.103Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.6132539090909092,
      "accuracy_std_dev": 0.1299958112139524,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.6145130344352616,
      "avg_duration_per_task_ms": 628297.9467401286,
      "avg_output_tokens_per_task": 40850.65381083563,
      "last_run_timestamp": "2026-10-03T10:13:27.517Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6128760749682338,
      "accuracy_std_dev": 0.007225322047320813,
      "run_count": 4,
      "total_questions": 787,
      "avg_cost_per_task_usd": 0.0033133367217280814,
      "avg_duration_per_task_ms": 404332.3011435832,
      "avg_output_tokens_per_task": 13398.312579415502,
      "last_run_timestamp": "2026-10-08T22:16:42.056Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Nebius",
      "accuracy": 0.6108306666666666,
      "accuracy_std_dev": 0.019131244624668038,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.005598868686868688,
      "avg_duration_per_task_ms": 615283.2525252525,
      "avg_output_tokens_per_task": 18654.404882154882,
      "last_run_timestamp": "2026-08-29T19:06:01.882Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Google",
      "accuracy": 0.6102353048048047,
      "accuracy_std_dev": 0.004626267235105652,
      "run_count": 7,
      "total_questions": 1332,
      "avg_cost_per_task_usd": 0.004039552552552552,
      "avg_duration_per_task_ms": 36383.835585585584,
      "avg_output_tokens_per_task": 2146.3363363363364,
      "last_run_timestamp": "2026-10-03T17:03:50.598Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Novita",
      "accuracy": 0.6096681428571429,
      "accuracy_std_dev": 0.010277069224089844,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0036427301587301586,
      "avg_duration_per_task_ms": 34466.44155844156,
      "avg_output_tokens_per_task": 2150.969696969697,
      "last_run_timestamp": "2026-10-03T17:03:29.436Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.6096471504551988,
      "accuracy_std_dev": 0.016445311419549056,
      "run_count": 11,
      "total_questions": 2087,
      "avg_cost_per_task_usd": 0.008342372783900336,
      "avg_duration_per_task_ms": 542298.5462386201,
      "avg_output_tokens_per_task": 13701.597987541927,
      "last_run_timestamp": "2026-10-08T22:50:36.237Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "auto-routing",
      "accuracy": 0.6084389949367088,
      "accuracy_std_dev": 0.010283734794972733,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.008408201687763715,
      "avg_duration_per_task_ms": 279347.2447257384,
      "avg_output_tokens_per_task": 17442.14936708861,
      "last_run_timestamp": "2026-08-29T19:52:22.642Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6077443333333332,
      "accuracy_std_dev": 0.001374572014200141,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0077788232323232325,
      "avg_duration_per_task_ms": 287812.98148148146,
      "avg_output_tokens_per_task": 15362.217171717171,
      "last_run_timestamp": "2026-10-08T22:57:53.933Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "CoreWeave",
      "accuracy": 0.607744,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.003953611111111112,
      "avg_duration_per_task_ms": 57224.242424242424,
      "avg_output_tokens_per_task": 2094.1161616161617,
      "last_run_timestamp": "2026-08-01T17:44:33.510Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "WandB",
      "accuracy": 0.607744,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.004114661616161616,
      "avg_duration_per_task_ms": 51474.35353535353,
      "avg_output_tokens_per_task": 2201.479797979798,
      "last_run_timestamp": "2026-07-18T08:26:39.830Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.6060605,
      "accuracy_std_dev": 0.001683499999999949,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.008235411616161616,
      "avg_duration_per_task_ms": 152037.7323232323,
      "avg_output_tokens_per_task": 15653.954545454546,
      "last_run_timestamp": "2026-08-01T17:51:29.206Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Z.AI",
      "accuracy": 0.605239731707317,
      "accuracy_std_dev": 0.09170649999999997,
      "run_count": 2,
      "total_questions": 369,
      "avg_cost_per_task_usd": 0.03298509756097561,
      "avg_duration_per_task_ms": 869238.1111111111,
      "avg_output_tokens_per_task": 34362.55826558266,
      "last_run_timestamp": "2026-08-01T18:05:37.080Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "auto-routing",
      "accuracy": 0.6046178571428571,
      "accuracy_std_dev": 0.006995035043894045,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0034120865800865803,
      "avg_duration_per_task_ms": 53141.78787878788,
      "avg_output_tokens_per_task": 2114.74531024531,
      "last_run_timestamp": "2026-10-03T16:57:23.266Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Alibaba",
      "accuracy": 0.6036557142857142,
      "accuracy_std_dev": 0.0048341465405498454,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.011036557720057719,
      "avg_duration_per_task_ms": 33894.917748917745,
      "avg_output_tokens_per_task": 2102.8852813852814,
      "last_run_timestamp": "2026-10-03T17:43:44.972Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6026559059760956,
      "accuracy_std_dev": 0.018857781098506346,
      "run_count": 7,
      "total_questions": 1255,
      "avg_cost_per_task_usd": 0.002205658167330677,
      "avg_duration_per_task_ms": 47812.13147410358,
      "avg_output_tokens_per_task": 2066.2470119521913,
      "last_run_timestamp": "2026-10-08T23:15:07.689Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Venice",
      "accuracy": 0.6002888571428572,
      "accuracy_std_dev": 0.011234048798548233,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.003266399711399712,
      "avg_duration_per_task_ms": 60828.217893217894,
      "avg_output_tokens_per_task": 2084.8766233766232,
      "last_run_timestamp": "2026-10-03T16:59:44.383Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Nebius",
      "accuracy": 0.599327,
      "accuracy_std_dev": 0.0033669999999999534,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.014991068181818182,
      "avg_duration_per_task_ms": 284433.01515151514,
      "avg_output_tokens_per_task": 12302.075757575758,
      "last_run_timestamp": "2026-08-15T15:26:03.892Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.599327,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.012440297979797978,
      "avg_duration_per_task_ms": 230792.08080808082,
      "avg_output_tokens_per_task": 41242.71717171717,
      "last_run_timestamp": "2026-10-04T03:37:16.537Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Novita",
      "accuracy": 0.5980126829268292,
      "accuracy_std_dev": 0.061359,
      "run_count": 2,
      "total_questions": 369,
      "avg_cost_per_task_usd": 0.03206006233062331,
      "avg_duration_per_task_ms": 831219.6937669377,
      "avg_output_tokens_per_task": 40208.840108401084,
      "last_run_timestamp": "2026-08-01T18:07:49.346Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "NextBit",
      "accuracy": 0.5973626666666666,
      "accuracy_std_dev": 0.014397638598819675,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.003072445286195286,
      "avg_duration_per_task_ms": 655562.0867003367,
      "avg_output_tokens_per_task": 13595.005050505051,
      "last_run_timestamp": "2026-10-07T13:51:26.485Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Novita",
      "accuracy": 0.5968335954198474,
      "accuracy_std_dev": 0.034825528389147715,
      "run_count": 6,
      "total_questions": 1179,
      "avg_cost_per_task_usd": 0.024714866836301953,
      "avg_duration_per_task_ms": 1074343.121289228,
      "avg_output_tokens_per_task": 64653.8074639525,
      "last_run_timestamp": "2026-08-30T00:32:50.853Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.591751,
      "accuracy_std_dev": 0.006627755628164546,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.017204867003367003,
      "avg_duration_per_task_ms": 118615.4696969697,
      "avg_output_tokens_per_task": 42873.59259259259,
      "last_run_timestamp": "2026-10-01T19:14:48.696Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "auto-routing",
      "accuracy": 0.591751,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0029208080808080807,
      "avg_duration_per_task_ms": 521963.51515151514,
      "avg_output_tokens_per_task": 11688.588383838383,
      "last_run_timestamp": "2026-10-01T16:32:22.405Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "Novita",
      "accuracy": 0.5878479469525959,
      "accuracy_std_dev": 0.06702947204028986,
      "run_count": 10,
      "total_questions": 1772,
      "avg_cost_per_task_usd": 0.009048836343115122,
      "avg_duration_per_task_ms": 528814.486455982,
      "avg_output_tokens_per_task": 45050.80643340858,
      "last_run_timestamp": "2026-10-03T17:04:03.802Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "CoreWeave",
      "accuracy": 0.5872781331360947,
      "accuracy_std_dev": 0.004687499999999956,
      "run_count": 2,
      "total_questions": 338,
      "avg_cost_per_task_usd": 0.09507673668639052,
      "avg_duration_per_task_ms": 191708.50591715975,
      "avg_output_tokens_per_task": 21455.86094674556,
      "last_run_timestamp": "2026-08-01T18:26:20.700Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Novita",
      "accuracy": 0.5860806989010988,
      "accuracy_std_dev": 0.014672256934467207,
      "run_count": 7,
      "total_questions": 1365,
      "avg_cost_per_task_usd": 0.006546478388278388,
      "avg_duration_per_task_ms": 728895.4901098901,
      "avg_output_tokens_per_task": 43384.46666666667,
      "last_run_timestamp": "2026-10-03T16:12:07.492Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Cerebras",
      "accuracy": 0.5836942857142858,
      "accuracy_std_dev": 0.01603982892422384,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.010374448051948051,
      "avg_duration_per_task_ms": 11818.474025974027,
      "avg_output_tokens_per_task": 13378.126262626263,
      "last_run_timestamp": "2026-10-03T17:02:12.175Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Darkbloom",
      "accuracy": 0.5767120715096482,
      "accuracy_std_dev": 0.20359371905450435,
      "run_count": 9,
      "total_questions": 1762,
      "avg_cost_per_task_usd": 0.023737154370034052,
      "avg_duration_per_task_ms": 751908.8337116913,
      "avg_output_tokens_per_task": 33435.202610669694,
      "last_run_timestamp": "2026-10-03T17:20:14.386Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Cloudflare",
      "accuracy": 0.5750844,
      "accuracy_std_dev": 0.02978589572666902,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0254069303030303,
      "avg_duration_per_task_ms": 1275373.0656565656,
      "avg_output_tokens_per_task": 63407.47878787879,
      "last_run_timestamp": "2026-08-30T00:49:28.653Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "SambaNova",
      "accuracy": 0.5717949931623931,
      "accuracy_std_dev": 0.029316541569280166,
      "run_count": 6,
      "total_questions": 585,
      "avg_cost_per_task_usd": 0.032133635897435904,
      "avg_duration_per_task_ms": 57054.20170940171,
      "avg_output_tokens_per_task": 6867.617094017094,
      "last_run_timestamp": "2026-10-04T00:00:34.891Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.5676551155419224,
      "accuracy_std_dev": 0.08425995197346976,
      "run_count": 7,
      "total_questions": 978,
      "avg_cost_per_task_usd": 0.09743748875255624,
      "avg_duration_per_task_ms": 1096781.8711656441,
      "avg_output_tokens_per_task": 27754.531697341514,
      "last_run_timestamp": "2026-10-08T23:22:48.508Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Tenstorrent",
      "accuracy": 0.564814935483871,
      "accuracy_std_dev": 0.015030477400564811,
      "run_count": 3,
      "total_questions": 558,
      "avg_cost_per_task_usd": 0.0051593978494623655,
      "avg_duration_per_task_ms": 615212.1326164874,
      "avg_output_tokens_per_task": 14268.974910394265,
      "last_run_timestamp": "2026-10-05T18:22:40.566Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.5633876206896551,
      "accuracy_std_dev": 0.09194318097738406,
      "run_count": 5,
      "total_questions": 986,
      "avg_cost_per_task_usd": 0.029487100405679513,
      "avg_duration_per_task_ms": 2640133.111561866,
      "avg_output_tokens_per_task": 23360.7738336714,
      "last_run_timestamp": "2026-08-22T18:19:05.094Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "DeepInfra",
      "accuracy": 0.5630111428571428,
      "accuracy_std_dev": 0.01687958484110801,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.004603796536796536,
      "avg_duration_per_task_ms": 419928.0497835498,
      "avg_output_tokens_per_task": 16214.744588744588,
      "last_run_timestamp": "2026-10-08T23:30:03.872Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DekaLLM",
      "accuracy": 0.56229,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0034265555555555555,
      "avg_duration_per_task_ms": 3024088.404040404,
      "avg_output_tokens_per_task": 30654.242424242424,
      "last_run_timestamp": "2026-07-18T16:00:09.483Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DeepInfra",
      "accuracy": 0.5602958771331057,
      "accuracy_std_dev": 0.05661860675431709,
      "run_count": 5,
      "total_questions": 586,
      "avg_cost_per_task_usd": 0.05058588907849829,
      "avg_duration_per_task_ms": 611241.4914675768,
      "avg_output_tokens_per_task": 24217.037542662118,
      "last_run_timestamp": "2026-08-29T19:07:49.385Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.5505053333333333,
      "accuracy_std_dev": 0.016890890559378122,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.017397131313131312,
      "avg_duration_per_task_ms": 129348.1271043771,
      "avg_output_tokens_per_task": 43400.656565656565,
      "last_run_timestamp": "2026-10-01T19:27:01.395Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.5488216666666668,
      "accuracy_std_dev": 0.006298902038362641,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0171891531986532,
      "avg_duration_per_task_ms": 121387.12121212122,
      "avg_output_tokens_per_task": 42892.78114478115,
      "last_run_timestamp": "2026-10-01T19:15:36.386Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "OpenAI",
      "accuracy": 0.542088,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.023943484848484846,
      "avg_duration_per_task_ms": 10793.646464646465,
      "avg_output_tokens_per_task": 1334.611111111111,
      "last_run_timestamp": "2026-08-01T08:53:34.329Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.538721,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.005473,
      "avg_duration_per_task_ms": 117831.5808080808,
      "avg_output_tokens_per_task": 11819.661616161617,
      "last_run_timestamp": "2026-10-04T03:06:12.419Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "DeepInfra",
      "accuracy": 0.535985,
      "accuracy_std_dev": 0.011835000612589788,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0018682651515151514,
      "avg_duration_per_task_ms": 92043.92676767676,
      "avg_output_tokens_per_task": 1839.9911616161617,
      "last_run_timestamp": "2026-10-08T22:10:22.349Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.5353534999999999,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.005200381313131313,
      "avg_duration_per_task_ms": 113966.90151515152,
      "avg_output_tokens_per_task": 11220.5,
      "last_run_timestamp": "2026-08-01T17:48:17.532Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "Alibaba",
      "accuracy": 0.5331086666666667,
      "accuracy_std_dev": 0.05754091454106565,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.005119237373737374,
      "avg_duration_per_task_ms": 101325.84511784512,
      "avg_output_tokens_per_task": 5420.193602693603,
      "last_run_timestamp": "2026-10-01T16:05:56.758Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Google",
      "accuracy": 0.5310693945086705,
      "accuracy_std_dev": 0.0817773924709448,
      "run_count": 7,
      "total_questions": 1384,
      "avg_cost_per_task_usd": 0.009087861994219654,
      "avg_duration_per_task_ms": 353941.95809248555,
      "avg_output_tokens_per_task": 26784.09176300578,
      "last_run_timestamp": "2026-10-03T16:13:54.660Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.5282142748691099,
      "accuracy_std_dev": 0.013706854874680634,
      "run_count": 6,
      "total_questions": 1146,
      "avg_cost_per_task_usd": 0.01651366666666667,
      "avg_duration_per_task_ms": 725814.7731239093,
      "avg_output_tokens_per_task": 41825.98516579407,
      "last_run_timestamp": "2026-08-30T00:11:29.373Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-2024-08-06",
      "model_slug": "openai/gpt-4o-2024-08-06",
      "model_name": "OpenAI: GPT-4o (2024-08-06)",
      "provider_name": "OpenAI",
      "accuracy": 0.5280586666666666,
      "accuracy_std_dev": 0.004827144451490511,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.01799902861952862,
      "avg_duration_per_task_ms": 14939.786195286195,
      "avg_output_tokens_per_task": 1607.080808080808,
      "last_run_timestamp": "2026-10-02T23:16:52.382Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "OpenAI",
      "accuracy": 0.522126,
      "accuracy_std_dev": 0.02474951208754282,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.01514130735930736,
      "avg_duration_per_task_ms": 13491.472582972583,
      "avg_output_tokens_per_task": 1320.8239538239538,
      "last_run_timestamp": "2026-10-02T23:17:37.980Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.5202023333333333,
      "accuracy_std_dev": 0.0023810642345154617,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0035723956228956226,
      "avg_duration_per_task_ms": 90157.08922558922,
      "avg_output_tokens_per_task": 2721.4242424242425,
      "last_run_timestamp": "2026-08-15T08:46:42.897Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-2024-08-06",
      "model_slug": "openai/gpt-4o-2024-08-06",
      "model_name": "OpenAI: GPT-4o (2024-08-06)",
      "provider_name": "auto-routing",
      "accuracy": 0.5202019999999999,
      "accuracy_std_dev": 0.0033669999999999534,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017511833333333334,
      "avg_duration_per_task_ms": 16177.118686868687,
      "avg_output_tokens_per_task": 1559.209595959596,
      "last_run_timestamp": "2026-10-01T16:04:12.703Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.5190796666666666,
      "accuracy_std_dev": 0.008837261503932561,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.004016927609427609,
      "avg_duration_per_task_ms": 28916.838383838385,
      "avg_output_tokens_per_task": 2584.1296296296296,
      "last_run_timestamp": "2026-08-15T09:25:42.974Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "DeepInfra",
      "accuracy": 0.517469628909552,
      "accuracy_std_dev": 0.07510589055164418,
      "run_count": 6,
      "total_questions": 1183,
      "avg_cost_per_task_usd": 0.003801546914623838,
      "avg_duration_per_task_ms": 137539.94759087067,
      "avg_output_tokens_per_task": 4071.0912933220625,
      "last_run_timestamp": "2026-08-29T20:34:10.117Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "Azure",
      "accuracy": 0.5168355,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017194939393939394,
      "avg_duration_per_task_ms": 13199.247474747475,
      "avg_output_tokens_per_task": 1523.2777777777778,
      "last_run_timestamp": "2026-10-01T16:28:34.286Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "DigitalOcean",
      "accuracy": 0.5157126666666666,
      "accuracy_std_dev": 0.02305568803763809,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.004710382154882156,
      "avg_duration_per_task_ms": 211036.34006734006,
      "avg_output_tokens_per_task": 2694.722222222222,
      "last_run_timestamp": "2026-08-15T09:11:29.276Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-2024-08-06",
      "model_slug": "openai/gpt-4o-2024-08-06",
      "model_name": "OpenAI: GPT-4o (2024-08-06)",
      "provider_name": "Azure",
      "accuracy": 0.5151519999999999,
      "accuracy_std_dev": 0.006733999999999962,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.018041959595959597,
      "avg_duration_per_task_ms": 13498.479797979799,
      "avg_output_tokens_per_task": 1611.5353535353536,
      "last_run_timestamp": "2026-10-01T16:26:08.836Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "SiliconFlow",
      "accuracy": 0.5134683333333333,
      "accuracy_std_dev": 0.00836132925370652,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0008388316498316498,
      "avg_duration_per_task_ms": 89120.7138047138,
      "avg_output_tokens_per_task": 2792.7188552188554,
      "last_run_timestamp": "2026-08-15T08:49:03.156Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.513468,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.006161969696969697,
      "avg_duration_per_task_ms": 165410.29292929292,
      "avg_output_tokens_per_task": 12691.05303030303,
      "last_run_timestamp": "2026-08-01T18:08:24.732Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "auto-routing",
      "accuracy": 0.5080647741935485,
      "accuracy_std_dev": 0.12443414088451235,
      "run_count": 6,
      "total_questions": 1116,
      "avg_cost_per_task_usd": 0.004283293010752688,
      "avg_duration_per_task_ms": 107626.95967741935,
      "avg_output_tokens_per_task": 3879.1989247311826,
      "last_run_timestamp": "2026-08-29T20:57:58.443Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "auto-routing",
      "accuracy": 0.506734,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0237055303030303,
      "avg_duration_per_task_ms": 11478.515151515152,
      "avg_output_tokens_per_task": 1318.7474747474748,
      "last_run_timestamp": "2026-08-01T08:34:23.935Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "auto-routing",
      "accuracy": 0.5056116666666668,
      "accuracy_std_dev": 0.007570922855827376,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.012942734006734005,
      "avg_duration_per_task_ms": 12612.308080808081,
      "avg_output_tokens_per_task": 1100.3198653198654,
      "last_run_timestamp": "2026-10-01T16:03:45.472Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-fin-20260827",
      "model_slug": "inclusionai/ling-3.0-flash-fin",
      "model_name": "inclusionAI: Ling 3.0 Flash Fin",
      "provider_name": "auto-routing",
      "accuracy": 0.505051,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.009755954545454545,
      "avg_duration_per_task_ms": 700759.0555555555,
      "avg_output_tokens_per_task": 64762.606060606064,
      "last_run_timestamp": "2026-10-01T08:24:17.043Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "Azure",
      "accuracy": 0.505051,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02292151515151515,
      "avg_duration_per_task_ms": 12797.353535353535,
      "avg_output_tokens_per_task": 1266.479797979798,
      "last_run_timestamp": "2026-08-01T08:47:47.980Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-nano-2025-04-14",
      "model_slug": "openai/gpt-4.1-nano",
      "model_name": "OpenAI: GPT-4.1 Nano",
      "provider_name": "auto-routing",
      "accuracy": 0.5,
      "accuracy_std_dev": 0.015120700182200542,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0008362104377104377,
      "avg_duration_per_task_ms": 19874.548821548822,
      "avg_output_tokens_per_task": 1898.5521885521885,
      "last_run_timestamp": "2026-10-01T19:13:37.127Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-nano-2025-04-14",
      "model_slug": "openai/gpt-4.1-nano",
      "model_name": "OpenAI: GPT-4.1 Nano",
      "provider_name": "Azure",
      "accuracy": 0.49764320000000006,
      "accuracy_std_dev": 0.01852187333290023,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0008495191919191919,
      "avg_duration_per_task_ms": 13033.880808080808,
      "avg_output_tokens_per_task": 1851.6353535353535,
      "last_run_timestamp": "2026-10-01T19:12:29.219Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-nano-2025-04-14",
      "model_slug": "openai/gpt-4.1-nano",
      "model_name": "OpenAI: GPT-4.1 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.4966332500000001,
      "accuracy_std_dev": 0.014773079898501197,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0008230315656565656,
      "avg_duration_per_task_ms": 17739.39898989899,
      "avg_output_tokens_per_task": 1863.909090909091,
      "last_run_timestamp": "2026-10-02T23:16:50.042Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.49601803668941985,
      "accuracy_std_dev": 0.017030085894772614,
      "run_count": 6,
      "total_questions": 1172,
      "avg_cost_per_task_usd": 0.01113635409556314,
      "avg_duration_per_task_ms": 463652.90699658706,
      "avg_output_tokens_per_task": 27776.748293515357,
      "last_run_timestamp": "2026-08-30T01:15:14.669Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Groq",
      "accuracy": 0.49537025,
      "accuracy_std_dev": 0.016425006010272865,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.002352652146464646,
      "avg_duration_per_task_ms": 10902.190025252525,
      "avg_output_tokens_per_task": 2331.179292929293,
      "last_run_timestamp": "2026-10-03T16:58:35.370Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Venice",
      "accuracy": 0.4939061636197441,
      "accuracy_std_dev": 0.024075339672231702,
      "run_count": 6,
      "total_questions": 1094,
      "avg_cost_per_task_usd": 0.010896170018281536,
      "avg_duration_per_task_ms": 477480.3711151737,
      "avg_output_tokens_per_task": 26830.819012797074,
      "last_run_timestamp": "2026-08-30T00:23:23.003Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "SambaNova",
      "accuracy": 0.48621612499999994,
      "accuracy_std_dev": 0.005119178558067202,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.002366635732323232,
      "avg_duration_per_task_ms": 13903.847853535353,
      "avg_output_tokens_per_task": 2226.974116161616,
      "last_run_timestamp": "2026-10-03T16:13:49.760Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "AkashML",
      "accuracy": 0.4840632105511069,
      "accuracy_std_dev": 0.017868444314808295,
      "run_count": 11,
      "total_questions": 2123,
      "avg_cost_per_task_usd": 0.0011087908619877532,
      "avg_duration_per_task_ms": 61751.02025435704,
      "avg_output_tokens_per_task": 2146.873763542157,
      "last_run_timestamp": "2026-10-09T15:26:28.059Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Cerebras",
      "accuracy": 0.48063949999999994,
      "accuracy_std_dev": 0.003857437653417109,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.04852650126262626,
      "avg_duration_per_task_ms": 45991.458333333336,
      "avg_output_tokens_per_task": 17009.132575757576,
      "last_run_timestamp": "2026-08-15T14:46:10.940Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "CoreWeave",
      "accuracy": 0.4772725,
      "accuracy_std_dev": 0.0074500120973056095,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.002467788720538721,
      "avg_duration_per_task_ms": 42647.26262626263,
      "avg_output_tokens_per_task": 2666.861952861953,
      "last_run_timestamp": "2026-10-03T16:39:56.573Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.47667114285714296,
      "accuracy_std_dev": 0.012154133775663925,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0013214949494949496,
      "avg_duration_per_task_ms": 47661.26911976912,
      "avg_output_tokens_per_task": 2366.4862914862915,
      "last_run_timestamp": "2026-10-03T16:14:03.229Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.4766039334298119,
      "accuracy_std_dev": 0.01707390743987169,
      "run_count": 7,
      "total_questions": 1382,
      "avg_cost_per_task_usd": 0.0013226222865412446,
      "avg_duration_per_task_ms": 98211.54196816208,
      "avg_output_tokens_per_task": 3034.6222865412446,
      "last_run_timestamp": "2026-10-03T17:16:32.671Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen-2.5-72b-instruct",
      "model_slug": "qwen/qwen-2.5-72b-instruct",
      "model_name": "Qwen2.5 72B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.4765793529411765,
      "accuracy_std_dev": 0.006944000000000006,
      "run_count": 2,
      "total_questions": 306,
      "avg_cost_per_task_usd": 0.0011232875816993464,
      "avg_duration_per_task_ms": 126677.51307189543,
      "avg_output_tokens_per_task": 2069.3627450980393,
      "last_run_timestamp": "2026-10-01T16:04:45.916Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Together",
      "accuracy": 0.474536875,
      "accuracy_std_dev": 0.01241241161537012,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.003145674873737374,
      "avg_duration_per_task_ms": 53376.7619949495,
      "avg_output_tokens_per_task": 2218.6376262626263,
      "last_run_timestamp": "2026-10-03T17:41:38.624Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Nebius",
      "accuracy": 0.4741861666666667,
      "accuracy_std_dev": 0.012808772416026279,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0010600345117845118,
      "avg_duration_per_task_ms": 199985.06313131313,
      "avg_output_tokens_per_task": 2395.604377104377,
      "last_run_timestamp": "2026-08-29T20:09:29.346Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.4713801428571428,
      "accuracy_std_dev": 0.009733517737746226,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0013883506493506494,
      "avg_duration_per_task_ms": 69013.97546897546,
      "avg_output_tokens_per_task": 2273.031746031746,
      "last_run_timestamp": "2026-10-03T16:57:23.026Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.469697,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.006110727272727273,
      "avg_duration_per_task_ms": 42879.368686868685,
      "avg_output_tokens_per_task": 2318.919191919192,
      "last_run_timestamp": "2026-10-04T03:02:23.034Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "WandB",
      "accuracy": 0.469697,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0022648939393939394,
      "avg_duration_per_task_ms": 33028.21717171717,
      "avg_output_tokens_per_task": 2380.9646464646466,
      "last_run_timestamp": "2026-07-18T16:13:31.071Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Crusoe",
      "accuracy": 0.46742483824768327,
      "accuracy_std_dev": 0.006948085005956674,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.002137747262005055,
      "avg_duration_per_task_ms": 34994.02274641955,
      "avg_output_tokens_per_task": 2587.217354675653,
      "last_run_timestamp": "2026-08-29T16:31:42.504Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen2.5-vl-72b-instruct",
      "model_slug": "qwen/qwen2.5-vl-72b-instruct",
      "model_name": "Qwen: Qwen2.5 VL 72B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.462542,
      "accuracy_std_dev": 0.005471000000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.001783989898989899,
      "avg_duration_per_task_ms": 27644.777777777777,
      "avg_output_tokens_per_task": 1397.5252525252524,
      "last_run_timestamp": "2026-09-01T13:03:42.840Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Ionstream",
      "accuracy": 0.46128619685039374,
      "accuracy_std_dev": 0.075159,
      "run_count": 2,
      "total_questions": 254,
      "avg_cost_per_task_usd": 0.00731248031496063,
      "avg_duration_per_task_ms": 1016477.8543307086,
      "avg_output_tokens_per_task": 20586.740157480315,
      "last_run_timestamp": "2026-08-08T10:11:59.748Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen-2.5-72b-instruct",
      "model_slug": "qwen/qwen-2.5-72b-instruct",
      "model_name": "Qwen2.5 72B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.461279,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0011232171717171719,
      "avg_duration_per_task_ms": 136788.7626262626,
      "avg_output_tokens_per_task": 2036.2272727272727,
      "last_run_timestamp": "2026-10-01T16:27:00.375Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.46106887500000004,
      "accuracy_std_dev": 0.01782538514617215,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0007930915404040404,
      "avg_duration_per_task_ms": 157279.59027777778,
      "avg_output_tokens_per_task": 2227.837121212121,
      "last_run_timestamp": "2026-10-08T22:58:25.155Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen-2.5-72b-instruct",
      "model_slug": "qwen/qwen-2.5-72b-instruct",
      "model_name": "Qwen2.5 72B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.45959600000000006,
      "accuracy_std_dev": 0.01346799999999998,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0010278964646464647,
      "avg_duration_per_task_ms": 96379.95959595959,
      "avg_output_tokens_per_task": 1848.388888888889,
      "last_run_timestamp": "2026-10-01T16:47:11.191Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.45862900000000006,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 141,
      "avg_cost_per_task_usd": 0.0011888297872340425,
      "avg_duration_per_task_ms": 278701.0070921986,
      "avg_output_tokens_per_task": 7719.879432624113,
      "last_run_timestamp": "2026-08-01T18:13:08.069Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen2.5-vl-72b-instruct",
      "model_slug": "qwen/qwen2.5-vl-72b-instruct",
      "model_name": "Qwen: Qwen2.5 VL 72B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.4444445,
      "accuracy_std_dev": 0.015151500000000012,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0015239469696969696,
      "avg_duration_per_task_ms": 59255.792929292926,
      "avg_output_tokens_per_task": 1592.9924242424242,
      "last_run_timestamp": "2026-09-01T13:21:35.512Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen2.5-vl-72b-instruct",
      "model_slug": "qwen/qwen2.5-vl-72b-instruct",
      "model_name": "Qwen: Qwen2.5 VL 72B Instruct",
      "provider_name": "Nebius",
      "accuracy": 0.4402355,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0013545,
      "avg_duration_per_task_ms": 68001.7196969697,
      "avg_output_tokens_per_task": 1535.1843434343434,
      "last_run_timestamp": "2026-09-01T12:41:25.955Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-mini",
      "model_slug": "openai/gpt-4o-mini",
      "model_name": "OpenAI: GPT-4o-mini",
      "provider_name": "auto-routing",
      "accuracy": 0.4320986666666667,
      "accuracy_std_dev": 0.00781631474687543,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0010060690235690235,
      "avg_duration_per_task_ms": 20918.510101010103,
      "avg_output_tokens_per_task": 1483.5286195286196,
      "last_run_timestamp": "2026-10-01T19:16:07.087Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-mini",
      "model_slug": "openai/gpt-4o-mini",
      "model_name": "OpenAI: GPT-4o-mini",
      "provider_name": "Azure",
      "accuracy": 0.4315373333333334,
      "accuracy_std_dev": 0.010588221658469791,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0010723358585858589,
      "avg_duration_per_task_ms": 21765.176767676767,
      "avg_output_tokens_per_task": 1505.858585858586,
      "last_run_timestamp": "2026-10-01T19:05:18.003Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-mini",
      "model_slug": "openai/gpt-4o-mini",
      "model_name": "OpenAI: GPT-4o-mini",
      "provider_name": "OpenAI",
      "accuracy": 0.4305555,
      "accuracy_std_dev": 0.007469952225416168,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0010262436868686867,
      "avg_duration_per_task_ms": 17602.7702020202,
      "avg_output_tokens_per_task": 1517.9292929292928,
      "last_run_timestamp": "2026-10-02T23:19:57.879Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AkashML",
      "accuracy": 0.40740725,
      "accuracy_std_dev": 0.3006428663691981,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.017353968434343434,
      "avg_duration_per_task_ms": 189056.44444444444,
      "avg_output_tokens_per_task": 19318.219696969696,
      "last_run_timestamp": "2026-08-15T15:24:35.389Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-3-12b-it",
      "model_slug": "google/gemma-3-12b-it",
      "model_name": "Google: Gemma 3 12B",
      "provider_name": "DeepInfra",
      "accuracy": 0.405724,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0003259848484848485,
      "avg_duration_per_task_ms": 29056.348484848484,
      "avg_output_tokens_per_task": 1917.1515151515152,
      "last_run_timestamp": "2026-10-08T22:01:16.852Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.3994878450704225,
      "accuracy_std_dev": 0.0901063557941392,
      "run_count": 4,
      "total_questions": 781,
      "avg_cost_per_task_usd": 0.016465559539052498,
      "avg_duration_per_task_ms": 3369255.4609475033,
      "avg_output_tokens_per_task": 16587.451984635085,
      "last_run_timestamp": "2026-08-22T17:37:26.215Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy-mt2-30b-a3b-20260521",
      "model_slug": "tencent/hy-mt2-30b-a3b",
      "model_name": "Tencent: Hy-MT2-30B-A3B",
      "provider_name": "Reka",
      "accuracy": 0.388889,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00023147979797979797,
      "avg_duration_per_task_ms": 8604.353535353535,
      "avg_output_tokens_per_task": 710.540404040404,
      "last_run_timestamp": "2026-09-24T06:00:03.613Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.38720499999999997,
      "accuracy_std_dev": 0.010647388881786938,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.00015740101010101008,
      "avg_duration_per_task_ms": 6591.242424242424,
      "avg_output_tokens_per_task": 937.0363636363636,
      "last_run_timestamp": "2026-10-01T19:39:32.831Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "Novita",
      "accuracy": 0.382155,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.05791383333333333,
      "avg_duration_per_task_ms": 1031309.4747474748,
      "avg_output_tokens_per_task": 22994.131313131315,
      "last_run_timestamp": "2026-10-03T19:39:23.691Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-3-12b-it",
      "model_slug": "google/gemma-3-12b-it",
      "model_name": "Google: Gemma 3 12B",
      "provider_name": "NextBit",
      "accuracy": 0.37878799999999996,
      "accuracy_std_dev": 0.01683499999999999,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.000344469696969697,
      "avg_duration_per_task_ms": 46431.74494949495,
      "avg_output_tokens_per_task": 2036.358585858586,
      "last_run_timestamp": "2026-10-07T13:35:26.756Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen-2.5-7b-instruct",
      "model_slug": "qwen/qwen-2.5-7b-instruct",
      "model_name": "Qwen: Qwen2.5 7B Instruct",
      "provider_name": "Phala",
      "accuracy": 0.36363599999999996,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00041437878787878784,
      "avg_duration_per_task_ms": 34981.92424242424,
      "avg_output_tokens_per_task": 1665.6666666666667,
      "last_run_timestamp": "2026-08-01T08:55:19.318Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "Mistral",
      "accuracy": 0.350168,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00011657070707070707,
      "avg_duration_per_task_ms": 7587.606060606061,
      "avg_output_tokens_per_task": 355.9040404040404,
      "last_run_timestamp": "2026-07-18T09:09:45.839Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Cloudflare",
      "accuracy": 0.3483847621145374,
      "accuracy_std_dev": 0.08501017676557487,
      "run_count": 7,
      "total_questions": 1362,
      "avg_cost_per_task_usd": 0.00338636490455213,
      "avg_duration_per_task_ms": 34858.14243759178,
      "avg_output_tokens_per_task": 1419.5543318649045,
      "last_run_timestamp": "2026-10-03T16:11:17.907Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Reka",
      "accuracy": 0.346801,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008616373737373737,
      "avg_duration_per_task_ms": 1179181.5303030303,
      "avg_output_tokens_per_task": 32995.232323232325,
      "last_run_timestamp": "2026-10-03T00:19:24.096Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "Novita",
      "accuracy": 0.344383,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 181,
      "avg_cost_per_task_usd": 0.00011365193370165745,
      "avg_duration_per_task_ms": 62171.872928176796,
      "avg_output_tokens_per_task": 485.27624309392263,
      "last_run_timestamp": "2026-07-18T08:00:15.379Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen-2.5-7b-instruct",
      "model_slug": "qwen/qwen-2.5-7b-instruct",
      "model_name": "Qwen: Qwen2.5 7B Instruct",
      "provider_name": "Together",
      "accuracy": 0.3417510000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0007054343434343434,
      "avg_duration_per_task_ms": 14064.868686868687,
      "avg_output_tokens_per_task": 1539,
      "last_run_timestamp": "2026-08-01T08:06:23.600Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "auto-routing",
      "accuracy": 0.338384,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00013078787878787878,
      "avg_duration_per_task_ms": 16627.883838383837,
      "avg_output_tokens_per_task": 667.8888888888889,
      "last_run_timestamp": "2026-07-18T09:08:52.364Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen-2.5-7b-instruct",
      "model_slug": "qwen/qwen-2.5-7b-instruct",
      "model_name": "Qwen: Qwen2.5 7B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.328283,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0005318181818181818,
      "avg_duration_per_task_ms": 36388.32323232323,
      "avg_output_tokens_per_task": 2041.050505050505,
      "last_run_timestamp": "2026-08-01T08:36:11.581Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "DeepInfra",
      "accuracy": 0.301347,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00002756060606060606,
      "avg_duration_per_task_ms": 15713.474747474747,
      "avg_output_tokens_per_task": 431.8787878787879,
      "last_run_timestamp": "2026-07-18T08:00:15.193Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "DekaLLM",
      "accuracy": 0.289562,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00002257070707070707,
      "avg_duration_per_task_ms": 19325.131313131315,
      "avg_output_tokens_per_task": 285.6515151515151,
      "last_run_timestamp": "2026-07-18T08:04:03.449Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sao10k/l3-lunaris-8b",
      "model_slug": "sao10k/l3-lunaris-8b",
      "model_name": "Sao10K: Llama 3 8B Lunaris",
      "provider_name": "DeepInfra",
      "accuracy": 0.2891415,
      "accuracy_std_dev": 0.003232916137174003,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.00006112626262626263,
      "avg_duration_per_task_ms": 10017.040404040405,
      "avg_output_tokens_per_task": 577.7083333333334,
      "last_run_timestamp": "2026-10-08T22:00:27.722Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.2886002857142857,
      "accuracy_std_dev": 0.01680762263901442,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0015827532467532469,
      "avg_duration_per_task_ms": 97706.50937950939,
      "avg_output_tokens_per_task": 20607.09163059163,
      "last_run_timestamp": "2026-10-03T17:14:12.203Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "Groq",
      "accuracy": 0.2783791428571428,
      "accuracy_std_dev": 0.011859313909672988,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0022877604617604613,
      "avg_duration_per_task_ms": 48962.33477633478,
      "avg_output_tokens_per_task": 28673.977633477632,
      "last_run_timestamp": "2026-10-03T16:55:38.344Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "CoreWeave",
      "accuracy": 0.27641208165057063,
      "accuracy_std_dev": 0.013856526257206988,
      "run_count": 6,
      "total_questions": 1139,
      "avg_cost_per_task_usd": 0.007524540825285337,
      "avg_duration_per_task_ms": 300907.412642669,
      "avg_output_tokens_per_task": 33503.4363476734,
      "last_run_timestamp": "2026-10-03T17:43:23.676Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.275533,
      "accuracy_std_dev": 0.009152576176501717,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0007405521885521885,
      "avg_duration_per_task_ms": 48360.410774410775,
      "avg_output_tokens_per_task": 2532.1531986531986,
      "last_run_timestamp": "2026-08-15T09:28:07.208Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sao10k/l3-lunaris-8b",
      "model_slug": "sao10k/l3-lunaris-8b",
      "model_name": "Sao10K: Llama 3 8B Lunaris",
      "provider_name": "Novita",
      "accuracy": 0.2739361560283688,
      "accuracy_std_dev": 0.008153796102974918,
      "run_count": 3,
      "total_questions": 564,
      "avg_cost_per_task_usd": 0.00006610283687943263,
      "avg_duration_per_task_ms": 8879.929078014184,
      "avg_output_tokens_per_task": 551.2074468085107,
      "last_run_timestamp": "2026-10-01T16:01:04.877Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sao10k/l3-lunaris-8b",
      "model_slug": "sao10k/l3-lunaris-8b",
      "model_name": "Sao10K: Llama 3 8B Lunaris",
      "provider_name": "auto-routing",
      "accuracy": 0.273148,
      "accuracy_std_dev": 0.006113657170630353,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.00006268181818181819,
      "avg_duration_per_task_ms": 9553.181818181818,
      "avg_output_tokens_per_task": 586.439393939394,
      "last_run_timestamp": "2026-10-01T20:43:57.173Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Venice",
      "accuracy": 0.27244666666666667,
      "accuracy_std_dev": 0.06424850635012633,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.008843218013468015,
      "avg_duration_per_task_ms": 589547.7979797979,
      "avg_output_tokens_per_task": 21876.559764309764,
      "last_run_timestamp": "2026-10-03T17:50:31.478Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sao10k/l3-lunaris-8b",
      "model_slug": "sao10k/l3-lunaris-8b",
      "model_name": "Sao10K: Llama 3 8B Lunaris",
      "provider_name": "Parasail",
      "accuracy": 0.2700681612244898,
      "accuracy_std_dev": 0.016752437809465226,
      "run_count": 5,
      "total_questions": 980,
      "avg_cost_per_task_usd": 0.00005762346938775511,
      "avg_duration_per_task_ms": 9734.870408163266,
      "avg_output_tokens_per_task": 576.8010204081633,
      "last_run_timestamp": "2026-10-01T20:47:49.128Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.2668783085714286,
      "accuracy_std_dev": 0.01699007316742264,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.00036382412698412705,
      "avg_duration_per_task_ms": 401076.15555555554,
      "avg_output_tokens_per_task": 8879.231746031746,
      "last_run_timestamp": "2026-10-08T23:14:52.443Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.25484025,
      "accuracy_std_dev": 0.015150020327956655,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0003991950757575757,
      "avg_duration_per_task_ms": 68227.3188131313,
      "avg_output_tokens_per_task": 7660.273358585859,
      "last_run_timestamp": "2026-10-03T16:12:07.494Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "WandB",
      "accuracy": 0.254209,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.001291050505050505,
      "avg_duration_per_task_ms": 40150.21212121212,
      "avg_output_tokens_per_task": 5062.732323232323,
      "last_run_timestamp": "2026-07-18T08:26:34.893Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "Cloudflare",
      "accuracy": 0.23148166666666667,
      "accuracy_std_dev": 0.10429433789947032,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0013719949494949493,
      "avg_duration_per_task_ms": 502712.0462962963,
      "avg_output_tokens_per_task": 4456.498316498317,
      "last_run_timestamp": "2026-08-29T17:16:14.290Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.2-3b-instruct",
      "model_slug": "meta-llama/llama-3.2-3b-instruct",
      "model_name": "Meta: Llama 3.2 3B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.21734899999999996,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 171,
      "avg_cost_per_task_usd": 0.0003523450292397661,
      "avg_duration_per_task_ms": 72727.70760233919,
      "avg_output_tokens_per_task": 4269.023391812865,
      "last_run_timestamp": "2026-08-01T08:05:56.858Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "Novita",
      "accuracy": 0.1910775,
      "accuracy_std_dev": 0.11363615432122239,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0014224511784511782,
      "avg_duration_per_task_ms": 46218.05892255892,
      "avg_output_tokens_per_task": 1225.9587542087543,
      "last_run_timestamp": "2026-08-29T18:43:09.114Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-fin-20260827",
      "model_slug": "inclusionai/ling-3.0-flash-fin",
      "model_name": "inclusionAI: Ling 3.0 Flash Fin",
      "provider_name": "DeepInfra",
      "accuracy": 0.17174282233502539,
      "accuracy_std_dev": 0.01788397279378631,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.019897788494077834,
      "avg_duration_per_task_ms": 1784246.4196277496,
      "avg_output_tokens_per_task": 110294.06598984772,
      "last_run_timestamp": "2026-10-08T23:13:58.126Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.2-3b-instruct",
      "model_slug": "meta-llama/llama-3.2-3b-instruct",
      "model_name": "Meta: Llama 3.2 3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.116162,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0001537979797979798,
      "avg_duration_per_task_ms": 29722.909090909092,
      "avg_output_tokens_per_task": 1836.7929292929293,
      "last_run_timestamp": "2026-08-01T09:03:02.970Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-flash-lite-image-20260630",
      "model_slug": "google/gemini-3.1-flash-lite-image",
      "model_name": "Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)",
      "provider_name": "Google AI Studio",
      "accuracy": 0.1161615,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017850737373737373,
      "avg_duration_per_task_ms": 46853.982323232325,
      "avg_output_tokens_per_task": 11776.444444444445,
      "last_run_timestamp": "2026-10-01T16:00:58.389Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-flash-lite-image-20260630",
      "model_slug": "google/gemini-3.1-flash-lite-image",
      "model_name": "Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)",
      "provider_name": "Google",
      "accuracy": 0.112795,
      "accuracy_std_dev": 0.006733999999999997,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.018038613636363634,
      "avg_duration_per_task_ms": 21125.79292929293,
      "avg_output_tokens_per_task": 11795.222222222223,
      "last_run_timestamp": "2026-10-01T16:00:58.364Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-flash-lite-image-20260630",
      "model_slug": "google/gemini-3.1-flash-lite-image",
      "model_name": "Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)",
      "provider_name": "auto-routing",
      "accuracy": 0.102694,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017984406565656568,
      "avg_duration_per_task_ms": 21554.590909090908,
      "avg_output_tokens_per_task": 11812.823232323231,
      "last_run_timestamp": "2026-10-01T16:02:13.023Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.2-3b-instruct",
      "model_slug": "meta-llama/llama-3.2-3b-instruct",
      "model_name": "Meta: Llama 3.2 3B Instruct",
      "provider_name": "Cloudflare",
      "accuracy": 0.060606,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.000040161616161616165,
      "avg_duration_per_task_ms": 3671.1060606060605,
      "avg_output_tokens_per_task": 93.72222222222223,
      "last_run_timestamp": "2026-08-01T08:05:34.450Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "SiliconFlow",
      "accuracy": 0.002245,
      "accuracy_std_dev": 0.0007933738084913065,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.003512888888888889,
      "avg_duration_per_task_ms": 175891.9090909091,
      "avg_output_tokens_per_task": 3312.4141414141413,
      "last_run_timestamp": "2026-08-22T17:29:06.480Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.001684,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0034544065656565656,
      "avg_duration_per_task_ms": 164367.952020202,
      "avg_output_tokens_per_task": 3253.931818181818,
      "last_run_timestamp": "2026-08-15T14:54:58.071Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "BaseTen",
      "accuracy": 0.822917,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.0087319375,
      "avg_duration_per_task_ms": 31283.875,
      "avg_output_tokens_per_task": 9903.5625,
      "last_run_timestamp": "2026-09-26T08:05:12.798Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Baidu",
      "accuracy": 0.806667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03810672,
      "avg_duration_per_task_ms": 107015.82,
      "avg_output_tokens_per_task": 17830.82,
      "last_run_timestamp": "2026-10-03T08:11:26.752Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.806667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.20567228,
      "avg_duration_per_task_ms": 69455.26,
      "avg_output_tokens_per_task": 15737.28,
      "last_run_timestamp": "2026-08-15T00:15:11.753Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Azure",
      "accuracy": 0.8051851111111111,
      "accuracy_std_dev": 0.020068562864143984,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.87284366,
      "avg_duration_per_task_ms": 135007.57555555555,
      "avg_output_tokens_per_task": 5778.728888888889,
      "last_run_timestamp": "2026-10-03T08:06:43.563Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8,
      "accuracy_std_dev": 0.033110611481316166,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.7322859599999999,
      "avg_duration_per_task_ms": 146639.07333333333,
      "avg_output_tokens_per_task": 5472.34,
      "last_run_timestamp": "2026-08-15T08:09:14.229Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "WandB",
      "accuracy": 0.8,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.030123700000000003,
      "avg_duration_per_task_ms": 294712.06,
      "avg_output_tokens_per_task": 14851.14,
      "last_run_timestamp": "2026-07-18T08:03:40.888Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "Xiaomi",
      "accuracy": 0.8,
      "accuracy_std_dev": 0.006667000000000034,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.053672830000000005,
      "avg_duration_per_task_ms": 853001.35,
      "avg_output_tokens_per_task": 37564.6,
      "last_run_timestamp": "2026-10-03T08:15:57.209Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7985185555555555,
      "accuracy_std_dev": 0.0278934726458882,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.9265346288888889,
      "avg_duration_per_task_ms": 128415.08888888889,
      "avg_output_tokens_per_task": 5765.7555555555555,
      "last_run_timestamp": "2026-10-03T08:17:58.914Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Sail Research",
      "accuracy": 0.7977776666666667,
      "accuracy_std_dev": 0.01911642045177097,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.008363246666666668,
      "avg_duration_per_task_ms": 272879.14,
      "avg_output_tokens_per_task": 12625.026666666667,
      "last_run_timestamp": "2026-08-22T16:20:38.832Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7952381428571429,
      "accuracy_std_dev": 0.02038185356374333,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.8750597885714285,
      "avg_duration_per_task_ms": 147569.17142857143,
      "avg_output_tokens_per_task": 5729.828571428571,
      "last_run_timestamp": "2026-10-03T08:21:31.946Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.793333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.020424579999999998,
      "avg_duration_per_task_ms": 413179.4,
      "avg_output_tokens_per_task": 22096.3,
      "last_run_timestamp": "2026-10-03T08:38:04.173Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Anthropic",
      "accuracy": 0.7925926666666666,
      "accuracy_std_dev": 0.01616105838118283,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.8220114666666667,
      "avg_duration_per_task_ms": 137835.87333333332,
      "avg_output_tokens_per_task": 5751.835555555555,
      "last_run_timestamp": "2026-10-03T08:17:58.906Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.79166675,
      "accuracy_std_dev": 0.012801703448662584,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.71616581,
      "avg_duration_per_task_ms": 111830.815,
      "avg_output_tokens_per_task": 6900.955,
      "last_run_timestamp": "2026-10-01T08:03:01.529Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Anthropic",
      "accuracy": 0.7906666,
      "accuracy_std_dev": 0.009977824303925176,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.814870324,
      "avg_duration_per_task_ms": 182126.652,
      "avg_output_tokens_per_task": 7046.908,
      "last_run_timestamp": "2026-10-01T08:05:22.427Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepSeek",
      "accuracy": 0.79,
      "accuracy_std_dev": 0.01666640000120001,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.016075525,
      "avg_duration_per_task_ms": 177598.75,
      "avg_output_tokens_per_task": 11878.23,
      "last_run_timestamp": "2026-08-08T08:29:14.821Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Google",
      "accuracy": 0.7896079411764706,
      "accuracy_std_dev": 0.020702777175085597,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.9446278929411763,
      "avg_duration_per_task_ms": 97552.96117647059,
      "avg_output_tokens_per_task": 5626.15294117647,
      "last_run_timestamp": "2026-10-03T08:25:13.024Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Parasail",
      "accuracy": 0.7871621216216216,
      "accuracy_std_dev": 0.015496260115123086,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.06734989864864864,
      "avg_duration_per_task_ms": 463154.4662162162,
      "avg_output_tokens_per_task": 9087.885135135135,
      "last_run_timestamp": "2026-08-01T17:45:43.989Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "WandB",
      "accuracy": 0.786667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.14809964,
      "avg_duration_per_task_ms": 328566.04,
      "avg_output_tokens_per_task": 11598.26,
      "last_run_timestamp": "2026-07-18T08:04:04.764Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7862320000000002,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 46,
      "avg_cost_per_task_usd": 0.029582717391304348,
      "avg_duration_per_task_ms": 2435445.304347826,
      "avg_output_tokens_per_task": 108721.58695652174,
      "last_run_timestamp": "2026-10-03T16:12:07.492Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Crusoe",
      "accuracy": 0.7859532642140468,
      "accuracy_std_dev": 0.022060806735148097,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.028781180602006688,
      "avg_duration_per_task_ms": 346700.6989966555,
      "avg_output_tokens_per_task": 11874.802675585284,
      "last_run_timestamp": "2026-08-29T17:06:21.366Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7859258888888889,
      "accuracy_std_dev": 0.023611001599746027,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.4867972777777778,
      "avg_duration_per_task_ms": 127433.78,
      "avg_output_tokens_per_task": 7132.633333333333,
      "last_run_timestamp": "2026-10-03T08:09:09.780Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Crusoe",
      "accuracy": 0.785,
      "accuracy_std_dev": 0.009860245661239886,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.08538871499999999,
      "avg_duration_per_task_ms": 41332.83,
      "avg_output_tokens_per_task": 4686.75,
      "last_run_timestamp": "2026-10-03T08:21:29.529Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google",
      "accuracy": 0.7847225,
      "accuracy_std_dev": 0.020833500000000005,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.11493318749999999,
      "avg_duration_per_task_ms": 147359.25,
      "avg_output_tokens_per_task": 14387.041666666666,
      "last_run_timestamp": "2026-10-01T16:00:58.391Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DeepInfra",
      "accuracy": 0.784471297188755,
      "accuracy_std_dev": 0.02769278737577711,
      "run_count": 5,
      "total_questions": 249,
      "avg_cost_per_task_usd": 0.09832240562248996,
      "avg_duration_per_task_ms": 406639.9036144578,
      "avg_output_tokens_per_task": 17335.586345381525,
      "last_run_timestamp": "2026-09-26T08:23:09.250Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Chutes",
      "accuracy": 0.784090875,
      "accuracy_std_dev": 0.025439439704308266,
      "run_count": 9,
      "total_questions": 440,
      "avg_cost_per_task_usd": 0.07952366136363637,
      "avg_duration_per_task_ms": 1164717.7386363635,
      "avg_output_tokens_per_task": 26062.15681818182,
      "last_run_timestamp": "2026-10-03T16:05:34.190Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "BaseTen",
      "accuracy": 0.7835159358372455,
      "accuracy_std_dev": 0.020461293571987414,
      "run_count": 13,
      "total_questions": 639,
      "avg_cost_per_task_usd": 0.08511225195618151,
      "avg_duration_per_task_ms": 86820.21752738654,
      "avg_output_tokens_per_task": 4360.830985915493,
      "last_run_timestamp": "2026-10-03T08:39:04.230Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Wafer",
      "accuracy": 0.7833639159049359,
      "accuracy_std_dev": 0.011285272864949418,
      "run_count": 11,
      "total_questions": 547,
      "avg_cost_per_task_usd": 0.10846238025594149,
      "avg_duration_per_task_ms": 142469.36563071297,
      "avg_output_tokens_per_task": 11291.52285191956,
      "last_run_timestamp": "2026-10-03T08:33:36.127Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "BaseTen",
      "accuracy": 0.7827005274261605,
      "accuracy_std_dev": 0.02374291531720566,
      "run_count": 5,
      "total_questions": 237,
      "avg_cost_per_task_usd": 0.10862543037974683,
      "avg_duration_per_task_ms": 71286.46835443038,
      "avg_output_tokens_per_task": 8041.662447257384,
      "last_run_timestamp": "2026-08-29T17:15:26.910Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Crusoe",
      "accuracy": 0.782313,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 49,
      "avg_cost_per_task_usd": 0.01855418367346939,
      "avg_duration_per_task_ms": 64182.816326530614,
      "avg_output_tokens_per_task": 11933.040816326531,
      "last_run_timestamp": "2026-10-08T18:21:28.790Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.78166675,
      "accuracy_std_dev": 0.022047820503113248,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.06611864500000002,
      "avg_duration_per_task_ms": 311561.6025,
      "avg_output_tokens_per_task": 15444.9325,
      "last_run_timestamp": "2026-10-03T16:18:22.812Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7811890906432747,
      "accuracy_std_dev": 0.04113861939038473,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.11509685380116959,
      "avg_duration_per_task_ms": 3147556.2339181285,
      "avg_output_tokens_per_task": 25903.42105263158,
      "last_run_timestamp": "2026-10-05T21:59:19.672Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Fireworks",
      "accuracy": 0.7811111666666667,
      "accuracy_std_dev": 0.023544066905957912,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09874701999999999,
      "avg_duration_per_task_ms": 77113.90666666666,
      "avg_output_tokens_per_task": 4731.806666666666,
      "last_run_timestamp": "2026-10-04T21:00:06.533Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Azure",
      "accuracy": 0.7807408333333333,
      "accuracy_std_dev": 0.022758947352268594,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.47270033222222213,
      "avg_duration_per_task_ms": 109166.16777777778,
      "avg_output_tokens_per_task": 7230.745555555555,
      "last_run_timestamp": "2026-10-03T08:13:54.015Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "auto-routing",
      "accuracy": 0.7800003333333333,
      "accuracy_std_dev": 0.018856180831641284,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 1.2757919799999997,
      "avg_duration_per_task_ms": 102078.39333333333,
      "avg_output_tokens_per_task": 5685.653333333334,
      "last_run_timestamp": "2026-10-01T19:04:24.509Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Reka",
      "accuracy": 0.78,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12438624000000001,
      "avg_duration_per_task_ms": 511620.62,
      "avg_output_tokens_per_task": 12374.38,
      "last_run_timestamp": "2026-10-03T13:23:27.233Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Near AI",
      "accuracy": 0.78,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06720102,
      "avg_duration_per_task_ms": 196852.88,
      "avg_output_tokens_per_task": 12773.4,
      "last_run_timestamp": "2026-10-09T15:19:58.090Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Xiaomi",
      "accuracy": 0.78,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01813584,
      "avg_duration_per_task_ms": 915938.12,
      "avg_output_tokens_per_task": 36481,
      "last_run_timestamp": "2026-10-03T08:17:09.266Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Phala",
      "accuracy": 0.78,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.12014696000000001,
      "avg_duration_per_task_ms": 363719.42,
      "avg_output_tokens_per_task": 9566.28,
      "last_run_timestamp": "2026-07-18T16:00:09.009Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7790475714285714,
      "accuracy_std_dev": 0.02348347998583042,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.4780597828571429,
      "avg_duration_per_task_ms": 128145.17428571428,
      "avg_output_tokens_per_task": 7892.805714285714,
      "last_run_timestamp": "2026-10-03T08:15:57.205Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.778333375,
      "accuracy_std_dev": 0.020749956493553795,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1860782475,
      "avg_duration_per_task_ms": 191849.3575,
      "avg_output_tokens_per_task": 9582.0075,
      "last_run_timestamp": "2026-10-03T16:10:17.831Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Azure",
      "accuracy": 0.77833325,
      "accuracy_std_dev": 0.011902275693643617,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.19232103999999997,
      "avg_duration_per_task_ms": 75116.48,
      "avg_output_tokens_per_task": 6244.94,
      "last_run_timestamp": "2026-10-03T08:17:58.904Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Google",
      "accuracy": 0.777857107142857,
      "accuracy_std_dev": 0.024498453027457068,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.5571781464285714,
      "avg_duration_per_task_ms": 116452.10642857142,
      "avg_output_tokens_per_task": 7231.447142857143,
      "last_run_timestamp": "2026-10-03T08:36:29.351Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ambient",
      "accuracy": 0.777778,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.012714583333333333,
      "avg_duration_per_task_ms": 516568.7291666667,
      "avg_output_tokens_per_task": 14230.208333333334,
      "last_run_timestamp": "2026-08-03T17:20:47.207Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Z.AI",
      "accuracy": 0.7777776666666667,
      "accuracy_std_dev": 0.021998866415239553,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08069113666666665,
      "avg_duration_per_task_ms": 131032.13666666667,
      "avg_output_tokens_per_task": 4578.13,
      "last_run_timestamp": "2026-10-03T08:38:05.155Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7772839629629629,
      "accuracy_std_dev": 0.02231524453767886,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.4874687607407408,
      "avg_duration_per_task_ms": 125767.80296296296,
      "avg_output_tokens_per_task": 7179.893333333333,
      "last_run_timestamp": "2026-10-03T08:32:03.356Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7766669999999999,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.18507052999999998,
      "avg_duration_per_task_ms": 70446.4,
      "avg_output_tokens_per_task": 6214.06,
      "last_run_timestamp": "2026-10-03T08:13:12.599Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Google",
      "accuracy": 0.7766666666666666,
      "accuracy_std_dev": 0.011385435794128478,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.18964581,
      "avg_duration_per_task_ms": 65753.47666666667,
      "avg_output_tokens_per_task": 6306.6466666666665,
      "last_run_timestamp": "2026-10-03T08:23:24.811Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7766666666666666,
      "accuracy_std_dev": 0.02841481095407032,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.21599287333333336,
      "avg_duration_per_task_ms": 589051.7166666667,
      "avg_output_tokens_per_task": 12452.323333333334,
      "last_run_timestamp": "2026-10-08T22:44:43.725Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Cloudflare",
      "accuracy": 0.7766666666666666,
      "accuracy_std_dev": 0.026874233841275463,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.15658131666666666,
      "avg_duration_per_task_ms": 581092.8233333334,
      "avg_output_tokens_per_task": 12459.696666666667,
      "last_run_timestamp": "2026-10-03T08:17:09.274Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepSeek",
      "accuracy": 0.7766665,
      "accuracy_std_dev": 0.019148513142887433,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.008114253333333333,
      "avg_duration_per_task_ms": 195047.99,
      "avg_output_tokens_per_task": 21425.716666666667,
      "last_run_timestamp": "2026-08-08T08:30:38.880Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.7766665,
      "accuracy_std_dev": 0.016666499999999973,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.017346550000000002,
      "avg_duration_per_task_ms": 335231.13,
      "avg_output_tokens_per_task": 20113.88,
      "last_run_timestamp": "2026-10-03T08:17:48.643Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "DeepInfra",
      "accuracy": 0.7762865704697987,
      "accuracy_std_dev": 0.02285360036308408,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.07428861073825505,
      "avg_duration_per_task_ms": 1222334.8724832216,
      "avg_output_tokens_per_task": 35964.32214765101,
      "last_run_timestamp": "2026-10-08T23:26:25.306Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Phala",
      "accuracy": 0.7761905714285714,
      "accuracy_std_dev": 0.024907983039620864,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.18391273428571428,
      "avg_duration_per_task_ms": 728509.9057142857,
      "avg_output_tokens_per_task": 29807.6,
      "last_run_timestamp": "2026-10-03T16:06:04.623Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7760001,
      "accuracy_std_dev": 0.021333425001391598,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.44147494200000004,
      "avg_duration_per_task_ms": 131608.992,
      "avg_output_tokens_per_task": 7283.092,
      "last_run_timestamp": "2026-10-03T08:36:29.563Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Together",
      "accuracy": 0.776,
      "accuracy_std_dev": 0.012364752775530953,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.074485596,
      "avg_duration_per_task_ms": 47627.076,
      "avg_output_tokens_per_task": 4707.808,
      "last_run_timestamp": "2026-10-03T08:37:28.530Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7755376129032258,
      "accuracy_std_dev": 0.01731650665001458,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.1579919798387097,
      "avg_duration_per_task_ms": 333040.58467741933,
      "avg_output_tokens_per_task": 20363.45564516129,
      "last_run_timestamp": "2026-10-03T08:32:03.329Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Friendli",
      "accuracy": 0.775,
      "accuracy_std_dev": 0.009860245661239886,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.029526945,
      "avg_duration_per_task_ms": 114391.885,
      "avg_output_tokens_per_task": 11970.57,
      "last_run_timestamp": "2026-10-03T16:15:04.036Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Google",
      "accuracy": 0.7746668,
      "accuracy_std_dev": 0.012927338819726197,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.758212416,
      "avg_duration_per_task_ms": 112131.876,
      "avg_output_tokens_per_task": 6866.944,
      "last_run_timestamp": "2026-10-01T08:05:27.141Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Cloudflare",
      "accuracy": 0.7744441666666667,
      "accuracy_std_dev": 0.015112691911289488,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09511451000000001,
      "avg_duration_per_task_ms": 133903.96,
      "avg_output_tokens_per_task": 4573.283333333334,
      "last_run_timestamp": "2026-10-03T08:23:53.940Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Nebius",
      "accuracy": 0.7744105858585859,
      "accuracy_std_dev": 0.019115499999999952,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.24473854545454546,
      "avg_duration_per_task_ms": 51240.79797979798,
      "avg_output_tokens_per_task": 4631.454545454545,
      "last_run_timestamp": "2026-10-03T08:37:57.937Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Alibaba",
      "accuracy": 0.7742857142857142,
      "accuracy_std_dev": 0.017612503291811795,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.12337239714285714,
      "avg_duration_per_task_ms": 82299.41428571429,
      "avg_output_tokens_per_task": 4508.834285714286,
      "last_run_timestamp": "2026-10-03T08:19:27.959Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7741665000000001,
      "accuracy_std_dev": 0.04514642535406319,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.12616785249999998,
      "avg_duration_per_task_ms": 205632.94,
      "avg_output_tokens_per_task": 7251.665,
      "last_run_timestamp": "2026-10-05T19:59:14.351Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Alibaba",
      "accuracy": 0.77375,
      "accuracy_std_dev": 0.02188582748035816,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.127833135,
      "avg_duration_per_task_ms": 303219.4325,
      "avg_output_tokens_per_task": 19468.2,
      "last_run_timestamp": "2026-10-09T22:14:36.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Reka",
      "accuracy": 0.7733335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.11407636,
      "avg_duration_per_task_ms": 360681.75,
      "avg_output_tokens_per_task": 19997.86,
      "last_run_timestamp": "2026-10-01T22:40:52.402Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Io Net",
      "accuracy": 0.7733335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.10867763,
      "avg_duration_per_task_ms": 521628.03,
      "avg_output_tokens_per_task": 12092.47,
      "last_run_timestamp": "2026-09-19T08:11:59.798Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Anthropic",
      "accuracy": 0.7733333333333333,
      "accuracy_std_dev": 0.02914423052490644,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.4720031333333333,
      "avg_duration_per_task_ms": 115130.24444444444,
      "avg_output_tokens_per_task": 7206.9,
      "last_run_timestamp": "2026-10-03T08:16:37.919Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Anthropic",
      "accuracy": 0.77333325,
      "accuracy_std_dev": 0.017638436566794497,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.4344371716666668,
      "avg_duration_per_task_ms": 95528.38,
      "avg_output_tokens_per_task": 5662.913333333333,
      "last_run_timestamp": "2026-10-03T08:06:28.791Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepSeek",
      "accuracy": 0.77333325,
      "accuracy_std_dev": 0.004713927363409425,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.007096445,
      "avg_duration_per_task_ms": 164831.185,
      "avg_output_tokens_per_task": 17141.055,
      "last_run_timestamp": "2026-08-08T08:27:33.509Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Reka",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.012345539999999999,
      "avg_duration_per_task_ms": 657478.84,
      "avg_output_tokens_per_task": 14196.68,
      "last_run_timestamp": "2026-10-03T00:00:06.435Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "WandB",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.11157742000000001,
      "avg_duration_per_task_ms": 52237.66,
      "avg_output_tokens_per_task": 12874.88,
      "last_run_timestamp": "2026-07-18T08:03:41.319Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "GMICloud",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.26629643999999997,
      "avg_duration_per_task_ms": 438741.58,
      "avg_output_tokens_per_task": 27876.64,
      "last_run_timestamp": "2026-08-29T16:58:30.455Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "ModelRun",
      "accuracy": 0.773333,
      "accuracy_std_dev": 0.019999999999999962,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06652479,
      "avg_duration_per_task_ms": 59089.94,
      "avg_output_tokens_per_task": 12443.51,
      "last_run_timestamp": "2026-10-05T17:41:05.783Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02012318,
      "avg_duration_per_task_ms": 229027.46,
      "avg_output_tokens_per_task": 10940.04,
      "last_run_timestamp": "2026-07-18T08:04:46.147Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DeepInfra",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.042562220000000005,
      "avg_duration_per_task_ms": 321425.02,
      "avg_output_tokens_per_task": 8985.7,
      "last_run_timestamp": "2026-07-18T16:00:10.793Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "StreamLake",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.057549260000000005,
      "avg_duration_per_task_ms": 140710.34,
      "avg_output_tokens_per_task": 4743.94,
      "last_run_timestamp": "2026-09-11T04:09:34.110Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Azure",
      "accuracy": 0.7729411176470589,
      "accuracy_std_dev": 0.02923599978246472,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.5054271811764706,
      "avg_duration_per_task_ms": 128787.48470588235,
      "avg_output_tokens_per_task": 7945.302352941177,
      "last_run_timestamp": "2026-10-03T08:15:30.352Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Anthropic",
      "accuracy": 0.77277775,
      "accuracy_std_dev": 0.019333070721456368,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.4942197533333333,
      "avg_duration_per_task_ms": 128626.26333333334,
      "avg_output_tokens_per_task": 8065.555,
      "last_run_timestamp": "2026-10-03T08:20:22.745Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Alibaba",
      "accuracy": 0.7727776666666667,
      "accuracy_std_dev": 0.03324520758278134,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.11460699333333334,
      "avg_duration_per_task_ms": 172737.43666666668,
      "avg_output_tokens_per_task": 9388.596666666666,
      "last_run_timestamp": "2026-10-03T16:16:11.149Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Alibaba",
      "accuracy": 0.7725,
      "accuracy_std_dev": 0.021457321984581383,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.17566200249999997,
      "avg_duration_per_task_ms": 370370.8425,
      "avg_output_tokens_per_task": 27669.585,
      "last_run_timestamp": "2026-10-03T16:19:10.407Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "auto-routing",
      "accuracy": 0.772381,
      "accuracy_std_dev": 0.020909200094011936,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.17040614571428572,
      "avg_duration_per_task_ms": 747500.4285714285,
      "avg_output_tokens_per_task": 28014.845714285715,
      "last_run_timestamp": "2026-10-03T16:14:18.634Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7722222222222223,
      "accuracy_std_dev": 0.03095202181433487,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.03593436666666667,
      "avg_duration_per_task_ms": 446735.1666666667,
      "avg_output_tokens_per_task": 13815.786666666667,
      "last_run_timestamp": "2026-10-08T22:42:47.098Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7722221666666667,
      "accuracy_std_dev": 0.0225805089137857,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.009277203333333336,
      "avg_duration_per_task_ms": 211225.32333333333,
      "avg_output_tokens_per_task": 4258.733333333334,
      "last_run_timestamp": "2026-10-03T08:53:37.283Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "BaseTen",
      "accuracy": 0.772109,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 49,
      "avg_cost_per_task_usd": 0.06601828571428571,
      "avg_duration_per_task_ms": 84701.85714285714,
      "avg_output_tokens_per_task": 5801.673469387755,
      "last_run_timestamp": "2026-07-18T08:06:26.187Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Azure",
      "accuracy": 0.772,
      "accuracy_std_dev": 0.02679938639596063,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.808827636,
      "avg_duration_per_task_ms": 183544.5,
      "avg_output_tokens_per_task": 6940.7,
      "last_run_timestamp": "2026-10-01T08:03:00.737Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "StreamLake",
      "accuracy": 0.7718518888888889,
      "accuracy_std_dev": 0.027176331186630626,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.04377958222222222,
      "avg_duration_per_task_ms": 219884.14888888889,
      "avg_output_tokens_per_task": 9569.68888888889,
      "last_run_timestamp": "2026-10-03T16:15:21.561Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7718518333333333,
      "accuracy_std_dev": 0.03201514259473338,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.43984256000000005,
      "avg_duration_per_task_ms": 80156.35111111111,
      "avg_output_tokens_per_task": 5674.794444444445,
      "last_run_timestamp": "2026-10-03T08:11:53.173Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Novita",
      "accuracy": 0.77166675,
      "accuracy_std_dev": 0.027233420032884263,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.08857231500000001,
      "avg_duration_per_task_ms": 275121.275,
      "avg_output_tokens_per_task": 19864.29,
      "last_run_timestamp": "2026-10-03T08:38:07.639Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.77166675,
      "accuracy_std_dev": 0.02020715632362706,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.09734501,
      "avg_duration_per_task_ms": 129281.66,
      "avg_output_tokens_per_task": 5411.39,
      "last_run_timestamp": "2026-10-03T08:23:24.809Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Makora",
      "accuracy": 0.77166675,
      "accuracy_std_dev": 0.009860203404975986,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.11018350999999998,
      "avg_duration_per_task_ms": 71664.89,
      "avg_output_tokens_per_task": 4808.815,
      "last_run_timestamp": "2026-10-03T08:21:31.953Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.771258387755102,
      "accuracy_std_dev": 0.022567909478948184,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.06584808673469388,
      "avg_duration_per_task_ms": 190152.3418367347,
      "avg_output_tokens_per_task": 10271.474489795919,
      "last_run_timestamp": "2026-10-08T23:02:20.692Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7711112222222222,
      "accuracy_std_dev": 0.021773400922418973,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.5318819755555556,
      "avg_duration_per_task_ms": 174829.24,
      "avg_output_tokens_per_task": 10576.931111111111,
      "last_run_timestamp": "2026-10-03T08:10:31.433Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Friendli",
      "accuracy": 0.7711111666666667,
      "accuracy_std_dev": 0.016629684897963515,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06979356333333334,
      "avg_duration_per_task_ms": 35038.676666666666,
      "avg_output_tokens_per_task": 4742.786666666667,
      "last_run_timestamp": "2026-10-03T08:30:16.613Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.771111,
      "accuracy_std_dev": 0.04085521560339636,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.4149127733333333,
      "avg_duration_per_task_ms": 472514.2266666667,
      "avg_output_tokens_per_task": 36470.113333333335,
      "last_run_timestamp": "2026-10-01T16:00:45.071Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.77083325,
      "accuracy_std_dev": 0.028662434281433592,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0230241875,
      "avg_duration_per_task_ms": 301656.3375,
      "avg_output_tokens_per_task": 11912.9475,
      "last_run_timestamp": "2026-10-03T16:14:24.154Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Parasail",
      "accuracy": 0.7706935637583894,
      "accuracy_std_dev": 0.013644816256244217,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.09587820805369127,
      "avg_duration_per_task_ms": 288021.6644295302,
      "avg_output_tokens_per_task": 13114.979865771813,
      "last_run_timestamp": "2026-08-29T08:48:12.957Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Wafer",
      "accuracy": 0.7706668,
      "accuracy_std_dev": 0.010832125062055002,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.100171232,
      "avg_duration_per_task_ms": 273464.56,
      "avg_output_tokens_per_task": 22455.776,
      "last_run_timestamp": "2026-10-03T08:37:40.434Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "NextBit",
      "accuracy": 0.7706666,
      "accuracy_std_dev": 0.0204828947622156,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.10916058,
      "avg_duration_per_task_ms": 258128.42,
      "avg_output_tokens_per_task": 11359.708,
      "last_run_timestamp": "2026-10-03T16:07:05.245Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Parasail",
      "accuracy": 0.770580068825911,
      "accuracy_std_dev": 0.030947081374501206,
      "run_count": 5,
      "total_questions": 247,
      "avg_cost_per_task_usd": 0.08480364777327934,
      "avg_duration_per_task_ms": 268608.14574898785,
      "avg_output_tokens_per_task": 10039.971659919029,
      "last_run_timestamp": "2026-10-03T16:17:43.913Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Parasail",
      "accuracy": 0.7704762857142857,
      "accuracy_std_dev": 0.03627188525453795,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07425714857142857,
      "avg_duration_per_task_ms": 258403.66,
      "avg_output_tokens_per_task": 11728.06,
      "last_run_timestamp": "2026-10-03T08:42:46.753Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.770476,
      "accuracy_std_dev": 0.02332836876300735,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07879788857142857,
      "avg_duration_per_task_ms": 228548.80857142858,
      "avg_output_tokens_per_task": 9318.671428571428,
      "last_run_timestamp": "2026-08-29T16:11:58.076Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.7700001666666666,
      "accuracy_std_dev": 0.02168818589629438,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08111510333333333,
      "avg_duration_per_task_ms": 263647.49333333335,
      "avg_output_tokens_per_task": 13148.976666666667,
      "last_run_timestamp": "2026-10-03T08:16:12.045Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.18774403,
      "avg_duration_per_task_ms": 78191.08,
      "avg_output_tokens_per_task": 6339.04,
      "last_run_timestamp": "2026-10-03T08:11:23.532Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Baidu",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.018257509872652396,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.049748467500000004,
      "avg_duration_per_task_ms": 175035.1425,
      "avg_output_tokens_per_task": 11732.075,
      "last_run_timestamp": "2026-10-03T16:16:48.795Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "SiliconFlow",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.03986093883114146,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1172345275,
      "avg_duration_per_task_ms": 283399.5725,
      "avg_output_tokens_per_task": 12335.785,
      "last_run_timestamp": "2026-10-03T16:16:31.766Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Baidu",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.017950332712793934,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.10653246,
      "avg_duration_per_task_ms": 360324.605,
      "avg_output_tokens_per_task": 20634.68,
      "last_run_timestamp": "2026-10-03T08:06:28.791Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "AkashML",
      "accuracy": 0.7696969898989898,
      "accuracy_std_dev": 0.023999951853285028,
      "run_count": 10,
      "total_questions": 495,
      "avg_cost_per_task_usd": 0.05881589494949495,
      "avg_duration_per_task_ms": 254508.298989899,
      "avg_output_tokens_per_task": 11128.349494949494,
      "last_run_timestamp": "2026-10-09T15:53:11.694Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Azure",
      "accuracy": 0.7696295555555556,
      "accuracy_std_dev": 0.019654008621998214,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.4308894377777778,
      "avg_duration_per_task_ms": 100253.29111111112,
      "avg_output_tokens_per_task": 5601.184444444444,
      "last_run_timestamp": "2026-10-03T08:11:39.222Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Alibaba",
      "accuracy": 0.7696295555555556,
      "accuracy_std_dev": 0.027952376707182647,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.08778874,
      "avg_duration_per_task_ms": 191583.55333333334,
      "avg_output_tokens_per_task": 11553.07111111111,
      "last_run_timestamp": "2026-10-10T01:08:05.932Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "auto-routing",
      "accuracy": 0.7695237142857143,
      "accuracy_std_dev": 0.01326512827695538,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.10013912285714285,
      "avg_duration_per_task_ms": 333739.5742857143,
      "avg_output_tokens_per_task": 20222.222857142857,
      "last_run_timestamp": "2026-10-03T08:17:48.665Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7694445,
      "accuracy_std_dev": 0.02899006747864978,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.551655145,
      "avg_duration_per_task_ms": 183576.22666666665,
      "avg_output_tokens_per_task": 10538.358333333334,
      "last_run_timestamp": "2026-10-03T08:10:31.444Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DeepInfra",
      "accuracy": 0.7693334,
      "accuracy_std_dev": 0.027194796889110983,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.047570167999999996,
      "avg_duration_per_task_ms": 114696.804,
      "avg_output_tokens_per_task": 4956.572,
      "last_run_timestamp": "2026-10-08T23:21:59.850Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.7688886666666668,
      "accuracy_std_dev": 0.021998731726069008,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09783254,
      "avg_duration_per_task_ms": 102292.53333333334,
      "avg_output_tokens_per_task": 23933.02,
      "last_run_timestamp": "2026-10-01T19:04:42.463Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Io Net",
      "accuracy": 0.7688886666666668,
      "accuracy_std_dev": 0.011331263428624758,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0562409,
      "avg_duration_per_task_ms": 124558.21333333333,
      "avg_output_tokens_per_task": 4508.253333333333,
      "last_run_timestamp": "2026-09-26T08:24:47.398Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "AkashML",
      "accuracy": 0.768707,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 49,
      "avg_cost_per_task_usd": 0.18590030612244896,
      "avg_duration_per_task_ms": 362695.9183673469,
      "avg_output_tokens_per_task": 6540.102040816327,
      "last_run_timestamp": "2026-09-29T21:00:05.041Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.7684563154362417,
      "accuracy_std_dev": 0.006017520216464235,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.1581116510067114,
      "avg_duration_per_task_ms": 537705.5637583892,
      "avg_output_tokens_per_task": 10979.71812080537,
      "last_run_timestamp": "2026-08-08T08:27:49.545Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Cloudflare",
      "accuracy": 0.7683335,
      "accuracy_std_dev": 0.017873086533948198,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.14871092,
      "avg_duration_per_task_ms": 286909.625,
      "avg_output_tokens_per_task": 13559.755,
      "last_run_timestamp": "2026-10-03T16:12:29.902Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7683335,
      "accuracy_std_dev": 0.01190259075789804,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.15200416500000002,
      "avg_duration_per_task_ms": 286481.905,
      "avg_output_tokens_per_task": 19800.085,
      "last_run_timestamp": "2026-10-03T08:10:31.466Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "AtlasCloud",
      "accuracy": 0.768333375,
      "accuracy_std_dev": 0.028625995860133417,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0977644525,
      "avg_duration_per_task_ms": 232629.3425,
      "avg_output_tokens_per_task": 11722.16,
      "last_run_timestamp": "2026-10-03T16:04:46.698Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.76833325,
      "accuracy_std_dev": 0.023746444179866155,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.038813765,
      "avg_duration_per_task_ms": 338234.72,
      "avg_output_tokens_per_task": 12670.62,
      "last_run_timestamp": "2026-10-09T00:07:53.060Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "GMICloud",
      "accuracy": 0.767778,
      "accuracy_std_dev": 0.009557997105391226,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.055267166666666666,
      "avg_duration_per_task_ms": 921928.7333333333,
      "avg_output_tokens_per_task": 38360.306666666664,
      "last_run_timestamp": "2026-10-03T08:15:16.206Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7677777777777778,
      "accuracy_std_dev": 0.03175926811456523,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.5635850022222223,
      "avg_duration_per_task_ms": 180702.17222222223,
      "avg_output_tokens_per_task": 10553.803333333333,
      "last_run_timestamp": "2026-10-03T08:12:06.145Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cohere",
      "accuracy": 0.7673375906040268,
      "accuracy_std_dev": 0.01364674745945795,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.02802279865771812,
      "avg_duration_per_task_ms": 90972.77852348993,
      "avg_output_tokens_per_task": 17409.81879194631,
      "last_run_timestamp": "2026-10-03T08:20:48.319Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0121027,
      "avg_duration_per_task_ms": 95348.66,
      "avg_output_tokens_per_task": 10736.12,
      "last_run_timestamp": "2026-10-08T06:08:26.323Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "Novita",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0337152,
      "avg_duration_per_task_ms": 192245.9,
      "avg_output_tokens_per_task": 17702.16,
      "last_run_timestamp": "2026-09-01T12:41:31.535Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Reka",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.025223939999999997,
      "avg_duration_per_task_ms": 165137.1,
      "avg_output_tokens_per_task": 14240.82,
      "last_run_timestamp": "2026-09-15T00:00:05.846Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "ModelRun",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0668136,
      "avg_duration_per_task_ms": 100184.26,
      "avg_output_tokens_per_task": 11544.38,
      "last_run_timestamp": "2026-07-18T08:04:03.432Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Novita",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01215033,
      "avg_duration_per_task_ms": 551566.9,
      "avg_output_tokens_per_task": 23566.32,
      "last_run_timestamp": "2026-10-03T08:18:05.266Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Parasail",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0912162,
      "avg_duration_per_task_ms": 191018.42,
      "avg_output_tokens_per_task": 9260.76,
      "last_run_timestamp": "2026-08-04T03:12:10.013Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "NextBit",
      "accuracy": 0.76666675,
      "accuracy_std_dev": 0.024267078752653787,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.036940722499999995,
      "avg_duration_per_task_ms": 209470.17,
      "avg_output_tokens_per_task": 17358.2575,
      "last_run_timestamp": "2026-10-03T08:39:48.671Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "Google",
      "accuracy": 0.7666666666666667,
      "accuracy_std_dev": 0.014401440259764175,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.3460653,
      "avg_duration_per_task_ms": 170899.55333333334,
      "avg_output_tokens_per_task": 12611.153333333334,
      "last_run_timestamp": "2026-10-01T19:04:40.361Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Mistral",
      "accuracy": 0.7666666666666667,
      "accuracy_std_dev": 0.020727473271535585,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.06962397,
      "avg_duration_per_task_ms": 57268.18666666667,
      "avg_output_tokens_per_task": 4623.046666666667,
      "last_run_timestamp": "2026-10-03T08:43:42.470Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7666666153846154,
      "accuracy_std_dev": 0.020085283979679725,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.5291991723076925,
      "avg_duration_per_task_ms": 121619.19230769231,
      "avg_output_tokens_per_task": 8086.011538461538,
      "last_run_timestamp": "2026-10-03T08:16:24.268Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Parasail",
      "accuracy": 0.7666666,
      "accuracy_std_dev": 0.009428137559454682,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.081377272,
      "avg_duration_per_task_ms": 89133.128,
      "avg_output_tokens_per_task": 4543.572,
      "last_run_timestamp": "2026-10-03T08:30:16.616Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Decart",
      "accuracy": 0.7666666,
      "accuracy_std_dev": 0.018856133692780202,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.061894704,
      "avg_duration_per_task_ms": 48771.608,
      "avg_output_tokens_per_task": 4659.024,
      "last_run_timestamp": "2026-10-03T08:19:53.303Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "StreamLake",
      "accuracy": 0.7666665714285714,
      "accuracy_std_dev": 0.01951809905012813,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08535362285714285,
      "avg_duration_per_task_ms": 334395.88857142854,
      "avg_output_tokens_per_task": 19919.711428571427,
      "last_run_timestamp": "2026-10-03T08:40:08.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "OpenInference",
      "accuracy": 0.7661904285714286,
      "accuracy_std_dev": 0.01872337959311179,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.019808859999999998,
      "avg_duration_per_task_ms": 510174.85142857145,
      "avg_output_tokens_per_task": 13353.934285714286,
      "last_run_timestamp": "2026-08-29T16:47:05.637Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "auto-routing",
      "accuracy": 0.765833375,
      "accuracy_std_dev": 0.03471108632676851,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.04402259,
      "avg_duration_per_task_ms": 213134.3675,
      "avg_output_tokens_per_task": 9446.67,
      "last_run_timestamp": "2026-10-03T16:14:17.312Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Io Net",
      "accuracy": 0.7653059591836735,
      "accuracy_std_dev": 0.005416499999999991,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.02433298979591837,
      "avg_duration_per_task_ms": 682809.275510204,
      "avg_output_tokens_per_task": 13469.224489795919,
      "last_run_timestamp": "2026-08-01T08:05:58.651Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7651006510067114,
      "accuracy_std_dev": 0.02769017312403245,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.035872944630872476,
      "avg_duration_per_task_ms": 526054.0553691275,
      "avg_output_tokens_per_task": 12761.159395973154,
      "last_run_timestamp": "2026-10-08T22:51:57.050Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.765,
      "accuracy_std_dev": 0.036930670783239215,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.032793829999999996,
      "avg_duration_per_task_ms": 164299.645,
      "avg_output_tokens_per_task": 15398.3,
      "last_run_timestamp": "2026-10-03T08:19:00.261Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.765,
      "accuracy_std_dev": 0.028819341500110663,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.011177185,
      "avg_duration_per_task_ms": 267604.5325,
      "avg_output_tokens_per_task": 6349.9575,
      "last_run_timestamp": "2026-10-03T08:39:56.920Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Venice",
      "accuracy": 0.7647619999999999,
      "accuracy_std_dev": 0.022172301981139845,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.011946834285714285,
      "avg_duration_per_task_ms": 275656.4971428571,
      "avg_output_tokens_per_task": 13715.774285714286,
      "last_run_timestamp": "2026-10-03T16:20:06.645Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7646564924623115,
      "accuracy_std_dev": 0.022445600325773765,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.10723165075376884,
      "avg_duration_per_task_ms": 358580.18341708544,
      "avg_output_tokens_per_task": 11741.600502512563,
      "last_run_timestamp": "2026-10-03T16:15:36.234Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7646153076923077,
      "accuracy_std_dev": 0.017615963863166925,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.019299193846153844,
      "avg_duration_per_task_ms": 239187.3676923077,
      "avg_output_tokens_per_task": 17117.927692307694,
      "last_run_timestamp": "2026-10-03T08:37:24.250Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "auto-routing",
      "accuracy": 0.7644446666666667,
      "accuracy_std_dev": 0.020122943113658996,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.4976620088888889,
      "avg_duration_per_task_ms": 132296.16222222222,
      "avg_output_tokens_per_task": 8163.442222222222,
      "last_run_timestamp": "2026-10-03T08:17:09.269Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Novita",
      "accuracy": 0.7644445,
      "accuracy_std_dev": 0.021999102116147078,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06149954333333334,
      "avg_duration_per_task_ms": 120941.48333333334,
      "avg_output_tokens_per_task": 4754.513333333333,
      "last_run_timestamp": "2026-10-03T08:41:33.560Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Modal",
      "accuracy": 0.7644445,
      "accuracy_std_dev": 0.03281078940303024,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08256609000000001,
      "avg_duration_per_task_ms": 48194.92,
      "avg_output_tokens_per_task": 4847.44,
      "last_run_timestamp": "2026-10-03T08:29:04.086Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Relace",
      "accuracy": 0.7644443333333334,
      "accuracy_std_dev": 0.00831485358994506,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.029735113333333334,
      "avg_duration_per_task_ms": 54092.806666666664,
      "avg_output_tokens_per_task": 4732.093333333333,
      "last_run_timestamp": "2026-10-03T08:37:59.129Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "auto-routing",
      "accuracy": 0.7643097626262627,
      "accuracy_std_dev": 0.041231633865880205,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02870519191919192,
      "avg_duration_per_task_ms": 478260.4166666667,
      "avg_output_tokens_per_task": 12840.517676767677,
      "last_run_timestamp": "2026-10-03T16:10:32.037Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7641723061224489,
      "accuracy_std_dev": 0.03540326230730724,
      "run_count": 3,
      "total_questions": 147,
      "avg_cost_per_task_usd": 0.01049295918367347,
      "avg_duration_per_task_ms": 1253925.7823129252,
      "avg_output_tokens_per_task": 17278.809523809523,
      "last_run_timestamp": "2026-10-08T23:09:38.449Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7641127822580646,
      "accuracy_std_dev": 0.03914012491663765,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.10629601008064517,
      "avg_duration_per_task_ms": 864547.8709677419,
      "avg_output_tokens_per_task": 10411.459677419354,
      "last_run_timestamp": "2026-10-05T20:07:40.127Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "GMICloud",
      "accuracy": 0.7641026153846153,
      "accuracy_std_dev": 0.016021001176525294,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.017410869230769233,
      "avg_duration_per_task_ms": 197270.92,
      "avg_output_tokens_per_task": 19327.204615384617,
      "last_run_timestamp": "2026-10-03T08:29:06.422Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Inceptron",
      "accuracy": 0.764,
      "accuracy_std_dev": 0.02585427538338681,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.050899012,
      "avg_duration_per_task_ms": 154368.784,
      "avg_output_tokens_per_task": 4603.632,
      "last_run_timestamp": "2026-10-03T08:39:30.463Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.7633335,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01259681,
      "avg_duration_per_task_ms": 657803.07,
      "avg_output_tokens_per_task": 25182.87,
      "last_run_timestamp": "2026-10-03T08:05:49.905Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Relace",
      "accuracy": 0.763333,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02542324,
      "avg_duration_per_task_ms": 138965.85,
      "avg_output_tokens_per_task": 13863.01,
      "last_run_timestamp": "2026-10-03T16:12:10.916Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "Novita",
      "accuracy": 0.763333,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06322074000000001,
      "avg_duration_per_task_ms": 968065,
      "avg_output_tokens_per_task": 43847.04,
      "last_run_timestamp": "2026-10-03T08:20:05.772Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7631579766081871,
      "accuracy_std_dev": 0.021601269198350176,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.025081707602339184,
      "avg_duration_per_task_ms": 429534.44444444444,
      "avg_output_tokens_per_task": 12901.88596491228,
      "last_run_timestamp": "2026-10-03T16:09:19.685Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "NextBit",
      "accuracy": 0.7628571428571429,
      "accuracy_std_dev": 0.023328310442945685,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11986269428571429,
      "avg_duration_per_task_ms": 415507.14,
      "avg_output_tokens_per_task": 18962.734285714287,
      "last_run_timestamp": "2026-10-03T08:33:31.578Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7627686612903226,
      "accuracy_std_dev": 0.02051209766552413,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.02082460887096774,
      "avg_duration_per_task_ms": 113248.77822580645,
      "avg_output_tokens_per_task": 14257.657258064517,
      "last_run_timestamp": "2026-10-03T08:37:24.028Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7626666,
      "accuracy_std_dev": 0.03227647784749755,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.146569104,
      "avg_duration_per_task_ms": 373419.58,
      "avg_output_tokens_per_task": 24831.152,
      "last_run_timestamp": "2026-08-29T08:26:04.467Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Venice",
      "accuracy": 0.762500125,
      "accuracy_std_dev": 0.027474623393403872,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.12900198,
      "avg_duration_per_task_ms": 265844.8625,
      "avg_output_tokens_per_task": 11594.9925,
      "last_run_timestamp": "2026-10-03T16:12:29.878Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7625,
      "accuracy_std_dev": 0.03072062282490381,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.160654465,
      "avg_duration_per_task_ms": 223937.135,
      "avg_output_tokens_per_task": 9632.975,
      "last_run_timestamp": "2026-10-03T16:12:07.499Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Parasail",
      "accuracy": 0.762381,
      "accuracy_std_dev": 0.013179314517184024,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.12107822285714286,
      "avg_duration_per_task_ms": 467340.5485714286,
      "avg_output_tokens_per_task": 21773.494285714285,
      "last_run_timestamp": "2026-10-03T08:54:39.113Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "DeepInfra",
      "accuracy": 0.7622223333333333,
      "accuracy_std_dev": 0.030469494967188978,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.029655940000000006,
      "avg_duration_per_task_ms": 256385.88,
      "avg_output_tokens_per_task": 18856.086666666666,
      "last_run_timestamp": "2026-10-08T22:27:05.549Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7622223333333333,
      "accuracy_std_dev": 0.011330936570096763,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03153908666666666,
      "avg_duration_per_task_ms": 312202.32,
      "avg_output_tokens_per_task": 39820.306666666664,
      "last_run_timestamp": "2026-10-03T08:09:38.144Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Sail Research",
      "accuracy": 0.7621211818181818,
      "accuracy_std_dev": 0.026372448546363295,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.05366102181818182,
      "avg_duration_per_task_ms": 93039.47818181818,
      "avg_output_tokens_per_task": 4281.278181818182,
      "last_run_timestamp": "2026-10-03T08:43:08.437Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.7620649764309765,
      "accuracy_std_dev": 0.008722071727647192,
      "run_count": 6,
      "total_questions": 297,
      "avg_cost_per_task_usd": 0.007020558922558921,
      "avg_duration_per_task_ms": 240106.04377104377,
      "avg_output_tokens_per_task": 12275.050505050505,
      "last_run_timestamp": "2026-08-29T16:33:29.856Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Chutes",
      "accuracy": 0.7619049795918367,
      "accuracy_std_dev": 0.01847199999999999,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.1408940612244898,
      "avg_duration_per_task_ms": 345711.98979591834,
      "avg_output_tokens_per_task": 8276.57142857143,
      "last_run_timestamp": "2026-08-01T08:06:00.902Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7619048571428572,
      "accuracy_std_dev": 0.017353196366158268,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.01491982857142857,
      "avg_duration_per_task_ms": 401525.7,
      "avg_output_tokens_per_task": 15852.922857142858,
      "last_run_timestamp": "2026-10-03T16:17:41.694Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Together",
      "accuracy": 0.7619047142857143,
      "accuracy_std_dev": 0.020067932249624297,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.026417691428571432,
      "avg_duration_per_task_ms": 160543.33428571429,
      "avg_output_tokens_per_task": 8753.874285714286,
      "last_run_timestamp": "2026-10-03T16:17:22.242Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Chutes",
      "accuracy": 0.7619047142857143,
      "accuracy_std_dev": 0.03611523004698906,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07077762857142857,
      "avg_duration_per_task_ms": 309481.20857142855,
      "avg_output_tokens_per_task": 12537.965714285714,
      "last_run_timestamp": "2026-10-03T08:21:29.527Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Baidu",
      "accuracy": 0.761666625,
      "accuracy_std_dev": 0.015184060540065538,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.05742875749999999,
      "avg_duration_per_task_ms": 229440.2175,
      "avg_output_tokens_per_task": 10184.775,
      "last_run_timestamp": "2026-10-03T16:05:34.170Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DekaLLM",
      "accuracy": 0.7613333999999999,
      "accuracy_std_dev": 0.012219939993306055,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.059212192000000004,
      "avg_duration_per_task_ms": 137664.628,
      "avg_output_tokens_per_task": 14571.664,
      "last_run_timestamp": "2026-10-09T02:14:11.888Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Phala",
      "accuracy": 0.7613333999999999,
      "accuracy_std_dev": 0.019955635240202208,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.054660096000000005,
      "avg_duration_per_task_ms": 124144.4,
      "avg_output_tokens_per_task": 4448.496,
      "last_run_timestamp": "2026-10-03T08:45:16.467Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "auto-routing",
      "accuracy": 0.7613333999999999,
      "accuracy_std_dev": 0.022469788112930676,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.06195499599999999,
      "avg_duration_per_task_ms": 143749.252,
      "avg_output_tokens_per_task": 6410.492,
      "last_run_timestamp": "2026-10-03T08:22:27.259Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "auto-routing",
      "accuracy": 0.7613332,
      "accuracy_std_dev": 0.026297374605842312,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.167040694,
      "avg_duration_per_task_ms": 594740.238,
      "avg_output_tokens_per_task": 37349.044,
      "last_run_timestamp": "2026-10-03T08:17:06.941Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Parasail",
      "accuracy": 0.7611113333333333,
      "accuracy_std_dev": 0.025288354320165317,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.17373087333333334,
      "avg_duration_per_task_ms": 477933.7,
      "avg_output_tokens_per_task": 24266.54,
      "last_run_timestamp": "2026-08-29T17:19:03.369Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Google",
      "accuracy": 0.7610256923076923,
      "accuracy_std_dev": 0.021218684472687738,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.589651923076923,
      "avg_duration_per_task_ms": 120803.91923076923,
      "avg_output_tokens_per_task": 8142.596923076923,
      "last_run_timestamp": "2026-10-03T08:38:25.156Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7609522857142857,
      "accuracy_std_dev": 0.0335097115932266,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.43242590285714283,
      "avg_duration_per_task_ms": 94424.92,
      "avg_output_tokens_per_task": 5553.762857142857,
      "last_run_timestamp": "2026-10-03T08:20:09.018Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Google",
      "accuracy": 0.7604938148148149,
      "accuracy_std_dev": 0.020602075859156715,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.656406777037037,
      "avg_duration_per_task_ms": 73567.76444444444,
      "avg_output_tokens_per_task": 5646.465185185185,
      "last_run_timestamp": "2026-10-03T08:42:35.471Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Alibaba",
      "accuracy": 0.7600001249999999,
      "accuracy_std_dev": 0.03349974295213884,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0116075025,
      "avg_duration_per_task_ms": 155751.2425,
      "avg_output_tokens_per_task": 14262.745,
      "last_run_timestamp": "2026-10-03T16:12:33.521Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.02735228340559523,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.5190067183333332,
      "avg_duration_per_task_ms": 221047.60166666665,
      "avg_output_tokens_per_task": 10200.993333333334,
      "last_run_timestamp": "2026-10-03T08:20:09.019Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01747962,
      "avg_duration_per_task_ms": 589736.28,
      "avg_output_tokens_per_task": 13675.16,
      "last_run_timestamp": "2026-08-01T08:06:01.204Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "StreamLake",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.020870148806369355,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.0067541,
      "avg_duration_per_task_ms": 260092.14,
      "avg_output_tokens_per_task": 15338.358,
      "last_run_timestamp": "2026-10-03T08:38:03.907Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.01333300000000004,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.054461335,
      "avg_duration_per_task_ms": 91417,
      "avg_output_tokens_per_task": 15101.33,
      "last_run_timestamp": "2026-10-03T08:13:12.597Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Novita",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.01817028797633262,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08873829714285716,
      "avg_duration_per_task_ms": 247182.05714285714,
      "avg_output_tokens_per_task": 16018.762857142858,
      "last_run_timestamp": "2026-10-03T16:22:04.066Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Azure",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.013984245106547586,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.16696927200000003,
      "avg_duration_per_task_ms": 223419.404,
      "avg_output_tokens_per_task": 11293.668,
      "last_run_timestamp": "2026-10-03T16:10:46.857Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Relace",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.026666999999999996,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12531742999999998,
      "avg_duration_per_task_ms": 312784.77,
      "avg_output_tokens_per_task": 20756.6,
      "last_run_timestamp": "2026-10-03T08:29:42.404Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Decart",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.01333300000000004,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.015769889999999998,
      "avg_duration_per_task_ms": 1152795.83,
      "avg_output_tokens_per_task": 36264.12,
      "last_run_timestamp": "2026-10-03T08:16:24.271Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "GLM 5.3",
      "provider_name": "digitalocean",
      "accuracy": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01597332,
      "avg_duration_per_task_ms": 42256.58,
      "avg_output_tokens_per_task": 1764.66,
      "last_run_timestamp": "2026-10-01T23:50:45.025Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "GMICloud",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.019626323972325197,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02215426,
      "avg_duration_per_task_ms": 122620.84,
      "avg_output_tokens_per_task": 8751.653333333334,
      "last_run_timestamp": "2026-10-03T16:07:50.330Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Novita",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.020470931774452417,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.09495435142857143,
      "avg_duration_per_task_ms": 338628.4542857143,
      "avg_output_tokens_per_task": 12897.748571428572,
      "last_run_timestamp": "2026-10-03T16:13:59.491Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.30272262,
      "avg_duration_per_task_ms": 355584.56,
      "avg_output_tokens_per_task": 10613.82,
      "last_run_timestamp": "2026-10-03T18:42:35.851Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "Venice",
      "accuracy": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.19103682,
      "avg_duration_per_task_ms": 233230.24,
      "avg_output_tokens_per_task": 15236.66,
      "last_run_timestamp": "2026-08-04T03:12:00.236Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "Novita",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.01333300000000004,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07523408,
      "avg_duration_per_task_ms": 676626.08,
      "avg_output_tokens_per_task": 16285.62,
      "last_run_timestamp": "2026-10-03T08:15:42.490Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Makora",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07038507,
      "avg_duration_per_task_ms": 256302.49,
      "avg_output_tokens_per_task": 10045.27,
      "last_run_timestamp": "2026-08-26T00:00:05.638Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Baidu",
      "accuracy": 0.76,
      "accuracy_std_dev": 1.1102230246251565e-16,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05899852666666667,
      "avg_duration_per_task_ms": 72162.04666666666,
      "avg_output_tokens_per_task": 4660.633333333333,
      "last_run_timestamp": "2026-10-03T08:13:12.618Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.020547857418718854,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.011870755,
      "avg_duration_per_task_ms": 198678.775,
      "avg_output_tokens_per_task": 4474.35,
      "last_run_timestamp": "2026-10-03T08:39:24.758Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "InferenceNet",
      "accuracy": 0.7599998333333334,
      "accuracy_std_dev": 0.01333333333437501,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03796188999999999,
      "avg_duration_per_task_ms": 83545.53666666667,
      "avg_output_tokens_per_task": 4558.626666666667,
      "last_run_timestamp": "2026-10-03T08:35:08.461Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7597841862348177,
      "accuracy_std_dev": 0.0190545430278451,
      "run_count": 5,
      "total_questions": 247,
      "avg_cost_per_task_usd": 0.164671975708502,
      "avg_duration_per_task_ms": 858653.5303643724,
      "avg_output_tokens_per_task": 31997.372469635626,
      "last_run_timestamp": "2026-10-09T00:08:43.362Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.7595079060402684,
      "accuracy_std_dev": 0.0216087995861768,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.1067814966442953,
      "avg_duration_per_task_ms": 41164.906040268455,
      "avg_output_tokens_per_task": 4181.802013422819,
      "last_run_timestamp": "2026-10-03T21:42:23.046Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Sail Research",
      "accuracy": 0.7594443333333334,
      "accuracy_std_dev": 0.015265178644731146,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.008844093333333332,
      "avg_duration_per_task_ms": 361189.67333333334,
      "avg_output_tokens_per_task": 14258.985,
      "last_run_timestamp": "2026-10-03T08:39:28.580Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.7593333,
      "accuracy_std_dev": 0.016452043441773433,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.083344808,
      "avg_duration_per_task_ms": 106115.21,
      "avg_output_tokens_per_task": 23741.31,
      "last_run_timestamp": "2026-10-01T19:05:47.886Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Reka",
      "accuracy": 0.7592593333333333,
      "accuracy_std_dev": 0.02381917591633544,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.050810793333333326,
      "avg_duration_per_task_ms": 122662.1288888889,
      "avg_output_tokens_per_task": 4548.155555555555,
      "last_run_timestamp": "2026-10-04T12:00:02.814Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "auto-routing",
      "accuracy": 0.7592592222222222,
      "accuracy_std_dev": 0.020714267497322178,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.41923416888888887,
      "avg_duration_per_task_ms": 97992.62444444445,
      "avg_output_tokens_per_task": 5633.653333333334,
      "last_run_timestamp": "2026-10-03T08:15:30.343Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "Google",
      "accuracy": 0.7590477142857143,
      "accuracy_std_dev": 0.013997181448055748,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.24529245999999996,
      "avg_duration_per_task_ms": 151053.55142857143,
      "avg_output_tokens_per_task": 23224.54,
      "last_run_timestamp": "2026-10-01T19:04:58.090Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "CoreWeave",
      "accuracy": 0.7590477142857143,
      "accuracy_std_dev": 0.034992667843071215,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08880086285714285,
      "avg_duration_per_task_ms": 68112.38857142857,
      "avg_output_tokens_per_task": 12261.222857142857,
      "last_run_timestamp": "2026-10-03T16:15:11.803Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Cloudflare",
      "accuracy": 0.7588384343434343,
      "accuracy_std_dev": 0.024611570108132058,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20684339646464647,
      "avg_duration_per_task_ms": 493166.0580808081,
      "avg_output_tokens_per_task": 20296.997474747473,
      "last_run_timestamp": "2026-10-03T08:19:53.308Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7586668,
      "accuracy_std_dev": 0.01654634971708262,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.071547176,
      "avg_duration_per_task_ms": 127553.92,
      "avg_output_tokens_per_task": 4587.528,
      "last_run_timestamp": "2026-10-03T09:01:39.121Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.7585565223214286,
      "accuracy_std_dev": 0.0371971128592224,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.19265237276785713,
      "avg_duration_per_task_ms": 474084.81696428574,
      "avg_output_tokens_per_task": 27864.78348214286,
      "last_run_timestamp": "2026-10-03T16:15:58.694Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.758333375,
      "accuracy_std_dev": 0.019364862941275235,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0383828675,
      "avg_duration_per_task_ms": 246490.5475,
      "avg_output_tokens_per_task": 12250.2775,
      "last_run_timestamp": "2026-10-03T16:07:44.182Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Phala",
      "accuracy": 0.758333375,
      "accuracy_std_dev": 0.012360224066916245,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.034803532500000005,
      "avg_duration_per_task_ms": 275650.13,
      "avg_output_tokens_per_task": 13717.48,
      "last_run_timestamp": "2026-10-03T08:20:09.018Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7583332500000001,
      "accuracy_std_dev": 0.025110926131616508,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.021590805,
      "avg_duration_per_task_ms": 448620.805,
      "avg_output_tokens_per_task": 13949.42,
      "last_run_timestamp": "2026-10-05T20:08:33.293Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "OpenAI",
      "accuracy": 0.7583332500000001,
      "accuracy_std_dev": 0.005527984369324845,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.40697005500000005,
      "avg_duration_per_task_ms": 402356.405,
      "avg_output_tokens_per_task": 35595.7,
      "last_run_timestamp": "2026-10-02T23:12:57.355Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Crusoe",
      "accuracy": 0.7583331666666667,
      "accuracy_std_dev": 0.013844473427047428,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09591996333333334,
      "avg_duration_per_task_ms": 214673.9,
      "avg_output_tokens_per_task": 9689.54,
      "last_run_timestamp": "2026-08-29T16:31:41.554Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DekaLLM",
      "accuracy": 0.7583329999999999,
      "accuracy_std_dev": 0.04499999999999999,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06691538,
      "avg_duration_per_task_ms": 350031.58,
      "avg_output_tokens_per_task": 54823.52,
      "last_run_timestamp": "2026-10-09T03:50:20.056Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7580951428571429,
      "accuracy_std_dev": 0.025125616384357164,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.5482415914285714,
      "avg_duration_per_task_ms": 178558.36285714287,
      "avg_output_tokens_per_task": 10788.502857142857,
      "last_run_timestamp": "2026-10-03T08:08:31.999Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Parasail",
      "accuracy": 0.7580951428571429,
      "accuracy_std_dev": 0.05160461492216857,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.02014420857142857,
      "avg_duration_per_task_ms": 210702.71142857143,
      "avg_output_tokens_per_task": 12469.117142857143,
      "last_run_timestamp": "2026-10-03T16:19:59.478Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.758,
      "accuracy_std_dev": 0.03709447398198281,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.079319272,
      "avg_duration_per_task_ms": 620750.224,
      "avg_output_tokens_per_task": 10895.356,
      "last_run_timestamp": "2026-08-29T17:19:42.622Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7579964141414142,
      "accuracy_std_dev": 0.023533015514113357,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20713992929292932,
      "avg_duration_per_task_ms": 664979.7727272727,
      "avg_output_tokens_per_task": 53099.67929292929,
      "last_run_timestamp": "2026-09-26T08:05:12.736Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Novita",
      "accuracy": 0.7577778333333334,
      "accuracy_std_dev": 0.024545289459260582,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.026035279999999997,
      "avg_duration_per_task_ms": 169995.66166666665,
      "avg_output_tokens_per_task": 18134.881666666668,
      "last_run_timestamp": "2026-10-03T08:33:35.373Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "Google AI Studio",
      "accuracy": 0.7577776666666667,
      "accuracy_std_dev": 0.04629468145358481,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12420428666666666,
      "avg_duration_per_task_ms": 116279.86,
      "avg_output_tokens_per_task": 19029.046666666665,
      "last_run_timestamp": "2026-10-01T19:05:00.104Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "auto-routing",
      "accuracy": 0.7577776666666667,
      "accuracy_std_dev": 0.017497908167041637,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.35916456,
      "avg_duration_per_task_ms": 165675.65333333332,
      "avg_output_tokens_per_task": 12758.866666666667,
      "last_run_timestamp": "2026-10-01T19:06:01.702Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "InferenceNet",
      "accuracy": 0.757619,
      "accuracy_std_dev": 0.014553057204587645,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.3625878914285714,
      "avg_duration_per_task_ms": 337462.2485714286,
      "avg_output_tokens_per_task": 10492.708571428571,
      "last_run_timestamp": "2026-10-03T08:37:29.554Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Novita",
      "accuracy": 0.757499875,
      "accuracy_std_dev": 0.03298787550918331,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.08512432500000001,
      "avg_duration_per_task_ms": 226499.755,
      "avg_output_tokens_per_task": 9952.425,
      "last_run_timestamp": "2026-10-03T16:09:05.689Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "NextBit",
      "accuracy": 0.7573333999999999,
      "accuracy_std_dev": 0.03991088294989224,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.023026752,
      "avg_duration_per_task_ms": 161149.652,
      "avg_output_tokens_per_task": 18473.992,
      "last_run_timestamp": "2026-10-03T08:55:31.322Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Io Net",
      "accuracy": 0.7572707181208054,
      "accuracy_std_dev": 0.01805706983120156,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.00953123489932886,
      "avg_duration_per_task_ms": 371076.8926174497,
      "avg_output_tokens_per_task": 15024.973154362417,
      "last_run_timestamp": "2026-10-05T08:35:06.671Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7571428571428571,
      "accuracy_std_dev": 0.03263217671225126,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.01064094,
      "avg_duration_per_task_ms": 1272417.7228571428,
      "avg_output_tokens_per_task": 12476.905714285715,
      "last_run_timestamp": "2026-10-05T20:14:35.950Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Google",
      "accuracy": 0.757037,
      "accuracy_std_dev": 0.028822348820933333,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.5493601955555556,
      "avg_duration_per_task_ms": 192839.5377777778,
      "avg_output_tokens_per_task": 10676.25111111111,
      "last_run_timestamp": "2026-10-03T08:20:09.018Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7566669999999999,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.17767661,
      "avg_duration_per_task_ms": 84052.21,
      "avg_output_tokens_per_task": 6226.55,
      "last_run_timestamp": "2026-10-03T08:17:58.907Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7566669999999999,
      "accuracy_std_dev": 0.030000000000000027,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02875953,
      "avg_duration_per_task_ms": 155469.3,
      "avg_output_tokens_per_task": 15776.48,
      "last_run_timestamp": "2026-10-03T08:38:11.307Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Reka",
      "accuracy": 0.756525067114094,
      "accuracy_std_dev": 0.028844409691384666,
      "run_count": 9,
      "total_questions": 447,
      "avg_cost_per_task_usd": 0.0700060134228188,
      "avg_duration_per_task_ms": 265527.27293064876,
      "avg_output_tokens_per_task": 11642.608501118568,
      "last_run_timestamp": "2026-10-03T13:30:10.047Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Alibaba",
      "accuracy": 0.7564103277591973,
      "accuracy_std_dev": 0.017294806587206737,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.0331466856187291,
      "avg_duration_per_task_ms": 189272.61872909698,
      "avg_output_tokens_per_task": 17663.859531772574,
      "last_run_timestamp": "2026-10-09T21:37:39.384Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7563351081871345,
      "accuracy_std_dev": 0.013338594901982123,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.033182684210526316,
      "avg_duration_per_task_ms": 2146298.769005848,
      "avg_output_tokens_per_task": 50213.23391812865,
      "last_run_timestamp": "2026-10-09T04:05:56.641Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7561905714285714,
      "accuracy_std_dev": 0.0243929472000011,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.1949313942857143,
      "avg_duration_per_task_ms": 124043.37428571428,
      "avg_output_tokens_per_task": 7684.874285714286,
      "last_run_timestamp": "2026-10-03T08:17:09.273Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7561538461538462,
      "accuracy_std_dev": 0.031169392555582855,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.2008702830769231,
      "avg_duration_per_task_ms": 120980.52153846154,
      "avg_output_tokens_per_task": 7797.5623076923075,
      "last_run_timestamp": "2026-10-03T08:16:37.922Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.756,
      "accuracy_std_dev": 0.02048314212224288,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.022952103999999997,
      "avg_duration_per_task_ms": 125535.02,
      "avg_output_tokens_per_task": 15581.832,
      "last_run_timestamp": "2026-10-03T08:16:37.919Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.756,
      "accuracy_std_dev": 0.021745967699782896,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.006672712000000001,
      "avg_duration_per_task_ms": 144287.16,
      "avg_output_tokens_per_task": 4265.224,
      "last_run_timestamp": "2026-10-03T08:12:06.166Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "BaseTen",
      "accuracy": 0.7557079178082191,
      "accuracy_std_dev": 0.01607886717264481,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.2608585205479452,
      "avg_duration_per_task_ms": 125025.47260273973,
      "avg_output_tokens_per_task": 7152.054794520548,
      "last_run_timestamp": "2026-10-03T08:05:09.514Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "DeepInfra",
      "accuracy": 0.7556891325301206,
      "accuracy_std_dev": 0.028686868714448425,
      "run_count": 5,
      "total_questions": 249,
      "avg_cost_per_task_usd": 0.05997379919678716,
      "avg_duration_per_task_ms": 208407.72289156626,
      "avg_output_tokens_per_task": 7587.301204819277,
      "last_run_timestamp": "2026-08-15T08:17:17.868Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "ModelRun",
      "accuracy": 0.7556161712204007,
      "accuracy_std_dev": 0.028091014564674013,
      "run_count": 11,
      "total_questions": 549,
      "avg_cost_per_task_usd": 0.08449569216757742,
      "avg_duration_per_task_ms": 138097.57559198543,
      "avg_output_tokens_per_task": 12733.64116575592,
      "last_run_timestamp": "2026-10-05T17:44:38.668Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7555803571428571,
      "accuracy_std_dev": 0.023882484930784247,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.044577171874999995,
      "avg_duration_per_task_ms": 507800.79910714284,
      "avg_output_tokens_per_task": 22983.540178571428,
      "last_run_timestamp": "2026-10-09T00:21:32.259Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Morph",
      "accuracy": 0.7555556666666666,
      "accuracy_std_dev": 0.00831485358994506,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.027460826666666667,
      "avg_duration_per_task_ms": 642340.9866666667,
      "avg_output_tokens_per_task": 17001.633333333335,
      "last_run_timestamp": "2026-08-08T08:29:47.435Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Azure",
      "accuracy": 0.7555555555555555,
      "accuracy_std_dev": 0.02266234163693451,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.34417582888888887,
      "avg_duration_per_task_ms": 219931.23555555556,
      "avg_output_tokens_per_task": 11911.208888888888,
      "last_run_timestamp": "2026-10-03T08:23:27.050Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Relace",
      "accuracy": 0.7555555,
      "accuracy_std_dev": 0.03119039105103792,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.0069151533333333326,
      "avg_duration_per_task_ms": 128985.71333333333,
      "avg_output_tokens_per_task": 4275.633333333333,
      "last_run_timestamp": "2026-10-03T08:51:15.407Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Venice",
      "accuracy": 0.75541675,
      "accuracy_std_dev": 0.033289157186499954,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0732536275,
      "avg_duration_per_task_ms": 407903.865,
      "avg_output_tokens_per_task": 9228.4025,
      "last_run_timestamp": "2026-10-03T16:19:32.420Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7552381428571429,
      "accuracy_std_dev": 0.032212325637550464,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07236689428571429,
      "avg_duration_per_task_ms": 290760.98,
      "avg_output_tokens_per_task": 11419.182857142858,
      "last_run_timestamp": "2026-10-03T16:21:33.429Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Morph",
      "accuracy": 0.7552081741071428,
      "accuracy_std_dev": 0.03871863477403911,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.12397573883928571,
      "avg_duration_per_task_ms": 416217.65178571426,
      "avg_output_tokens_per_task": 20474.676339285714,
      "last_run_timestamp": "2026-10-03T16:14:16.547Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "StreamLake",
      "accuracy": 0.755050595959596,
      "accuracy_std_dev": 0.048907866785416025,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07672445454545455,
      "avg_duration_per_task_ms": 362896.6767676768,
      "avg_output_tokens_per_task": 12711.555555555555,
      "last_run_timestamp": "2026-10-03T16:20:30.141Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Wafer",
      "accuracy": 0.755,
      "accuracy_std_dev": 0.009860245661239886,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07160828000000001,
      "avg_duration_per_task_ms": 331234.47,
      "avg_output_tokens_per_task": 13566.125,
      "last_run_timestamp": "2026-10-03T08:43:45.086Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.7549997500000001,
      "accuracy_std_dev": 0.022791366543880153,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0787181275,
      "avg_duration_per_task_ms": 246265.255,
      "avg_output_tokens_per_task": 9770.0025,
      "last_run_timestamp": "2026-10-03T16:07:50.332Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Parasail",
      "accuracy": 0.7549019380804953,
      "accuracy_std_dev": 0.022500075722423546,
      "run_count": 13,
      "total_questions": 646,
      "avg_cost_per_task_usd": 0.020211260061919506,
      "avg_duration_per_task_ms": 487189.79102167184,
      "avg_output_tokens_per_task": 17152.354489164085,
      "last_run_timestamp": "2026-10-03T08:34:10.987Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.7548035989974938,
      "accuracy_std_dev": 0.028446815979286225,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.0686145037593985,
      "avg_duration_per_task_ms": 407411.1979949875,
      "avg_output_tokens_per_task": 10663.82456140351,
      "last_run_timestamp": "2026-10-08T23:43:42.862Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "auto-routing",
      "accuracy": 0.7546668,
      "accuracy_std_dev": 0.012927338819726197,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.19784133199999998,
      "avg_duration_per_task_ms": 328733.652,
      "avg_output_tokens_per_task": 12530.504,
      "last_run_timestamp": "2026-10-03T08:11:26.748Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Azure",
      "accuracy": 0.7546666,
      "accuracy_std_dev": 0.0308113572151569,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.025647999999999997,
      "avg_duration_per_task_ms": 294521.988,
      "avg_output_tokens_per_task": 13284.328,
      "last_run_timestamp": "2026-10-03T16:12:20.035Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Azure",
      "accuracy": 0.7545098823529411,
      "accuracy_std_dev": 0.035662554927909464,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.19809168117647058,
      "avg_duration_per_task_ms": 103146.24235294118,
      "avg_output_tokens_per_task": 7666.7011764705885,
      "last_run_timestamp": "2026-10-03T08:20:09.021Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Venice",
      "accuracy": 0.7545043243243243,
      "accuracy_std_dev": 0.0263378748741976,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.023569141891891895,
      "avg_duration_per_task_ms": 316460.4864864865,
      "avg_output_tokens_per_task": 4132.989864864865,
      "last_run_timestamp": "2026-10-03T08:34:09.439Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Novita",
      "accuracy": 0.7544445,
      "accuracy_std_dev": 0.02678215501504689,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08861642333333332,
      "avg_duration_per_task_ms": 293005.13333333336,
      "avg_output_tokens_per_task": 12023.27,
      "last_run_timestamp": "2026-10-03T08:37:04.257Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.7544445,
      "accuracy_std_dev": 0.025067718163340406,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.008447913333333333,
      "avg_duration_per_task_ms": 189202.61333333334,
      "avg_output_tokens_per_task": 4437.176666666666,
      "last_run_timestamp": "2026-10-03T08:32:50.838Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.7544443333333334,
      "accuracy_std_dev": 0.0343008332341294,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.009552396666666668,
      "avg_duration_per_task_ms": 212261.59,
      "avg_output_tokens_per_task": 15474.373333333333,
      "last_run_timestamp": "2026-10-03T16:05:34.199Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Z.AI",
      "accuracy": 0.7544443333333334,
      "accuracy_std_dev": 0.02786659578340269,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.008521756666666666,
      "avg_duration_per_task_ms": 190718.40666666668,
      "avg_output_tokens_per_task": 4439.933333333333,
      "last_run_timestamp": "2026-10-03T08:37:24.517Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Ionstream",
      "accuracy": 0.7542953608247424,
      "accuracy_std_dev": 0.02652500000000002,
      "run_count": 2,
      "total_questions": 97,
      "avg_cost_per_task_usd": 0.021299783505154636,
      "avg_duration_per_task_ms": 633495.5257731959,
      "avg_output_tokens_per_task": 26267.113402061856,
      "last_run_timestamp": "2026-08-08T08:30:53.738Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Novita",
      "accuracy": 0.75416675,
      "accuracy_std_dev": 0.03169964729989123,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.014154627500000003,
      "avg_duration_per_task_ms": 169480.67,
      "avg_output_tokens_per_task": 14972.985,
      "last_run_timestamp": "2026-10-03T16:15:22.012Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Phala",
      "accuracy": 0.75416675,
      "accuracy_std_dev": 0.012219036835507957,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1638202875,
      "avg_duration_per_task_ms": 453197.5425,
      "avg_output_tokens_per_task": 9528.3525,
      "last_run_timestamp": "2026-10-03T16:16:10.656Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Anthropic",
      "accuracy": 0.754074,
      "accuracy_std_dev": 0.039182303332556186,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.19633914666666663,
      "avg_duration_per_task_ms": 104140.85111111111,
      "avg_output_tokens_per_task": 7619.131111111111,
      "last_run_timestamp": "2026-10-03T08:06:43.541Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "CoreWeave",
      "accuracy": 0.754,
      "accuracy_std_dev": 0.021385259722528513,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.022165832,
      "avg_duration_per_task_ms": 128895.106,
      "avg_output_tokens_per_task": 17780.99,
      "last_run_timestamp": "2026-10-03T08:11:39.228Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Morph",
      "accuracy": 0.753914899328859,
      "accuracy_std_dev": 0.040895046741628756,
      "run_count": 9,
      "total_questions": 447,
      "avg_cost_per_task_usd": 0.07723322147651007,
      "avg_duration_per_task_ms": 229349.62863534674,
      "avg_output_tokens_per_task": 14123.306487695749,
      "last_run_timestamp": "2026-10-03T08:50:19.608Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Together",
      "accuracy": 0.7538975456570156,
      "accuracy_std_dev": 0.023333857993043206,
      "run_count": 9,
      "total_questions": 449,
      "avg_cost_per_task_usd": 0.013686612472160357,
      "avg_duration_per_task_ms": 214524.49888641425,
      "avg_output_tokens_per_task": 16219.260579064588,
      "last_run_timestamp": "2026-10-03T08:34:35.934Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cloudflare",
      "accuracy": 0.7538888333333333,
      "accuracy_std_dev": 0.019140560322838583,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.025353155,
      "avg_duration_per_task_ms": 294929.6716666667,
      "avg_output_tokens_per_task": 17480.86,
      "last_run_timestamp": "2026-10-03T08:34:34.719Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Relace",
      "accuracy": 0.7533335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06322421,
      "avg_duration_per_task_ms": 228589.91,
      "avg_output_tokens_per_task": 10641.07,
      "last_run_timestamp": "2026-10-03T16:12:33.507Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Phala",
      "accuracy": 0.7533335,
      "accuracy_std_dev": 0.01805358280000584,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.15314664666666666,
      "avg_duration_per_task_ms": 312626.37,
      "avg_output_tokens_per_task": 20191.006666666668,
      "last_run_timestamp": "2026-10-03T08:43:11.974Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DekaLLM",
      "accuracy": 0.7533335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01376083,
      "avg_duration_per_task_ms": 347095.91,
      "avg_output_tokens_per_task": 14466.76,
      "last_run_timestamp": "2026-10-09T02:54:47.806Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.7533335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.33087047999999997,
      "avg_duration_per_task_ms": 241177.17,
      "avg_output_tokens_per_task": 11982.77,
      "last_run_timestamp": "2026-10-01T16:01:02.676Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Venice",
      "accuracy": 0.7533333999999999,
      "accuracy_std_dev": 0.015776240928687683,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.138680804,
      "avg_duration_per_task_ms": 204157.96,
      "avg_output_tokens_per_task": 8340.072,
      "last_run_timestamp": "2026-10-03T08:52:57.109Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Azure",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.027755626724844277,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.5350316133333334,
      "avg_duration_per_task_ms": 188978.14666666667,
      "avg_output_tokens_per_task": 10298.875555555556,
      "last_run_timestamp": "2026-10-03T08:09:23.860Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.02739730544739352,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.19939614444444445,
      "avg_duration_per_task_ms": 129460.44888888889,
      "avg_output_tokens_per_task": 7850.873333333333,
      "last_run_timestamp": "2026-10-03T08:12:20.019Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "Google",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.028803034820811656,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.11451421333333334,
      "avg_duration_per_task_ms": 146387.56666666668,
      "avg_output_tokens_per_task": 17546.113333333335,
      "last_run_timestamp": "2026-10-01T19:05:47.885Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Morph",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.014401903168987377,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.021082753333333332,
      "avg_duration_per_task_ms": 915986.2133333334,
      "avg_output_tokens_per_task": 13268.026666666667,
      "last_run_timestamp": "2026-08-15T08:15:52.453Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "auto-routing",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.29158895333333334,
      "avg_duration_per_task_ms": 171961.5,
      "avg_output_tokens_per_task": 10696.2,
      "last_run_timestamp": "2026-10-01T19:04:36.169Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.018856180831641284,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06088320666666667,
      "avg_duration_per_task_ms": 1112758.12,
      "avg_output_tokens_per_task": 52398.04,
      "last_run_timestamp": "2026-10-03T16:14:15.492Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Wafer",
      "accuracy": 0.7533333,
      "accuracy_std_dev": 0.023475765282733617,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.021858692000000002,
      "avg_duration_per_task_ms": 119951.828,
      "avg_output_tokens_per_task": 15133.982,
      "last_run_timestamp": "2026-10-03T08:40:00.936Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Venice",
      "accuracy": 0.7533332500000001,
      "accuracy_std_dev": 0.014907082584714572,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.243283165,
      "avg_duration_per_task_ms": 338011.175,
      "avg_output_tokens_per_task": 17182.585,
      "last_run_timestamp": "2026-10-03T08:32:03.359Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.012521599999999999,
      "avg_duration_per_task_ms": 89897.54,
      "avg_output_tokens_per_task": 11032.16,
      "last_run_timestamp": "2026-10-08T06:07:32.522Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "auto-routing",
      "accuracy": 0.753333,
      "accuracy_std_dev": 0.032659863237109024,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12209446666666665,
      "avg_duration_per_task_ms": 147223.81333333332,
      "avg_output_tokens_per_task": 19831.306666666667,
      "last_run_timestamp": "2026-10-01T19:04:26.702Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.8-flash-20260902",
      "model_slug": "google/gemini-3.8-flash",
      "model_name": "Google: Gemini 3.8 Flash",
      "provider_name": "Google",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.2515843,
      "avg_duration_per_task_ms": 421006.06,
      "avg_output_tokens_per_task": 32340.48,
      "last_run_timestamp": "2026-10-01T08:00:24.867Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.3-codex-20260224",
      "model_slug": "openai/gpt-5.3-codex",
      "model_name": "OpenAI: GPT-5.3-Codex",
      "provider_name": "OpenAI",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.29971074000000003,
      "avg_duration_per_task_ms": 180697.5,
      "avg_output_tokens_per_task": 13569.24,
      "last_run_timestamp": "2026-10-02T23:12:05.500Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0196027,
      "avg_duration_per_task_ms": 76385.02,
      "avg_output_tokens_per_task": 9950,
      "last_run_timestamp": "2026-07-18T08:04:02.230Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "Tencent",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.07618375999999999,
      "avg_duration_per_task_ms": 574082.66,
      "avg_output_tokens_per_task": 14524.72,
      "last_run_timestamp": "2026-10-03T08:19:27.962Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "CoreWeave",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.10139794,
      "avg_duration_per_task_ms": 119629.34,
      "avg_output_tokens_per_task": 9506.02,
      "last_run_timestamp": "2026-08-01T17:46:04.356Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Ionstream",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.09723594,
      "avg_duration_per_task_ms": 178772.1,
      "avg_output_tokens_per_task": 8310.58,
      "last_run_timestamp": "2026-08-01T08:05:58.203Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Together",
      "accuracy": 0.7528571428571429,
      "accuracy_std_dev": 0.043949114567478,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.1576592571428571,
      "avg_duration_per_task_ms": 349242.77428571426,
      "avg_output_tokens_per_task": 22490.9,
      "last_run_timestamp": "2026-10-03T08:08:18.114Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Azure",
      "accuracy": 0.7525925555555556,
      "accuracy_std_dev": 0.03021638651133634,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.5157013133333334,
      "avg_duration_per_task_ms": 178548.1511111111,
      "avg_output_tokens_per_task": 7818.084444444445,
      "last_run_timestamp": "2026-10-01T19:05:47.886Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.752499875,
      "accuracy_std_dev": 0.03850520617863741,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.07637632,
      "avg_duration_per_task_ms": 229071.57,
      "avg_output_tokens_per_task": 9485.04,
      "last_run_timestamp": "2026-10-03T16:14:20.292Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Google",
      "accuracy": 0.7523456666666667,
      "accuracy_std_dev": 0.03581868589830568,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.22564089555555555,
      "avg_duration_per_task_ms": 98845.40814814816,
      "avg_output_tokens_per_task": 7557.676296296297,
      "last_run_timestamp": "2026-10-03T08:43:57.161Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "Azure",
      "accuracy": 0.7522221666666666,
      "accuracy_std_dev": 0.01511278994115765,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.44058496666666663,
      "avg_duration_per_task_ms": 458389.25,
      "avg_output_tokens_per_task": 35681.92333333333,
      "last_run_timestamp": "2026-10-01T16:01:13.647Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Modal",
      "accuracy": 0.7519045714285715,
      "accuracy_std_dev": 0.03226166192184128,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.016922877142857146,
      "avg_duration_per_task_ms": 77003.04285714286,
      "avg_output_tokens_per_task": 4241.571428571428,
      "last_run_timestamp": "2026-10-03T08:48:31.438Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Venice",
      "accuracy": 0.751852,
      "accuracy_std_dev": 0.03366914875008805,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021374442222222224,
      "avg_duration_per_task_ms": 173854.87555555557,
      "avg_output_tokens_per_task": 14192.557777777778,
      "last_run_timestamp": "2026-10-03T08:32:03.332Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "AkashML",
      "accuracy": 0.7517005306122448,
      "accuracy_std_dev": 0.012796557796532647,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.09461716326530611,
      "avg_duration_per_task_ms": 209290.9224489796,
      "avg_output_tokens_per_task": 7253.518367346939,
      "last_run_timestamp": "2026-08-15T08:10:05.152Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7513333999999999,
      "accuracy_std_dev": 0.030521588717496353,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.062658464,
      "avg_duration_per_task_ms": 1700876.012,
      "avg_output_tokens_per_task": 8868.604,
      "last_run_timestamp": "2026-08-29T16:20:58.762Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "CoreWeave",
      "accuracy": 0.7511113333333332,
      "accuracy_std_dev": 0.014740562999952068,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.14565638333333333,
      "avg_duration_per_task_ms": 139837.15,
      "avg_output_tokens_per_task": 20029.903333333332,
      "last_run_timestamp": "2026-10-03T08:32:03.353Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7511113333333332,
      "accuracy_std_dev": 0.02199873172606905,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.48754586,
      "avg_duration_per_task_ms": 141651.29333333333,
      "avg_output_tokens_per_task": 7913.286666666667,
      "last_run_timestamp": "2026-10-01T19:04:40.362Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7510256923076923,
      "accuracy_std_dev": 0.024103851303776277,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.3703844146153846,
      "avg_duration_per_task_ms": 173913.4323076923,
      "avg_output_tokens_per_task": 11840.180769230768,
      "last_run_timestamp": "2026-10-03T08:17:09.274Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "Novita",
      "accuracy": 0.7508382676056339,
      "accuracy_std_dev": 0.0329423758499596,
      "run_count": 10,
      "total_questions": 497,
      "avg_cost_per_task_usd": 0.19496027364185115,
      "avg_duration_per_task_ms": 585792.7323943662,
      "avg_output_tokens_per_task": 27449.199195171026,
      "last_run_timestamp": "2026-10-03T08:06:43.538Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Fireworks",
      "accuracy": 0.7505555,
      "accuracy_std_dev": 0.03274012621004996,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.02124632666666667,
      "avg_duration_per_task_ms": 354797.86666666664,
      "avg_output_tokens_per_task": 18494.39,
      "last_run_timestamp": "2026-09-19T08:13:24.062Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "BaseTen",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.28014248,
      "avg_duration_per_task_ms": 223020.56,
      "avg_output_tokens_per_task": 18715.38,
      "last_run_timestamp": "2026-09-12T08:12:13.731Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Chutes",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.05066206,
      "avg_duration_per_task_ms": 368063.6,
      "avg_output_tokens_per_task": 8379.6,
      "last_run_timestamp": "2026-07-25T08:07:26.285Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "OpenAI",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.04384306609373936,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.50077502,
      "avg_duration_per_task_ms": 139265.845,
      "avg_output_tokens_per_task": 7986.47,
      "last_run_timestamp": "2026-10-02T23:08:08.058Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Venice",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.016666999999999987,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01952235,
      "avg_duration_per_task_ms": 2267441.92,
      "avg_output_tokens_per_task": 19667.19,
      "last_run_timestamp": "2026-10-03T08:13:52.062Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.025963164791681315,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012758686666666668,
      "avg_duration_per_task_ms": 426185.5333333333,
      "avg_output_tokens_per_task": 4133.86,
      "last_run_timestamp": "2026-10-09T04:18:40.640Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Parasail",
      "accuracy": 0.7499999,
      "accuracy_std_dev": 0.0217561485582812,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.03801625199999999,
      "avg_duration_per_task_ms": 294150.938,
      "avg_output_tokens_per_task": 23410.556,
      "last_run_timestamp": "2026-10-03T16:08:50.061Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.7495237142857143,
      "accuracy_std_dev": 0.027338579222746164,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11086225999999999,
      "avg_duration_per_task_ms": 273444.3714285714,
      "avg_output_tokens_per_task": 12185.84,
      "last_run_timestamp": "2026-10-03T08:09:23.862Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Modal",
      "accuracy": 0.7493332,
      "accuracy_std_dev": 0.011623875814890659,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.177717976,
      "avg_duration_per_task_ms": 166159.384,
      "avg_output_tokens_per_task": 12813.752,
      "last_run_timestamp": "2026-10-03T08:12:20.018Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7491665000000001,
      "accuracy_std_dev": 0.024480644282779816,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.014165765000000002,
      "avg_duration_per_task_ms": 216043.245,
      "avg_output_tokens_per_task": 15278.05,
      "last_run_timestamp": "2026-10-03T16:07:05.244Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.7490475714285715,
      "accuracy_std_dev": 0.017155993645346933,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.034581794285714285,
      "avg_duration_per_task_ms": 410319.02857142856,
      "avg_output_tokens_per_task": 9408.788571428571,
      "last_run_timestamp": "2026-10-09T02:05:38.689Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "NextBit",
      "accuracy": 0.7490367167630058,
      "accuracy_std_dev": 0.02179500605877504,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.011687421965317918,
      "avg_duration_per_task_ms": 205541.59826589594,
      "avg_output_tokens_per_task": 4094.841040462428,
      "last_run_timestamp": "2026-10-03T08:38:03.473Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.748889,
      "accuracy_std_dev": 0.015713326891527487,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.013287926666666667,
      "avg_duration_per_task_ms": 103204.87333333334,
      "avg_output_tokens_per_task": 11122.173333333334,
      "last_run_timestamp": "2026-10-08T06:08:21.516Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Wafer",
      "accuracy": 0.748889,
      "accuracy_std_dev": 0.019626097517789383,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.24296280222222222,
      "avg_duration_per_task_ms": 310167.38,
      "avg_output_tokens_per_task": 9997.322222222223,
      "last_run_timestamp": "2026-10-03T08:39:17.397Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "Google",
      "accuracy": 0.7488737702702704,
      "accuracy_std_dev": 0.035681240826082414,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.3075426081081082,
      "avg_duration_per_task_ms": 149494.7972972973,
      "avg_output_tokens_per_task": 20582.41554054054,
      "last_run_timestamp": "2026-10-01T19:05:00.071Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepInfra",
      "accuracy": 0.7485673982808023,
      "accuracy_std_dev": 0.01859093254678699,
      "run_count": 14,
      "total_questions": 698,
      "avg_cost_per_task_usd": 0.006995356733524355,
      "avg_duration_per_task_ms": 195318.86962750717,
      "avg_output_tokens_per_task": 12847.305157593124,
      "last_run_timestamp": "2026-10-08T23:16:43.713Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7485381177944861,
      "accuracy_std_dev": 0.021900710861750612,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.028614501253132835,
      "avg_duration_per_task_ms": 631491.8922305765,
      "avg_output_tokens_per_task": 22158.218045112782,
      "last_run_timestamp": "2026-10-03T16:12:26.817Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Together",
      "accuracy": 0.748333375,
      "accuracy_std_dev": 0.018782375754530515,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0418515775,
      "avg_duration_per_task_ms": 191146.0675,
      "avg_output_tokens_per_task": 12827.9725,
      "last_run_timestamp": "2026-10-03T08:33:09.729Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.7483332500000001,
      "accuracy_std_dev": 0.02803765820916874,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.338020715,
      "avg_duration_per_task_ms": 209359.69166666668,
      "avg_output_tokens_per_task": 11805.883333333333,
      "last_run_timestamp": "2026-10-03T08:02:27.539Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7481482222222222,
      "accuracy_std_dev": 0.03186030551146598,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.5489024666666666,
      "avg_duration_per_task_ms": 183789.9622222222,
      "avg_output_tokens_per_task": 10558.302222222223,
      "last_run_timestamp": "2026-10-03T08:38:04.046Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.7480002,
      "accuracy_std_dev": 0.020396130352593862,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.030171100000000003,
      "avg_duration_per_task_ms": 197857.312,
      "avg_output_tokens_per_task": 18663.832,
      "last_run_timestamp": "2026-10-03T08:18:33.163Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Together",
      "accuracy": 0.748,
      "accuracy_std_dev": 0.029333454545961698,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.07858999599999998,
      "avg_duration_per_task_ms": 166655.096,
      "avg_output_tokens_per_task": 8640.336,
      "last_run_timestamp": "2026-08-22T08:39:22.990Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.7477778333333334,
      "accuracy_std_dev": 0.022249863365997446,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.020253733333333336,
      "avg_duration_per_task_ms": 168378.26333333334,
      "avg_output_tokens_per_task": 4203.816666666667,
      "last_run_timestamp": "2026-10-08T06:59:18.400Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Alibaba",
      "accuracy": 0.7477776666666667,
      "accuracy_std_dev": 0.011166340517625106,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.17740497333333335,
      "avg_duration_per_task_ms": 310994.1566666667,
      "avg_output_tokens_per_task": 10783.98,
      "last_run_timestamp": "2026-10-03T08:17:48.657Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "GMICloud",
      "accuracy": 0.7477776666666667,
      "accuracy_std_dev": 0.016063016338436865,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07122743666666666,
      "avg_duration_per_task_ms": 145473.51333333334,
      "avg_output_tokens_per_task": 4717.823333333334,
      "last_run_timestamp": "2026-10-03T08:39:53.772Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Venice",
      "accuracy": 0.747619,
      "accuracy_std_dev": 0.029477692093222922,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.021878034285714287,
      "avg_duration_per_task_ms": 445632.1057142857,
      "avg_output_tokens_per_task": 15072.191428571428,
      "last_run_timestamp": "2026-10-03T16:20:50.532Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "xAI",
      "accuracy": 0.747619,
      "accuracy_std_dev": 0.025305279849978234,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.33171734571428574,
      "avg_duration_per_task_ms": 235739.5942857143,
      "avg_output_tokens_per_task": 12164.34,
      "last_run_timestamp": "2026-10-01T16:01:04.865Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Z.AI",
      "accuracy": 0.7476188571428571,
      "accuracy_std_dev": 0.023211532438045718,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11681679714285716,
      "avg_duration_per_task_ms": 334025.72571428574,
      "avg_output_tokens_per_task": 9744.797142857144,
      "last_run_timestamp": "2026-10-03T16:14:01.845Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.7474074444444444,
      "accuracy_std_dev": 0.025809327224909778,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.4897610111111111,
      "avg_duration_per_task_ms": 194827.99555555556,
      "avg_output_tokens_per_task": 10055.273333333333,
      "last_run_timestamp": "2026-10-03T08:17:06.939Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.7470687587939698,
      "accuracy_std_dev": 0.027067095132974672,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.005836223618090453,
      "avg_duration_per_task_ms": 273469.26130653266,
      "avg_output_tokens_per_task": 12660.396984924622,
      "last_run_timestamp": "2026-10-03T16:13:42.336Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Google",
      "accuracy": 0.7470370000000001,
      "accuracy_std_dev": 0.028671978645437867,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.5621734177777777,
      "avg_duration_per_task_ms": 202509.61777777778,
      "avg_output_tokens_per_task": 9856.906666666666,
      "last_run_timestamp": "2026-10-03T08:21:15.324Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.746667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.1787217,
      "avg_duration_per_task_ms": 84226.38,
      "avg_output_tokens_per_task": 6412.78,
      "last_run_timestamp": "2026-10-03T08:15:16.203Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AkashML",
      "accuracy": 0.746667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02836242,
      "avg_duration_per_task_ms": 881974.84,
      "avg_output_tokens_per_task": 13870.46,
      "last_run_timestamp": "2026-07-18T08:03:38.106Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "StepFun",
      "accuracy": 0.746667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02698528,
      "avg_duration_per_task_ms": 147907.84,
      "avg_output_tokens_per_task": 11065.38,
      "last_run_timestamp": "2026-07-18T08:04:04.303Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "DeepInfra",
      "accuracy": 0.746667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.07051216,
      "avg_duration_per_task_ms": 742250.02,
      "avg_output_tokens_per_task": 12898.82,
      "last_run_timestamp": "2026-10-08T23:33:36.803Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "InferenceNet",
      "accuracy": 0.746667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.07074242,
      "avg_duration_per_task_ms": 800280.44,
      "avg_output_tokens_per_task": 15183.42,
      "last_run_timestamp": "2026-10-02T06:00:07.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Io Net",
      "accuracy": 0.746667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.38669936,
      "avg_duration_per_task_ms": 338232.82,
      "avg_output_tokens_per_task": 9439.24,
      "last_run_timestamp": "2026-07-18T08:04:09.215Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Baidu",
      "accuracy": 0.7466666666666667,
      "accuracy_std_dev": 0.027034372269390698,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.015178880000000002,
      "avg_duration_per_task_ms": 145373.45333333334,
      "avg_output_tokens_per_task": 16435.604444444445,
      "last_run_timestamp": "2026-10-03T08:25:27.328Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7466666666666667,
      "accuracy_std_dev": 0.023726809342082992,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.21412848666666662,
      "avg_duration_per_task_ms": 142371.99333333335,
      "avg_output_tokens_per_task": 22756.126666666667,
      "last_run_timestamp": "2026-10-01T19:04:26.676Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "auto-routing",
      "accuracy": 0.7466666666666667,
      "accuracy_std_dev": 0.028803034820811656,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.18399340666666666,
      "avg_duration_per_task_ms": 116870.40666666666,
      "avg_output_tokens_per_task": 5117.2266666666665,
      "last_run_timestamp": "2026-10-01T19:04:55.696Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.746666625,
      "accuracy_std_dev": 0.03349969319985448,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.087697375,
      "avg_duration_per_task_ms": 258190.44,
      "avg_output_tokens_per_task": 12176.68,
      "last_run_timestamp": "2026-10-03T08:20:48.321Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "auto-routing",
      "accuracy": 0.7466665714285714,
      "accuracy_std_dev": 0.03380611384628165,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03743281428571429,
      "avg_duration_per_task_ms": 197093.65714285715,
      "avg_output_tokens_per_task": 13601.3,
      "last_run_timestamp": "2026-10-03T08:15:57.213Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7466665555555556,
      "accuracy_std_dev": 0.012957518436208841,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.3578504977777778,
      "avg_duration_per_task_ms": 113715.46888888889,
      "avg_output_tokens_per_task": 7110.575555555555,
      "last_run_timestamp": "2026-10-03T08:17:21.944Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Ionstream",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.037271729999999996,
      "avg_duration_per_task_ms": 168918.11,
      "avg_output_tokens_per_task": 24042.94,
      "last_run_timestamp": "2026-10-03T08:36:42.222Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Fireworks",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.011547101609639251,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04422623666666667,
      "avg_duration_per_task_ms": 266653.04,
      "avg_output_tokens_per_task": 13360.296666666667,
      "last_run_timestamp": "2026-09-19T08:01:36.427Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.58250779,
      "avg_duration_per_task_ms": 102747.13,
      "avg_output_tokens_per_task": 7546.85,
      "last_run_timestamp": "2026-10-01T19:04:42.459Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Novita",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.02854502616189613,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.17205141333333332,
      "avg_duration_per_task_ms": 322261.24,
      "avg_output_tokens_per_task": 11142.893333333333,
      "last_run_timestamp": "2026-10-03T08:20:41.479Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Sail Research",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.019999888889274717,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06283585333333332,
      "avg_duration_per_task_ms": 361191.69,
      "avg_output_tokens_per_task": 8869.28,
      "last_run_timestamp": "2026-08-29T08:25:32.049Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AkashML",
      "accuracy": 0.7465427067238912,
      "accuracy_std_dev": 0.015385569261983973,
      "run_count": 14,
      "total_questions": 699,
      "avg_cost_per_task_usd": 0.048925171673819746,
      "avg_duration_per_task_ms": 263945.0042918455,
      "avg_output_tokens_per_task": 28450.080114449214,
      "last_run_timestamp": "2026-10-05T17:03:10.825Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Wafer",
      "accuracy": 0.7463769999999998,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 46,
      "avg_cost_per_task_usd": 0.05924082608695652,
      "avg_duration_per_task_ms": 289784.32608695654,
      "avg_output_tokens_per_task": 9531.695652173914,
      "last_run_timestamp": "2026-08-15T14:23:29.125Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "auto-routing",
      "accuracy": 0.7462311155778895,
      "accuracy_std_dev": 0.026033403023507623,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.031126798994974874,
      "avg_duration_per_task_ms": 552773.9422110553,
      "avg_output_tokens_per_task": 14499.40703517588,
      "last_run_timestamp": "2026-10-04T00:00:28.210Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.745926,
      "accuracy_std_dev": 0.02095136179610079,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.3237732711111111,
      "avg_duration_per_task_ms": 198551.82222222222,
      "avg_output_tokens_per_task": 11463.67111111111,
      "last_run_timestamp": "2026-10-03T08:12:20.076Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Z.AI",
      "accuracy": 0.745833375,
      "accuracy_std_dev": 0.053274708014069624,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.08002613749999998,
      "avg_duration_per_task_ms": 231305.2225,
      "avg_output_tokens_per_task": 9750.2375,
      "last_run_timestamp": "2026-10-03T16:12:22.573Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "AtlasCloud",
      "accuracy": 0.74583325,
      "accuracy_std_dev": 0.04047453074388262,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.11309086500000001,
      "avg_duration_per_task_ms": 292417.6425,
      "avg_output_tokens_per_task": 9222.4525,
      "last_run_timestamp": "2026-10-03T16:18:39.856Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7455556666666666,
      "accuracy_std_dev": 0.02506781419913237,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.3528521766666667,
      "avg_duration_per_task_ms": 231482.565,
      "avg_output_tokens_per_task": 10841.498333333333,
      "last_run_timestamp": "2026-10-03T08:06:43.559Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7455555,
      "accuracy_std_dev": 0.01739159368536039,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.014102986666666668,
      "avg_duration_per_task_ms": 390602.61333333334,
      "avg_output_tokens_per_task": 16178.07,
      "last_run_timestamp": "2026-10-08T23:19:45.038Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Reka",
      "accuracy": 0.7451851111111111,
      "accuracy_std_dev": 0.023471128692475918,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.01492189111111111,
      "avg_duration_per_task_ms": 122706.42666666667,
      "avg_output_tokens_per_task": 15563.857777777777,
      "last_run_timestamp": "2026-10-03T13:25:39.404Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.745,
      "accuracy_std_dev": 0.019649399202011284,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.019549249999999997,
      "avg_duration_per_task_ms": 83357.6,
      "avg_output_tokens_per_task": 4168.05,
      "last_run_timestamp": "2026-10-03T08:30:42.966Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Makora",
      "accuracy": 0.7448975714285715,
      "accuracy_std_dev": 0.011805499999999969,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.26889934693877554,
      "avg_duration_per_task_ms": 528124.887755102,
      "avg_output_tokens_per_task": 13421.112244897959,
      "last_run_timestamp": "2026-10-03T08:42:17.390Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "Xiaomi",
      "accuracy": 0.7448486363636363,
      "accuracy_std_dev": 0.04090809875423143,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.021699527272727274,
      "avg_duration_per_task_ms": 329455.08181818185,
      "avg_output_tokens_per_task": 12517.08909090909,
      "last_run_timestamp": "2026-10-03T08:11:53.178Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Io Net",
      "accuracy": 0.7448276551724138,
      "accuracy_std_dev": 0.0035438860340341496,
      "run_count": 3,
      "total_questions": 145,
      "avg_cost_per_task_usd": 0.022715406896551726,
      "avg_duration_per_task_ms": 289516.0620689655,
      "avg_output_tokens_per_task": 10103.63448275862,
      "last_run_timestamp": "2026-08-22T08:32:23.660Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7447619999999999,
      "accuracy_std_dev": 0.026836107941140986,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.3472446428571429,
      "avg_duration_per_task_ms": 201649.75142857144,
      "avg_output_tokens_per_task": 11858.474285714286,
      "last_run_timestamp": "2026-10-03T08:17:06.935Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Cloudflare",
      "accuracy": 0.7447619999999999,
      "accuracy_std_dev": 0.02512556764959325,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.02136247142857143,
      "avg_duration_per_task_ms": 216896.61714285714,
      "avg_output_tokens_per_task": 12936.834285714285,
      "last_run_timestamp": "2026-10-04T03:00:10.240Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7447618571428571,
      "accuracy_std_dev": 0.01621859463198753,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.05547540571428572,
      "avg_duration_per_task_ms": 220891.45428571428,
      "avg_output_tokens_per_task": 9176.125714285714,
      "last_run_timestamp": "2026-10-04T00:00:34.896Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Together",
      "accuracy": 0.744681,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 47,
      "avg_cost_per_task_usd": 0.05595836170212766,
      "avg_duration_per_task_ms": 34218.255319148935,
      "avg_output_tokens_per_task": 6745.085106382979,
      "last_run_timestamp": "2026-08-04T03:12:18.736Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Darkbloom",
      "accuracy": 0.744681,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 47,
      "avg_cost_per_task_usd": 0.02665859574468085,
      "avg_duration_per_task_ms": 538486.1063829787,
      "avg_output_tokens_per_task": 9135.659574468085,
      "last_run_timestamp": "2026-10-03T08:11:53.173Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Venice",
      "accuracy": 0.7446235483870968,
      "accuracy_std_dev": 0.021113320137770862,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.14085775806451614,
      "avg_duration_per_task_ms": 569919.2116935484,
      "avg_output_tokens_per_task": 27271.122983870966,
      "last_run_timestamp": "2026-10-03T16:05:51.295Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.7444445,
      "accuracy_std_dev": 0.03235589896113743,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.007050623333333332,
      "avg_duration_per_task_ms": 202444.91666666666,
      "avg_output_tokens_per_task": 4335.703333333333,
      "last_run_timestamp": "2026-10-03T08:42:41.346Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Azure",
      "accuracy": 0.7444444444444445,
      "accuracy_std_dev": 0.028803274848959327,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.5183121866666667,
      "avg_duration_per_task_ms": 226393.7577777778,
      "avg_output_tokens_per_task": 10022.293333333333,
      "last_run_timestamp": "2026-10-03T08:15:30.348Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "Anthropic",
      "accuracy": 0.7444443333333334,
      "accuracy_std_dev": 0.00831485358994506,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 1.2319787333333332,
      "avg_duration_per_task_ms": 98531.56,
      "avg_output_tokens_per_task": 5524.42,
      "last_run_timestamp": "2026-10-01T19:04:26.700Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7444443333333334,
      "accuracy_std_dev": 0.00831485358994506,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.34168689999999996,
      "avg_duration_per_task_ms": 130878.06,
      "avg_output_tokens_per_task": 7080.526666666667,
      "last_run_timestamp": "2026-10-03T08:24:43.704Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Reka",
      "accuracy": 0.7444443333333334,
      "accuracy_std_dev": 0.006285236475706844,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.04041968666666667,
      "avg_duration_per_task_ms": 291588.68666666665,
      "avg_output_tokens_per_task": 24330.48,
      "last_run_timestamp": "2026-09-30T11:25:05.946Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.7444443333333334,
      "accuracy_std_dev": 0.02845816509467109,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.00811217,
      "avg_duration_per_task_ms": 250640.82333333333,
      "avg_output_tokens_per_task": 5287.056666666666,
      "last_run_timestamp": "2026-10-03T08:42:16.119Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7442424545454546,
      "accuracy_std_dev": 0.04942282999763578,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.08388880545454544,
      "avg_duration_per_task_ms": 185197.90181818182,
      "avg_output_tokens_per_task": 8027.018181818182,
      "last_run_timestamp": "2026-10-03T16:25:24.079Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Sail Research",
      "accuracy": 0.7441947955056181,
      "accuracy_std_dev": 0.013929720069329346,
      "run_count": 9,
      "total_questions": 445,
      "avg_cost_per_task_usd": 0.10463494606741572,
      "avg_duration_per_task_ms": 943614.8921348314,
      "avg_output_tokens_per_task": 21227.357303370787,
      "last_run_timestamp": "2026-09-26T08:24:43.106Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.74416675,
      "accuracy_std_dev": 0.006400770358909954,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.08563862,
      "avg_duration_per_task_ms": 309336.805,
      "avg_output_tokens_per_task": 9480.84,
      "last_run_timestamp": "2026-08-22T16:12:20.461Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Z.AI",
      "accuracy": 0.744,
      "accuracy_std_dev": 0.03855146055858323,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.047344880000000006,
      "avg_duration_per_task_ms": 518226.236,
      "avg_output_tokens_per_task": 9205.268,
      "last_run_timestamp": "2026-08-29T16:10:33.029Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Ambient",
      "accuracy": 0.7439998000000001,
      "accuracy_std_dev": 0.025508095447524134,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.010447856000000002,
      "avg_duration_per_task_ms": 427694.796,
      "avg_output_tokens_per_task": 12591.384,
      "last_run_timestamp": "2026-08-29T09:31:13.940Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Baidu",
      "accuracy": 0.7438095714285714,
      "accuracy_std_dev": 0.03184416704990163,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08113430285714286,
      "avg_duration_per_task_ms": 191849.82857142857,
      "avg_output_tokens_per_task": 9843.122857142856,
      "last_run_timestamp": "2026-10-03T16:14:23.910Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Baidu",
      "accuracy": 0.7438094285714286,
      "accuracy_std_dev": 0.015474279737285213,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.008623625714285715,
      "avg_duration_per_task_ms": 150720.6542857143,
      "avg_output_tokens_per_task": 13331.717142857144,
      "last_run_timestamp": "2026-10-03T16:14:52.862Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Io Net",
      "accuracy": 0.743738,
      "accuracy_std_dev": 0.01688456971344439,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.00938970520231214,
      "avg_duration_per_task_ms": 347914.32658959535,
      "avg_output_tokens_per_task": 3877.916184971098,
      "last_run_timestamp": "2026-10-03T08:46:08.334Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7433335,
      "accuracy_std_dev": 0.043333500000000025,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.64823893,
      "avg_duration_per_task_ms": 264133.83,
      "avg_output_tokens_per_task": 10511,
      "last_run_timestamp": "2026-10-01T16:00:37.003Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Venice",
      "accuracy": 0.743333375,
      "accuracy_std_dev": 0.019436727836350823,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.3260761375,
      "avg_duration_per_task_ms": 567167.5325,
      "avg_output_tokens_per_task": 26014.51,
      "last_run_timestamp": "2026-10-03T16:09:05.690Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Venice",
      "accuracy": 0.743333375,
      "accuracy_std_dev": 0.038151762010218804,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0479863775,
      "avg_duration_per_task_ms": 502843.075,
      "avg_output_tokens_per_task": 9263.875,
      "last_run_timestamp": "2026-10-04T00:00:36.971Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Cloudflare",
      "accuracy": 0.7433333333333333,
      "accuracy_std_dev": 0.017532868606008445,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.012350623333333333,
      "avg_duration_per_task_ms": 125234.99333333333,
      "avg_output_tokens_per_task": 4198.8,
      "last_run_timestamp": "2026-10-03T08:23:09.816Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "StreamLake",
      "accuracy": 0.7433332500000001,
      "accuracy_std_dev": 0.030368011314333716,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.073473185,
      "avg_duration_per_task_ms": 258433.37,
      "avg_output_tokens_per_task": 9863.935,
      "last_run_timestamp": "2026-10-03T16:21:25.001Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7433330000000001,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.34579436,
      "avg_duration_per_task_ms": 133736.57,
      "avg_output_tokens_per_task": 7063.35,
      "last_run_timestamp": "2026-10-03T08:17:21.926Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Nebius",
      "accuracy": 0.7428572857142857,
      "accuracy_std_dev": 0.030207079708071675,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.28986799428571425,
      "avg_duration_per_task_ms": 509147.33714285714,
      "avg_output_tokens_per_task": 9527.642857142857,
      "last_run_timestamp": "2026-10-03T16:04:32.682Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AkashML",
      "accuracy": 0.7426668,
      "accuracy_std_dev": 0.03690245640821219,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.011645404,
      "avg_duration_per_task_ms": 585840.916,
      "avg_output_tokens_per_task": 18794.172,
      "last_run_timestamp": "2026-08-29T08:43:07.210Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Ionstream",
      "accuracy": 0.7425475772357722,
      "accuracy_std_dev": 0.026593049538554243,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.05760138211382114,
      "avg_duration_per_task_ms": 427993.5081300813,
      "avg_output_tokens_per_task": 12666.333333333334,
      "last_run_timestamp": "2026-10-03T08:41:56.167Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.74250025,
      "accuracy_std_dev": 0.024874694303397962,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.01637653,
      "avg_duration_per_task_ms": 283609.14,
      "avg_output_tokens_per_task": 16652.845,
      "last_run_timestamp": "2026-10-03T08:46:48.915Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.7425,
      "accuracy_std_dev": 0.020258742970628775,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.010828945,
      "avg_duration_per_task_ms": 216783.6,
      "avg_output_tokens_per_task": 14322.1425,
      "last_run_timestamp": "2026-10-03T16:11:33.008Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Cloudflare",
      "accuracy": 0.7425,
      "accuracy_std_dev": 0.023318376160230372,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.07715312749999999,
      "avg_duration_per_task_ms": 252677.67,
      "avg_output_tokens_per_task": 11166.4925,
      "last_run_timestamp": "2026-10-03T16:11:33.007Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Google",
      "accuracy": 0.74239996,
      "accuracy_std_dev": 0.0339871010313972,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.3750872199999999,
      "avg_duration_per_task_ms": 202842.4192,
      "avg_output_tokens_per_task": 11926.788,
      "last_run_timestamp": "2026-10-03T08:42:18.933Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "Alibaba",
      "accuracy": 0.7423934827586207,
      "accuracy_std_dev": 0.04075671767696707,
      "run_count": 10,
      "total_questions": 493,
      "avg_cost_per_task_usd": 0.11995382961460448,
      "avg_duration_per_task_ms": 491199.4949290061,
      "avg_output_tokens_per_task": 23541.371196754564,
      "last_run_timestamp": "2026-10-03T08:18:33.162Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "Google AI Studio",
      "accuracy": 0.7422223333333333,
      "accuracy_std_dev": 0.006285707880227596,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.38184122000000004,
      "avg_duration_per_task_ms": 149071.25333333333,
      "avg_output_tokens_per_task": 12794.173333333334,
      "last_run_timestamp": "2026-10-01T19:04:55.722Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7422223333333333,
      "accuracy_std_dev": 0.003142853940113798,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0028365266666666666,
      "avg_duration_per_task_ms": 68565.14666666667,
      "avg_output_tokens_per_task": 11675.026666666667,
      "last_run_timestamp": "2026-10-01T16:01:02.671Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Phala",
      "accuracy": 0.7422222222222222,
      "accuracy_std_dev": 0.023306958640323044,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.08909326444444444,
      "avg_duration_per_task_ms": 804074.62,
      "avg_output_tokens_per_task": 37512.27111111111,
      "last_run_timestamp": "2026-10-03T16:15:37.291Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.7422221111111111,
      "accuracy_std_dev": 0.02012307808597685,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.009145446666666668,
      "avg_duration_per_task_ms": 281815.08222222223,
      "avg_output_tokens_per_task": 4191.742222222222,
      "last_run_timestamp": "2026-10-03T08:37:33.027Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7421122716049383,
      "accuracy_std_dev": 0.04949233809146621,
      "run_count": 5,
      "total_questions": 243,
      "avg_cost_per_task_usd": 0.07618280658436215,
      "avg_duration_per_task_ms": 605963.6255144033,
      "avg_output_tokens_per_task": 7686.909465020576,
      "last_run_timestamp": "2026-08-29T17:13:58.312Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Venice",
      "accuracy": 0.7419048571428571,
      "accuracy_std_dev": 0.03181571083343471,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07701604857142856,
      "avg_duration_per_task_ms": 557828.82,
      "avg_output_tokens_per_task": 11690.26,
      "last_run_timestamp": "2026-10-03T16:12:13.504Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7419047142857143,
      "accuracy_std_dev": 0.02753696505589889,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.34922034571428573,
      "avg_duration_per_task_ms": 221709.42857142858,
      "avg_output_tokens_per_task": 10867.765714285715,
      "last_run_timestamp": "2026-10-03T08:22:39.770Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Phala",
      "accuracy": 0.7416665,
      "accuracy_std_dev": 0.015898969219732467,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.011650780000000001,
      "avg_duration_per_task_ms": 547605.185,
      "avg_output_tokens_per_task": 5560.365,
      "last_run_timestamp": "2026-10-03T08:36:39.549Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7413794367816092,
      "accuracy_std_dev": 0.07093542318941382,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.2491148678160919,
      "avg_duration_per_task_ms": 429289.867816092,
      "avg_output_tokens_per_task": 27858.181034482757,
      "last_run_timestamp": "2026-10-03T16:11:17.912Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.7413334,
      "accuracy_std_dev": 0.04420155150715866,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.13771882,
      "avg_duration_per_task_ms": 864032.39,
      "avg_output_tokens_per_task": 32740.022,
      "last_run_timestamp": "2026-10-03T16:09:32.721Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Venice",
      "accuracy": 0.7412120909090909,
      "accuracy_std_dev": 0.04686698917236574,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.12050856545454547,
      "avg_duration_per_task_ms": 159418.54,
      "avg_output_tokens_per_task": 9545.092727272728,
      "last_run_timestamp": "2026-10-03T16:12:33.493Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Google",
      "accuracy": 0.7411615606060606,
      "accuracy_std_dev": 0.022128401427803163,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.11188127777777777,
      "avg_duration_per_task_ms": 105277.7601010101,
      "avg_output_tokens_per_task": 9824.214646464647,
      "last_run_timestamp": "2026-10-04T00:00:36.942Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Venice",
      "accuracy": 0.7409522857142857,
      "accuracy_std_dev": 0.029905244158911024,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11990698857142856,
      "avg_duration_per_task_ms": 408401.10285714286,
      "avg_output_tokens_per_task": 9217.857142857143,
      "last_run_timestamp": "2026-10-03T16:14:05.348Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Morph",
      "accuracy": 0.7406779661016949,
      "accuracy_std_dev": 0.0192449897289312,
      "run_count": 6,
      "total_questions": 295,
      "avg_cost_per_task_usd": 0.02004586440677966,
      "avg_duration_per_task_ms": 1117850.5423728814,
      "avg_output_tokens_per_task": 22657.922033898307,
      "last_run_timestamp": "2026-09-12T08:11:08.573Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.7403703333333334,
      "accuracy_std_dev": 0.030325141571529955,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.3409674066666667,
      "avg_duration_per_task_ms": 128710.22,
      "avg_output_tokens_per_task": 5246.528888888889,
      "last_run_timestamp": "2026-10-01T19:05:16.562Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7402943534136547,
      "accuracy_std_dev": 0.02302499266101948,
      "run_count": 5,
      "total_questions": 249,
      "avg_cost_per_task_usd": 0.06053981927710843,
      "avg_duration_per_task_ms": 161020.49799196786,
      "avg_output_tokens_per_task": 6902.349397590361,
      "last_run_timestamp": "2026-08-15T08:13:33.738Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "GMICloud",
      "accuracy": 0.7400002,
      "accuracy_std_dev": 0.01264900523203309,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.034869184,
      "avg_duration_per_task_ms": 129959.044,
      "avg_output_tokens_per_task": 16349.368,
      "last_run_timestamp": "2026-10-03T08:21:01.017Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Phala",
      "accuracy": 0.7400001428571428,
      "accuracy_std_dev": 0.03454918738936282,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.12511439714285713,
      "avg_duration_per_task_ms": 410478.8742857143,
      "avg_output_tokens_per_task": 11800.674285714285,
      "last_run_timestamp": "2026-10-03T16:34:12.025Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.014401594564491825,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.3455118266666667,
      "avg_duration_per_task_ms": 100747.32666666666,
      "avg_output_tokens_per_task": 6984.213333333333,
      "last_run_timestamp": "2026-10-03T08:20:22.717Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Fireworks",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.2696536,
      "avg_duration_per_task_ms": 310259.6,
      "avg_output_tokens_per_task": 11435.58,
      "last_run_timestamp": "2026-07-18T08:04:29.408Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Ambient",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05110574,
      "avg_duration_per_task_ms": 158677.44,
      "avg_output_tokens_per_task": 8249.06,
      "last_run_timestamp": "2026-07-18T16:00:09.002Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.044555500139339316,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 1.0935073933333332,
      "avg_duration_per_task_ms": 295552.94,
      "avg_output_tokens_per_task": 22860.126666666667,
      "last_run_timestamp": "2026-10-02T23:08:29.026Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "Novita",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02561014,
      "avg_duration_per_task_ms": 147210.62,
      "avg_output_tokens_per_task": 10787.18,
      "last_run_timestamp": "2026-07-18T08:04:42.215Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Makora",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01199366,
      "avg_duration_per_task_ms": 376489.52,
      "avg_output_tokens_per_task": 20183.24,
      "last_run_timestamp": "2026-10-07T21:32:26.985Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Chutes",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.0354338036978815,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.09767998,
      "avg_duration_per_task_ms": 450597.1575,
      "avg_output_tokens_per_task": 9690.325,
      "last_run_timestamp": "2026-10-03T16:05:51.299Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "WandB",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.12237688000000001,
      "avg_duration_per_task_ms": 90244.42,
      "avg_output_tokens_per_task": 9936.14,
      "last_run_timestamp": "2026-07-18T08:04:45.990Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DekaLLM",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01290664,
      "avg_duration_per_task_ms": 114271.64,
      "avg_output_tokens_per_task": 4375.16,
      "last_run_timestamp": "2026-10-09T04:02:01.544Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03319066,
      "avg_duration_per_task_ms": 77162.14,
      "avg_output_tokens_per_task": 4453.24,
      "last_run_timestamp": "2026-10-04T03:00:16.404Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "Novita",
      "accuracy": 0.7399999090909091,
      "accuracy_std_dev": 0.025266298870790318,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.023259952727272725,
      "avg_duration_per_task_ms": 396353.68727272726,
      "avg_output_tokens_per_task": 12260.443636363636,
      "last_run_timestamp": "2026-10-03T08:16:24.263Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "auto-routing",
      "accuracy": 0.7399998750000001,
      "accuracy_std_dev": 0.02185807758265525,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0728203575,
      "avg_duration_per_task_ms": 345522.1125,
      "avg_output_tokens_per_task": 12908.68,
      "last_run_timestamp": "2026-10-03T16:14:15.894Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Together",
      "accuracy": 0.7399998333333334,
      "accuracy_std_dev": 0.02108185106844802,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.015687136666666667,
      "avg_duration_per_task_ms": 84185.48333333334,
      "avg_output_tokens_per_task": 4443.856666666667,
      "last_run_timestamp": "2026-10-03T08:49:53.707Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.7396624683544304,
      "accuracy_std_dev": 0.02347519769052809,
      "run_count": 8,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.019082217721518985,
      "avg_duration_per_task_ms": 310670.8101265823,
      "avg_output_tokens_per_task": 15967.20506329114,
      "last_run_timestamp": "2026-10-03T08:37:28.611Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Morph",
      "accuracy": 0.7396296666666666,
      "accuracy_std_dev": 0.019177150471908547,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.48443006888888884,
      "avg_duration_per_task_ms": 673286.5122222222,
      "avg_output_tokens_per_task": 9807.494444444445,
      "last_run_timestamp": "2026-10-03T08:24:07.563Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Reka",
      "accuracy": 0.7392593333333333,
      "accuracy_std_dev": 0.014553267185679581,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.015606891111111108,
      "avg_duration_per_task_ms": 184036.77333333335,
      "avg_output_tokens_per_task": 4234.36,
      "last_run_timestamp": "2026-10-03T13:17:20.799Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.7392472258064516,
      "accuracy_std_dev": 0.020530217177614083,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.04685905645161291,
      "avg_duration_per_task_ms": 881254.7661290322,
      "avg_output_tokens_per_task": 14245.16935483871,
      "last_run_timestamp": "2026-10-09T01:09:41.440Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.7391550136986301,
      "accuracy_std_dev": 0.013576518152268981,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.0136173698630137,
      "avg_duration_per_task_ms": 57786.109589041094,
      "avg_output_tokens_per_task": 3732.6301369863013,
      "last_run_timestamp": "2026-10-03T08:21:31.962Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.738889,
      "accuracy_std_dev": 0.017812306747489716,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.022993023333333338,
      "avg_duration_per_task_ms": 282964.01666666666,
      "avg_output_tokens_per_task": 13071.093333333334,
      "last_run_timestamp": "2026-08-29T16:33:29.042Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7388888333333333,
      "accuracy_std_dev": 0.030164102613408253,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.06139170666666667,
      "avg_duration_per_task_ms": 280149.0466666667,
      "avg_output_tokens_per_task": 9746.213333333333,
      "last_run_timestamp": "2026-10-05T19:59:12.757Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Phala",
      "accuracy": 0.7385184444444445,
      "accuracy_std_dev": 0.03154897514486578,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.03057054888888889,
      "avg_duration_per_task_ms": 269640.66222222225,
      "avg_output_tokens_per_task": 17038.204444444444,
      "last_run_timestamp": "2026-10-03T08:35:45.438Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Azure",
      "accuracy": 0.7385184444444445,
      "accuracy_std_dev": 0.017992647486688756,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.3103866155555556,
      "avg_duration_per_task_ms": 191198.25333333333,
      "avg_output_tokens_per_task": 10879.702222222222,
      "last_run_timestamp": "2026-10-01T19:05:18.005Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Fireworks",
      "accuracy": 0.7384840804020101,
      "accuracy_std_dev": 0.03165357517370423,
      "run_count": 16,
      "total_questions": 796,
      "avg_cost_per_task_usd": 0.10895770226130652,
      "avg_duration_per_task_ms": 110251.2148241206,
      "avg_output_tokens_per_task": 8058.163316582914,
      "last_run_timestamp": "2026-09-26T08:26:47.303Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AkashML",
      "accuracy": 0.7382865983935742,
      "accuracy_std_dev": 0.027461257286584682,
      "run_count": 5,
      "total_questions": 249,
      "avg_cost_per_task_usd": 0.07498479919678715,
      "avg_duration_per_task_ms": 91080.29718875502,
      "avg_output_tokens_per_task": 4367.955823293173,
      "last_run_timestamp": "2026-09-26T08:11:26.495Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "auto-routing",
      "accuracy": 0.7381818181818182,
      "accuracy_std_dev": 0.02622221420793717,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.023726459999999998,
      "avg_duration_per_task_ms": 166279.69454545455,
      "avg_output_tokens_per_task": 7657.0654545454545,
      "last_run_timestamp": "2026-10-03T16:10:46.859Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Friendli",
      "accuracy": 0.7381758445945945,
      "accuracy_std_dev": 0.015122946755951125,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.020533699324324327,
      "avg_duration_per_task_ms": 105846.9831081081,
      "avg_output_tokens_per_task": 3919.314189189189,
      "last_run_timestamp": "2026-10-03T08:06:57.005Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Novita",
      "accuracy": 0.7380951428571428,
      "accuracy_std_dev": 0.030594855288470453,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.12595829142857143,
      "avg_duration_per_task_ms": 255162.36285714287,
      "avg_output_tokens_per_task": 10029.605714285713,
      "last_run_timestamp": "2026-10-03T16:12:24.690Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Ambient",
      "accuracy": 0.7377778333333334,
      "accuracy_std_dev": 0.02484511527723086,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06635720666666667,
      "avg_duration_per_task_ms": 282495.5133333333,
      "avg_output_tokens_per_task": 8158.293333333333,
      "last_run_timestamp": "2026-08-29T08:12:20.454Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Modal",
      "accuracy": 0.737500125,
      "accuracy_std_dev": 0.01854046812675925,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.386765285,
      "avg_duration_per_task_ms": 179778.535,
      "avg_output_tokens_per_task": 10745.6125,
      "last_run_timestamp": "2026-10-03T08:40:00.236Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Inceptron",
      "accuracy": 0.7374686954887218,
      "accuracy_std_dev": 0.031950697455106813,
      "run_count": 11,
      "total_questions": 532,
      "avg_cost_per_task_usd": 0.0578316992481203,
      "avg_duration_per_task_ms": 124019.84022556391,
      "avg_output_tokens_per_task": 7638.883458646616,
      "last_run_timestamp": "2026-10-03T16:12:10.890Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7371428571428571,
      "accuracy_std_dev": 0.033210782885032,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.4983071857142858,
      "avg_duration_per_task_ms": 191159.05714285714,
      "avg_output_tokens_per_task": 9902.92,
      "last_run_timestamp": "2026-10-03T08:21:29.528Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.7369696363636363,
      "accuracy_std_dev": 0.0327047385954244,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.027384705454545456,
      "avg_duration_per_task_ms": 418295.4309090909,
      "avg_output_tokens_per_task": 12820.870909090909,
      "last_run_timestamp": "2026-10-03T08:16:37.922Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7369696363636363,
      "accuracy_std_dev": 0.041813872418074854,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.09632115999999999,
      "avg_duration_per_task_ms": 149694.9690909091,
      "avg_output_tokens_per_task": 9129.581818181818,
      "last_run_timestamp": "2026-10-03T16:15:37.342Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Decart",
      "accuracy": 0.736667,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.37755702999999996,
      "avg_duration_per_task_ms": 287090.13,
      "avg_output_tokens_per_task": 9995.37,
      "last_run_timestamp": "2026-10-03T08:10:45.812Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.736667,
      "accuracy_std_dev": 0.02999999999999997,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 1.2879983000000002,
      "avg_duration_per_task_ms": 310547.73,
      "avg_output_tokens_per_task": 22913.4,
      "last_run_timestamp": "2026-10-01T16:00:34.660Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "SiliconFlow",
      "accuracy": 0.736667,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06833741,
      "avg_duration_per_task_ms": 266166.56,
      "avg_output_tokens_per_task": 13745.94,
      "last_run_timestamp": "2026-10-03T08:19:53.305Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Azure",
      "accuracy": 0.7366668333333333,
      "accuracy_std_dev": 0.023333261905247794,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.35079288666666664,
      "avg_duration_per_task_ms": 89188.23666666666,
      "avg_output_tokens_per_task": 6914.113333333334,
      "last_run_timestamp": "2026-10-03T08:12:46.438Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Venice",
      "accuracy": 0.7366668,
      "accuracy_std_dev": 0.022900761755015906,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.063241744,
      "avg_duration_per_task_ms": 1530537.112,
      "avg_output_tokens_per_task": 25222.716,
      "last_run_timestamp": "2026-10-03T08:06:43.536Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "GMICloud",
      "accuracy": 0.7366667499999999,
      "accuracy_std_dev": 0.0328295429291591,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.10887684250000002,
      "avg_duration_per_task_ms": 323590.875,
      "avg_output_tokens_per_task": 18425.0325,
      "last_run_timestamp": "2026-10-03T08:19:27.960Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AtlasCloud",
      "accuracy": 0.736666625,
      "accuracy_std_dev": 0.03915774721219767,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0416237425,
      "avg_duration_per_task_ms": 215980.8525,
      "avg_output_tokens_per_task": 20025.32,
      "last_run_timestamp": "2026-10-03T08:21:01.017Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Fireworks",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.17216302,
      "avg_duration_per_task_ms": 234612.69,
      "avg_output_tokens_per_task": 17080.8,
      "last_run_timestamp": "2026-09-19T08:12:02.180Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Phala",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.020005850000000002,
      "avg_duration_per_task_ms": 329585.74,
      "avg_output_tokens_per_task": 11689.47,
      "last_run_timestamp": "2026-08-29T16:43:42.382Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Morph",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.023333357143440252,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.019450683333333333,
      "avg_duration_per_task_ms": 216938.31,
      "avg_output_tokens_per_task": 4564.706666666667,
      "last_run_timestamp": "2026-10-03T08:34:35.340Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Crusoe",
      "accuracy": 0.7365900057471265,
      "accuracy_std_dev": 0.030044845419567656,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.09514021264367815,
      "avg_duration_per_task_ms": 383890.3362068966,
      "avg_output_tokens_per_task": 11244.488505747126,
      "last_run_timestamp": "2026-10-03T16:13:47.481Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Together",
      "accuracy": 0.736,
      "accuracy_std_dev": 0.01717892766152766,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.23393872400000001,
      "avg_duration_per_task_ms": 250693.228,
      "avg_output_tokens_per_task": 12185.72,
      "last_run_timestamp": "2026-10-03T08:11:53.179Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Mancer 2",
      "accuracy": 0.736,
      "accuracy_std_dev": 0.006798758063058283,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.091687124,
      "avg_duration_per_task_ms": 467869.472,
      "avg_output_tokens_per_task": 12583.56,
      "last_run_timestamp": "2026-10-03T08:09:52.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Cerebras",
      "accuracy": 0.736,
      "accuracy_std_dev": 0.03542738426133094,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.42442627200000005,
      "avg_duration_per_task_ms": 39355.976,
      "avg_output_tokens_per_task": 9202.272,
      "last_run_timestamp": "2026-08-15T14:16:11.865Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.7358334999999999,
      "accuracy_std_dev": 0.025097105854062156,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.015653559999999997,
      "avg_duration_per_task_ms": 644029.88,
      "avg_output_tokens_per_task": 23956.73,
      "last_run_timestamp": "2026-10-03T08:19:27.959Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Relace",
      "accuracy": 0.73583325,
      "accuracy_std_dev": 0.04026825845424036,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0081671125,
      "avg_duration_per_task_ms": 333111.9,
      "avg_output_tokens_per_task": 15655.8425,
      "last_run_timestamp": "2026-10-03T08:38:11.814Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Fireworks",
      "accuracy": 0.7355556666666666,
      "accuracy_std_dev": 0.012570472951413688,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.13620342666666665,
      "avg_duration_per_task_ms": 101042.81333333334,
      "avg_output_tokens_per_task": 8769.966666666667,
      "last_run_timestamp": "2026-08-26T21:00:03.913Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Google",
      "accuracy": 0.7355554444444444,
      "accuracy_std_dev": 0.012171531247693556,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.3527382688888889,
      "avg_duration_per_task_ms": 87063.18222222223,
      "avg_output_tokens_per_task": 7051.828888888889,
      "last_run_timestamp": "2026-10-03T08:18:17.834Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7353034451858913,
      "accuracy_std_dev": 0.02679997092672807,
      "run_count": 21,
      "total_questions": 1049,
      "avg_cost_per_task_usd": 0.1235421267874166,
      "avg_duration_per_task_ms": 175822.2707340324,
      "avg_output_tokens_per_task": 9155.737845567206,
      "last_run_timestamp": "2026-10-03T16:13:53.559Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7352381428571428,
      "accuracy_std_dev": 0.03375250188156883,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.05847352000000001,
      "avg_duration_per_task_ms": 654076.3628571428,
      "avg_output_tokens_per_task": 14818.84,
      "last_run_timestamp": "2026-10-04T00:00:34.892Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Novita",
      "accuracy": 0.7352381428571428,
      "accuracy_std_dev": 0.031615541975556306,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.047925131428571424,
      "avg_duration_per_task_ms": 368302.0857142857,
      "avg_output_tokens_per_task": 10123.34857142857,
      "last_run_timestamp": "2026-10-03T16:15:19.157Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Ionstream",
      "accuracy": 0.735194730964467,
      "accuracy_std_dev": 0.02636439034355808,
      "run_count": 4,
      "total_questions": 197,
      "avg_cost_per_task_usd": 0.11292960406091371,
      "avg_duration_per_task_ms": 484722.62436548225,
      "avg_output_tokens_per_task": 29978.527918781725,
      "last_run_timestamp": "2026-10-03T08:25:13.022Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Parasail",
      "accuracy": 0.7351928099173552,
      "accuracy_std_dev": 0.026445667029780136,
      "run_count": 10,
      "total_questions": 484,
      "avg_cost_per_task_usd": 0.019751508264462807,
      "avg_duration_per_task_ms": 93530.47314049587,
      "avg_output_tokens_per_task": 6382.072314049587,
      "last_run_timestamp": "2026-10-03T16:12:29.900Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.7349997500000001,
      "accuracy_std_dev": 0.002886895683515417,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.12295462999999998,
      "avg_duration_per_task_ms": 89986.35,
      "avg_output_tokens_per_task": 23350.75,
      "last_run_timestamp": "2026-10-01T19:05:46.177Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.8-flash-20260902",
      "model_slug": "google/gemini-3.8-flash",
      "model_name": "Google: Gemini 3.8 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.734694,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 49,
      "avg_cost_per_task_usd": 0.25339910204081634,
      "avg_duration_per_task_ms": 146531.16326530612,
      "avg_output_tokens_per_task": 29831.102040816328,
      "last_run_timestamp": "2026-10-01T08:03:06.861Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "ModelRun",
      "accuracy": 0.7346668,
      "accuracy_std_dev": 0.020177018832325044,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.067634488,
      "avg_duration_per_task_ms": 150581.522,
      "avg_output_tokens_per_task": 9458.92,
      "last_run_timestamp": "2026-10-05T17:43:17.994Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Parasail",
      "accuracy": 0.7346666000000001,
      "accuracy_std_dev": 0.025785270559759498,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.382410584,
      "avg_duration_per_task_ms": 202893.328,
      "avg_output_tokens_per_task": 8639.172,
      "last_run_timestamp": "2026-10-03T08:18:05.263Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Inceptron",
      "accuracy": 0.7345953452115812,
      "accuracy_std_dev": 0.020421322845629582,
      "run_count": 9,
      "total_questions": 449,
      "avg_cost_per_task_usd": 0.01083965478841871,
      "avg_duration_per_task_ms": 494548.15144766145,
      "avg_output_tokens_per_task": 14266.481069042316,
      "last_run_timestamp": "2026-10-03T08:38:09.636Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Alibaba",
      "accuracy": 0.7342569118387909,
      "accuracy_std_dev": 0.03174336773937983,
      "run_count": 8,
      "total_questions": 397,
      "avg_cost_per_task_usd": 0.07517244332493701,
      "avg_duration_per_task_ms": 212080.59697733,
      "avg_output_tokens_per_task": 8222.277078085643,
      "last_run_timestamp": "2026-10-03T16:14:02.962Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "auto-routing",
      "accuracy": 0.7340741111111111,
      "accuracy_std_dev": 0.04903445388804455,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.00916024,
      "avg_duration_per_task_ms": 334894.9577777778,
      "avg_output_tokens_per_task": 15526.06,
      "last_run_timestamp": "2026-10-03T08:11:53.180Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Xiaomi",
      "accuracy": 0.7339392727272728,
      "accuracy_std_dev": 0.021733921472251252,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.007870807272727274,
      "avg_duration_per_task_ms": 338594.0509090909,
      "avg_output_tokens_per_task": 14023.138181818182,
      "last_run_timestamp": "2026-10-03T08:21:29.543Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.7333336666666668,
      "accuracy_std_dev": 0.03771236166328251,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.44470873333333333,
      "avg_duration_per_task_ms": 130557.96,
      "avg_output_tokens_per_task": 20380.893333333333,
      "last_run_timestamp": "2026-10-01T19:04:52.658Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "StreamLake",
      "accuracy": 0.7333335714285714,
      "accuracy_std_dev": 0.023904731551109535,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.019937122857142855,
      "avg_duration_per_task_ms": 479765.5571428572,
      "avg_output_tokens_per_task": 14687.234285714285,
      "last_run_timestamp": "2026-08-29T16:31:42.782Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Decart",
      "accuracy": 0.7333333333333333,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.008480786666666665,
      "avg_duration_per_task_ms": 397846.0266666667,
      "avg_output_tokens_per_task": 19127.326666666668,
      "last_run_timestamp": "2026-08-22T08:38:00.859Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Venice",
      "accuracy": 0.7333332727272728,
      "accuracy_std_dev": 0.022201865425683618,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.07251096727272727,
      "avg_duration_per_task_ms": 663094.6745454546,
      "avg_output_tokens_per_task": 8454.441818181818,
      "last_run_timestamp": "2026-10-03T16:16:13.260Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Phala",
      "accuracy": 0.73333325,
      "accuracy_std_dev": 0.012472414008021857,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.025873325,
      "avg_duration_per_task_ms": 147169.53,
      "avg_output_tokens_per_task": 18134.845,
      "last_run_timestamp": "2026-10-03T08:34:36.496Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.73333325,
      "accuracy_std_dev": 0.012472414008021857,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.17039939499999998,
      "avg_duration_per_task_ms": 114141.445,
      "avg_output_tokens_per_task": 5083.04,
      "last_run_timestamp": "2026-10-02T23:08:37.533Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Google",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": 0.016329931618554536,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012738966666666665,
      "avg_duration_per_task_ms": 70118.46666666666,
      "avg_output_tokens_per_task": 10694.486666666666,
      "last_run_timestamp": "2026-10-08T06:08:43.119Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "NextBit",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01186947,
      "avg_duration_per_task_ms": 176514.74,
      "avg_output_tokens_per_task": 14201.61,
      "last_run_timestamp": "2026-08-29T16:50:07.898Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.058005,
      "avg_duration_per_task_ms": 191243.62,
      "avg_output_tokens_per_task": 30812.32,
      "last_run_timestamp": "2026-09-12T08:09:12.993Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04628359,
      "avg_duration_per_task_ms": 154349.4,
      "avg_output_tokens_per_task": 38415.97,
      "last_run_timestamp": "2026-10-03T08:10:45.808Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.2-20251211",
      "model_slug": "openai/gpt-5.2",
      "model_name": "OpenAI: GPT-5.2",
      "provider_name": "auto-routing",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.2412942,
      "avg_duration_per_task_ms": 207134.84,
      "avg_output_tokens_per_task": 12538.34,
      "last_run_timestamp": "2026-08-01T08:06:12.346Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "ModelRun",
      "accuracy": 0.7328571428571429,
      "accuracy_std_dev": 0.02503509736024079,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.06745117142857142,
      "avg_duration_per_task_ms": 76562.92857142857,
      "avg_output_tokens_per_task": 8496.594285714285,
      "last_run_timestamp": "2026-09-26T08:08:19.642Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Google",
      "accuracy": 0.7325925555555556,
      "accuracy_std_dev": 0.02443313150262758,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.3673924444444444,
      "avg_duration_per_task_ms": 230445.1311111111,
      "avg_output_tokens_per_task": 10853.041111111112,
      "last_run_timestamp": "2026-10-03T08:12:59.963Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Baidu",
      "accuracy": 0.732381,
      "accuracy_std_dev": 0.01034532550064454,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08187367142857144,
      "avg_duration_per_task_ms": 300173.6142857143,
      "avg_output_tokens_per_task": 13347.99142857143,
      "last_run_timestamp": "2026-10-03T16:10:32.007Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.7320002,
      "accuracy_std_dev": 0.03435759143129796,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.016945259999999997,
      "avg_duration_per_task_ms": 153216.504,
      "avg_output_tokens_per_task": 14926.164,
      "last_run_timestamp": "2026-10-03T08:34:50.552Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "SambaNova",
      "accuracy": 0.7319933969849246,
      "accuracy_std_dev": 0.06840954592516678,
      "run_count": 4,
      "total_questions": 199,
      "avg_cost_per_task_usd": 0.07448894472361808,
      "avg_duration_per_task_ms": 91970.93467336684,
      "avg_output_tokens_per_task": 13988.37688442211,
      "last_run_timestamp": "2026-08-15T14:12:46.929Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "CoreWeave",
      "accuracy": 0.73166675,
      "accuracy_std_dev": 0.01280170344866263,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.021476995,
      "avg_duration_per_task_ms": 208483.135,
      "avg_output_tokens_per_task": 8547.82,
      "last_run_timestamp": "2026-10-03T16:05:47.173Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "OpenAI",
      "accuracy": 0.73166675,
      "accuracy_std_dev": 0.015184321229067165,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.29879598,
      "avg_duration_per_task_ms": 177541.565,
      "avg_output_tokens_per_task": 10870.785,
      "last_run_timestamp": "2026-10-02T23:13:07.333Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Inceptron",
      "accuracy": 0.7315437382550335,
      "accuracy_std_dev": 0.03437878873504547,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.014048604026845638,
      "avg_duration_per_task_ms": 181440.27516778524,
      "avg_output_tokens_per_task": 3805.51677852349,
      "last_run_timestamp": "2026-10-03T08:39:48.962Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Novita",
      "accuracy": 0.7315152727272727,
      "accuracy_std_dev": 0.033162046414239345,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.009558798181818184,
      "avg_duration_per_task_ms": 364447.56727272726,
      "avg_output_tokens_per_task": 14249.385454545454,
      "last_run_timestamp": "2026-10-03T08:10:45.809Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Minimax",
      "accuracy": 0.731515090909091,
      "accuracy_std_dev": 0.023799046212571194,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.02329288181818182,
      "avg_duration_per_task_ms": 89903.6490909091,
      "avg_output_tokens_per_task": 7512.998181818181,
      "last_run_timestamp": "2026-10-03T16:13:58.374Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7313987589285714,
      "accuracy_std_dev": 0.03616424235329221,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.08183186607142856,
      "avg_duration_per_task_ms": 746817.6116071428,
      "avg_output_tokens_per_task": 14329.484375,
      "last_run_timestamp": "2026-10-05T20:34:16.393Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7313334,
      "accuracy_std_dev": 0.11123746291443364,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.14579965199999997,
      "avg_duration_per_task_ms": 618150.212,
      "avg_output_tokens_per_task": 27105.312,
      "last_run_timestamp": "2026-10-03T08:15:42.485Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "WandB",
      "accuracy": 0.731293,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 49,
      "avg_cost_per_task_usd": 0.12389479591836736,
      "avg_duration_per_task_ms": 269993.51020408166,
      "avg_output_tokens_per_task": 23771,
      "last_run_timestamp": "2026-07-18T08:00:15.297Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "auto-routing",
      "accuracy": 0.7311113333333332,
      "accuracy_std_dev": 0.011331263428624758,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.10536965333333334,
      "avg_duration_per_task_ms": 88766.42666666667,
      "avg_output_tokens_per_task": 4017.0533333333333,
      "last_run_timestamp": "2026-10-01T19:04:40.360Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7311110000000001,
      "accuracy_std_dev": 0.022662199996175744,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.3988358066666667,
      "avg_duration_per_task_ms": 172929.82666666666,
      "avg_output_tokens_per_task": 27148.5,
      "last_run_timestamp": "2026-10-01T19:04:40.362Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DeepInfra",
      "accuracy": 0.7311110000000001,
      "accuracy_std_dev": 0.019116342955701554,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.4446010933333333,
      "avg_duration_per_task_ms": 752160.54,
      "avg_output_tokens_per_task": 9652.03,
      "last_run_timestamp": "2026-10-08T22:43:35.390Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Makora",
      "accuracy": 0.7308334999999999,
      "accuracy_std_dev": 0.023731723962872993,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.043652735000000005,
      "avg_duration_per_task_ms": 134583.575,
      "avg_output_tokens_per_task": 27937.025,
      "last_run_timestamp": "2026-10-03T08:15:00.917Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "auto-routing",
      "accuracy": 0.730303,
      "accuracy_std_dev": 0.02244879491400172,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.07475869636363637,
      "avg_duration_per_task_ms": 189076.56181818183,
      "avg_output_tokens_per_task": 8589.47090909091,
      "last_run_timestamp": "2026-10-03T16:12:07.465Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7300587002518891,
      "accuracy_std_dev": 0.027075268354798905,
      "run_count": 8,
      "total_questions": 397,
      "avg_cost_per_task_usd": 0.37323221158690173,
      "avg_duration_per_task_ms": 401791.3123425693,
      "avg_output_tokens_per_task": 9265.304785894206,
      "last_run_timestamp": "2026-10-03T08:39:26.414Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Crusoe",
      "accuracy": 0.7300001666666667,
      "accuracy_std_dev": 0.03605552816336059,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.11662036,
      "avg_duration_per_task_ms": 82058.93666666666,
      "avg_output_tokens_per_task": 9160.97,
      "last_run_timestamp": "2026-09-12T08:12:28.966Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Ambient",
      "accuracy": 0.730000125,
      "accuracy_std_dev": 0.04988867885461965,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.07825786500000001,
      "avg_duration_per_task_ms": 212986.7025,
      "avg_output_tokens_per_task": 8293.9625,
      "last_run_timestamp": "2026-09-19T08:12:24.315Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-haiku-5.5-20261007",
      "model_slug": "anthropic/claude-haiku-5.5",
      "model_name": "Anthropic: Claude Haiku 5.5",
      "provider_name": "Azure",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.012764169999999998,
      "avg_duration_per_task_ms": 81526.21,
      "avg_output_tokens_per_task": 10652.99,
      "last_run_timestamp": "2026-10-08T06:07:22.827Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Parasail",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.03241608458334048,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07427336285714285,
      "avg_duration_per_task_ms": 226239.70857142858,
      "avg_output_tokens_per_task": 11230.525714285714,
      "last_run_timestamp": "2026-10-03T16:17:46.210Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Alibaba",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.036666636365138984,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3388476266666666,
      "avg_duration_per_task_ms": 404706.78,
      "avg_output_tokens_per_task": 9319.38,
      "last_run_timestamp": "2026-10-03T08:15:30.345Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "StreamLake",
      "accuracy": 0.7299516521739131,
      "accuracy_std_dev": 0.017388525374334593,
      "run_count": 7,
      "total_questions": 345,
      "avg_cost_per_task_usd": 0.020714037681159418,
      "avg_duration_per_task_ms": 279014.0231884058,
      "avg_output_tokens_per_task": 10794.56811594203,
      "last_run_timestamp": "2026-10-03T08:13:52.062Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7295238571428572,
      "accuracy_std_dev": 0.034215143421598336,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.3555786895238095,
      "avg_duration_per_task_ms": 206423.8238095238,
      "avg_output_tokens_per_task": 10296.156190476191,
      "last_run_timestamp": "2026-10-03T08:16:52.730Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Relace",
      "accuracy": 0.7293334,
      "accuracy_std_dev": 0.013727396775791089,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.21736060000000001,
      "avg_duration_per_task_ms": 395127.28,
      "avg_output_tokens_per_task": 9443.484,
      "last_run_timestamp": "2026-10-03T08:37:59.299Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7291667620967742,
      "accuracy_std_dev": 0.06506350304902127,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.1482808629032258,
      "avg_duration_per_task_ms": 1327131.8165322582,
      "avg_output_tokens_per_task": 27401.167338709678,
      "last_run_timestamp": "2026-10-03T16:07:44.271Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Mara",
      "accuracy": 0.7290600769230768,
      "accuracy_std_dev": 0.010018378923258995,
      "run_count": 4,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.1270635282051282,
      "avg_duration_per_task_ms": 193981.8564102564,
      "avg_output_tokens_per_task": 10184.36923076923,
      "last_run_timestamp": "2026-10-03T16:07:34.513Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Inceptron",
      "accuracy": 0.7289057819548871,
      "accuracy_std_dev": 0.03087378380499384,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.06965342355889724,
      "avg_duration_per_task_ms": 217209.81203007518,
      "avg_output_tokens_per_task": 11564.82456140351,
      "last_run_timestamp": "2026-10-03T16:23:41.889Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Phala",
      "accuracy": 0.728889,
      "accuracy_std_dev": 0.020608136952831695,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06727646666666667,
      "avg_duration_per_task_ms": 501138.4733333333,
      "avg_output_tokens_per_task": 9493.433333333332,
      "last_run_timestamp": "2026-08-29T16:44:16.336Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.728889,
      "accuracy_std_dev": 0.052305021830285715,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.024965873333333333,
      "avg_duration_per_task_ms": 476087.61333333334,
      "avg_output_tokens_per_task": 18580.853333333333,
      "last_run_timestamp": "2026-10-03T08:29:04.085Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7285712857142858,
      "accuracy_std_dev": 0.02217236334277612,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.05063508571428572,
      "avg_duration_per_task_ms": 392975.47428571427,
      "avg_output_tokens_per_task": 14780.671428571428,
      "last_run_timestamp": "2026-10-04T00:00:34.887Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7284848181818182,
      "accuracy_std_dev": 0.02815423218248235,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.07946297454545455,
      "avg_duration_per_task_ms": 222298.18181818182,
      "avg_output_tokens_per_task": 8681.263636363636,
      "last_run_timestamp": "2026-10-03T16:14:07.111Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Novita",
      "accuracy": 0.7283335,
      "accuracy_std_dev": 0.023746461725907702,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.05840676,
      "avg_duration_per_task_ms": 311422.2825,
      "avg_output_tokens_per_task": 9732.815,
      "last_run_timestamp": "2026-10-04T00:00:36.973Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Modal",
      "accuracy": 0.72833325,
      "accuracy_std_dev": 0.022298511411919422,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.044024975,
      "avg_duration_per_task_ms": 136603.385,
      "avg_output_tokens_per_task": 25927.71,
      "last_run_timestamp": "2026-10-03T08:49:47.265Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Novita",
      "accuracy": 0.727878909090909,
      "accuracy_std_dev": 0.03385259734752348,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.022492889090909095,
      "avg_duration_per_task_ms": 121138.31272727273,
      "avg_output_tokens_per_task": 7028.689090909091,
      "last_run_timestamp": "2026-10-03T16:15:50.542Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Venice",
      "accuracy": 0.7278787272727273,
      "accuracy_std_dev": 0.027970785626470766,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.02821320545454546,
      "avg_duration_per_task_ms": 208497.22363636363,
      "avg_output_tokens_per_task": 8992.374545454546,
      "last_run_timestamp": "2026-10-03T16:24:32.945Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DeepInfra",
      "accuracy": 0.7278566673427991,
      "accuracy_std_dev": 0.03572243647191496,
      "run_count": 10,
      "total_questions": 493,
      "avg_cost_per_task_usd": 0.10080718052738337,
      "avg_duration_per_task_ms": 324790.75456389453,
      "avg_output_tokens_per_task": 14787.588235294117,
      "last_run_timestamp": "2026-09-26T08:12:00.292Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.72750025,
      "accuracy_std_dev": 0.014790298059454364,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.013181285000000001,
      "avg_duration_per_task_ms": 37840.265,
      "avg_output_tokens_per_task": 3828.01,
      "last_run_timestamp": "2026-10-03T08:06:57.006Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Sail Research",
      "accuracy": 0.7275,
      "accuracy_std_dev": 0.011273272229038026,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.25819418250000004,
      "avg_duration_per_task_ms": 292629.7475,
      "avg_output_tokens_per_task": 10249.3025,
      "last_run_timestamp": "2026-09-26T08:16:17.361Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Morph",
      "accuracy": 0.7272728080808081,
      "accuracy_std_dev": 0.03401092910650194,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.12744196464646465,
      "avg_duration_per_task_ms": 652406.3232323233,
      "avg_output_tokens_per_task": 27175.661616161615,
      "last_run_timestamp": "2026-08-15T08:15:35.382Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "GMICloud",
      "accuracy": 0.7271427142857143,
      "accuracy_std_dev": 0.03945770950094827,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.00689044,
      "avg_duration_per_task_ms": 394490.1285714286,
      "avg_output_tokens_per_task": 14238.282857142856,
      "last_run_timestamp": "2026-10-03T08:19:00.264Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Phala",
      "accuracy": 0.7270693489932887,
      "accuracy_std_dev": 0.056280655844911175,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.2920954026845638,
      "avg_duration_per_task_ms": 417994.4194630872,
      "avg_output_tokens_per_task": 10865.342281879195,
      "last_run_timestamp": "2026-10-03T08:36:28.640Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.2-20251211",
      "model_slug": "openai/gpt-5.2",
      "model_name": "OpenAI: GPT-5.2",
      "provider_name": "OpenAI",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.27023132,
      "avg_duration_per_task_ms": 228964.68,
      "avg_output_tokens_per_task": 14058.24,
      "last_run_timestamp": "2026-08-01T08:05:59.194Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-astra-pro-20260903",
      "model_slug": "openai/gpt-6-astra-pro",
      "model_name": "OpenAI: GPT-6 Astra Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 2.91364046,
      "avg_duration_per_task_ms": 331183.04,
      "avg_output_tokens_per_task": 12422.4,
      "last_run_timestamp": "2026-10-01T08:02:52.271Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.11300302000000001,
      "avg_duration_per_task_ms": 353321.78,
      "avg_output_tokens_per_task": 3658.14,
      "last_run_timestamp": "2026-10-03T00:00:06.432Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.7-plus-20260602",
      "model_slug": "qwen/qwen3.7-plus",
      "model_name": "Qwen: Qwen3.7 Plus",
      "provider_name": "Alibaba",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05448868,
      "avg_duration_per_task_ms": 350679.6,
      "avg_output_tokens_per_task": 17629,
      "last_run_timestamp": "2026-10-01T08:04:38.753Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Z.AI",
      "accuracy": 0.7266667,
      "accuracy_std_dev": 0.04152637180888791,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.12686799,
      "avg_duration_per_task_ms": 183230.478,
      "avg_output_tokens_per_task": 12462.162,
      "last_run_timestamp": "2026-10-03T16:12:35.768Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7266666666666667,
      "accuracy_std_dev": 0.050625568127402036,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.17683866,
      "avg_duration_per_task_ms": 285023.12333333335,
      "avg_output_tokens_per_task": 10017.94,
      "last_run_timestamp": "2026-10-03T08:23:53.937Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7266665,
      "accuracy_std_dev": 0.033665214491370776,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.027464944999999998,
      "avg_duration_per_task_ms": 167583.24,
      "avg_output_tokens_per_task": 16249.64,
      "last_run_timestamp": "2026-10-03T08:11:53.177Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7266665,
      "accuracy_std_dev": 0.03162256578378802,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.325251515,
      "avg_duration_per_task_ms": 113433.705,
      "avg_output_tokens_per_task": 10925.445,
      "last_run_timestamp": "2026-10-01T19:04:58.091Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.7264311717171718,
      "accuracy_std_dev": 0.023522929859766603,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.04122346464646465,
      "avg_duration_per_task_ms": 213769.9797979798,
      "avg_output_tokens_per_task": 25500.383838383837,
      "last_run_timestamp": "2026-10-03T08:37:32.589Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Moonshot AI",
      "accuracy": 0.725714,
      "accuracy_std_dev": 0.027004721179823374,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.04600672857142857,
      "avg_duration_per_task_ms": 172219.76,
      "avg_output_tokens_per_task": 8290.077142857142,
      "last_run_timestamp": "2026-08-29T17:36:31.073Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Wafer",
      "accuracy": 0.7254901363636364,
      "accuracy_std_dev": 0.03196228807593238,
      "run_count": 15,
      "total_questions": 748,
      "avg_cost_per_task_usd": 0.11537986497326203,
      "avg_duration_per_task_ms": 109572.1256684492,
      "avg_output_tokens_per_task": 9160.804812834225,
      "last_run_timestamp": "2026-10-08T07:00:33.849Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "DeepInfra",
      "accuracy": 0.7251207652173913,
      "accuracy_std_dev": 0.03173202093295975,
      "run_count": 7,
      "total_questions": 345,
      "avg_cost_per_task_usd": 0.020070034782608694,
      "avg_duration_per_task_ms": 306185.0115942029,
      "avg_output_tokens_per_task": 7045.43768115942,
      "last_run_timestamp": "2026-10-09T02:01:31.527Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Novita",
      "accuracy": 0.7248995421686747,
      "accuracy_std_dev": 0.02769605902019275,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.043781132530120476,
      "avg_duration_per_task_ms": 222582.8293172691,
      "avg_output_tokens_per_task": 9978.564257028112,
      "last_run_timestamp": "2026-10-03T16:16:50.387Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Mancer 2",
      "accuracy": 0.7247618571428571,
      "accuracy_std_dev": 0.04843106328860957,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.04376539714285715,
      "avg_duration_per_task_ms": 289052.61714285717,
      "avg_output_tokens_per_task": 14151.488571428572,
      "last_run_timestamp": "2026-10-03T08:24:26.924Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Novita",
      "accuracy": 0.7247618571428571,
      "accuracy_std_dev": 0.02776646862894973,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.05254494857142857,
      "avg_duration_per_task_ms": 272263.16,
      "avg_output_tokens_per_task": 8601.185714285713,
      "last_run_timestamp": "2026-08-29T17:21:26.058Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.7247618571428571,
      "accuracy_std_dev": 0.03161563234689254,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.034889062857142855,
      "avg_duration_per_task_ms": 356175.16285714286,
      "avg_output_tokens_per_task": 8342.902857142857,
      "last_run_timestamp": "2026-08-29T16:11:10.137Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Cloudflare",
      "accuracy": 0.7246665,
      "accuracy_std_dev": 0.02459454772607134,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.070817994,
      "avg_duration_per_task_ms": 195208.492,
      "avg_output_tokens_per_task": 8421.906,
      "last_run_timestamp": "2026-10-03T16:25:05.402Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.7244897346938776,
      "accuracy_std_dev": 0.009027500000000022,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.08190116326530611,
      "avg_duration_per_task_ms": 1045226.0918367347,
      "avg_output_tokens_per_task": 15985.285714285714,
      "last_run_timestamp": "2026-08-01T17:38:07.305Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Novita",
      "accuracy": 0.7244446666666666,
      "accuracy_std_dev": 0.01911642045177101,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05245255333333333,
      "avg_duration_per_task_ms": 276706.37333333335,
      "avg_output_tokens_per_task": 8448.833333333334,
      "last_run_timestamp": "2026-08-01T17:38:23.802Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Novita",
      "accuracy": 0.7242422727272728,
      "accuracy_std_dev": 0.021700055647565487,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.0786679290909091,
      "avg_duration_per_task_ms": 259230.04545454544,
      "avg_output_tokens_per_task": 9112.834545454545,
      "last_run_timestamp": "2026-10-03T16:06:47.444Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "auto-routing",
      "accuracy": 0.724,
      "accuracy_std_dev": 0.028315782800410083,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.034526184,
      "avg_duration_per_task_ms": 154106.95,
      "avg_output_tokens_per_task": 8565.566,
      "last_run_timestamp": "2026-10-03T16:08:50.062Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7238494100418409,
      "accuracy_std_dev": 0.024050891639188762,
      "run_count": 5,
      "total_questions": 239,
      "avg_cost_per_task_usd": 0.1394956820083682,
      "avg_duration_per_task_ms": 649605.6820083682,
      "avg_output_tokens_per_task": 27120.7949790795,
      "last_run_timestamp": "2026-10-08T23:41:49.599Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Chutes",
      "accuracy": 0.7238092857142858,
      "accuracy_std_dev": 0.03648997874451985,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.06535770571428572,
      "avg_duration_per_task_ms": 383455.27142857143,
      "avg_output_tokens_per_task": 11487.757142857143,
      "last_run_timestamp": "2026-10-03T16:10:02.630Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7237036666666666,
      "accuracy_std_dev": 0.03501115714867033,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.3328492511111112,
      "avg_duration_per_task_ms": 204813.59333333332,
      "avg_output_tokens_per_task": 10311.322222222223,
      "last_run_timestamp": "2026-10-03T08:11:53.180Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Together",
      "accuracy": 0.7237035555555555,
      "accuracy_std_dev": 0.03371788875396401,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.29463488,
      "avg_duration_per_task_ms": 344302.6911111111,
      "avg_output_tokens_per_task": 9396.468888888889,
      "last_run_timestamp": "2026-10-03T08:36:28.986Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7236843198380566,
      "accuracy_std_dev": 0.04197719449284337,
      "run_count": 10,
      "total_questions": 494,
      "avg_cost_per_task_usd": 0.02163497165991903,
      "avg_duration_per_task_ms": 102577.22267206477,
      "avg_output_tokens_per_task": 6911.8421052631575,
      "last_run_timestamp": "2026-10-03T16:04:32.681Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Phala",
      "accuracy": 0.7233335,
      "accuracy_std_dev": 0.023333500000000007,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02063872,
      "avg_duration_per_task_ms": 638594.61,
      "avg_output_tokens_per_task": 13867.61,
      "last_run_timestamp": "2026-08-29T16:33:36.025Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.72333325,
      "accuracy_std_dev": 0.01374358436462263,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.011429815,
      "avg_duration_per_task_ms": 101515.345,
      "avg_output_tokens_per_task": 5183.985,
      "last_run_timestamp": "2026-10-02T23:07:05.152Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "SambaNova",
      "accuracy": 0.7233330000000001,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12502769,
      "avg_duration_per_task_ms": 371456.84,
      "avg_output_tokens_per_task": 16328.88,
      "last_run_timestamp": "2026-10-03T16:15:26.163Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "StreamLake",
      "accuracy": 0.7226666,
      "accuracy_std_dev": 0.02619600308902105,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.065411432,
      "avg_duration_per_task_ms": 217655.416,
      "avg_output_tokens_per_task": 9279.2,
      "last_run_timestamp": "2026-10-03T16:12:07.493Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "GMICloud",
      "accuracy": 0.7225001249999999,
      "accuracy_std_dev": 0.02623366025947152,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0554660425,
      "avg_duration_per_task_ms": 314525.895,
      "avg_output_tokens_per_task": 11294.4225,
      "last_run_timestamp": "2026-10-03T16:05:51.299Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Friendli",
      "accuracy": 0.7225,
      "accuracy_std_dev": 0.017618787323479452,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.10703709,
      "avg_duration_per_task_ms": 327457.75,
      "avg_output_tokens_per_task": 14810.6125,
      "last_run_timestamp": "2026-10-04T00:00:34.897Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Alibaba",
      "accuracy": 0.7224880669856459,
      "accuracy_std_dev": 0.03220370562155287,
      "run_count": 21,
      "total_questions": 1045,
      "avg_cost_per_task_usd": 0.11458385071770336,
      "avg_duration_per_task_ms": 139702.72918660287,
      "avg_output_tokens_per_task": 8362.703349282297,
      "last_run_timestamp": "2026-10-09T22:51:18.482Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.7224242727272727,
      "accuracy_std_dev": 0.03827310252072574,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.047232909090909096,
      "avg_duration_per_task_ms": 150332.00363636363,
      "avg_output_tokens_per_task": 8525.521818181818,
      "last_run_timestamp": "2026-10-08T22:58:29.432Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Io Net",
      "accuracy": 0.7222223333333333,
      "accuracy_std_dev": 0.011330936570096713,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.018792733333333332,
      "avg_duration_per_task_ms": 301192.17333333334,
      "avg_output_tokens_per_task": 19692.3,
      "last_run_timestamp": "2026-10-01T16:00:06.404Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Decart",
      "accuracy": 0.7222223333333333,
      "accuracy_std_dev": 0.031427125187575726,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.009912693333333332,
      "avg_duration_per_task_ms": 115398.83333333333,
      "avg_output_tokens_per_task": 5230.286666666667,
      "last_run_timestamp": "2026-10-03T08:11:53.182Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.7222221428571428,
      "accuracy_std_dev": 0.024312166707227047,
      "run_count": 3,
      "total_questions": 147,
      "avg_cost_per_task_usd": 0.046834047619047615,
      "avg_duration_per_task_ms": 58108.6462585034,
      "avg_output_tokens_per_task": 13243.625850340137,
      "last_run_timestamp": "2026-09-29T21:00:05.013Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Chutes",
      "accuracy": 0.7222221111111111,
      "accuracy_std_dev": 0.027917975508621698,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.13745050505050505,
      "avg_duration_per_task_ms": 712939.898989899,
      "avg_output_tokens_per_task": 27938.80808080808,
      "last_run_timestamp": "2026-08-15T14:15:18.888Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Makora",
      "accuracy": 0.722222,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.019033145833333334,
      "avg_duration_per_task_ms": 65993.54166666667,
      "avg_output_tokens_per_task": 3355.9791666666665,
      "last_run_timestamp": "2026-09-12T08:12:22.607Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.72166675,
      "accuracy_std_dev": 0.018782342477377516,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0501920125,
      "avg_duration_per_task_ms": 390041.62,
      "avg_output_tokens_per_task": 9293.87,
      "last_run_timestamp": "2026-10-03T16:12:08.817Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Baidu",
      "accuracy": 0.7214285714285714,
      "accuracy_std_dev": 0.03730415283414796,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.057019652857142854,
      "avg_duration_per_task_ms": 116171.83142857143,
      "avg_output_tokens_per_task": 8636.995714285715,
      "last_run_timestamp": "2026-10-03T16:13:38.527Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Phala",
      "accuracy": 0.7211088376753507,
      "accuracy_std_dev": 0.03778148726612018,
      "run_count": 10,
      "total_questions": 499,
      "avg_cost_per_task_usd": 0.1016803647294589,
      "avg_duration_per_task_ms": 212461.24448897794,
      "avg_output_tokens_per_task": 10204.877755511023,
      "last_run_timestamp": "2026-10-03T16:20:53.360Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Makora",
      "accuracy": 0.7210882448979592,
      "accuracy_std_dev": 0.0076780598851532815,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.009489710204081632,
      "avg_duration_per_task_ms": 145499.9755102041,
      "avg_output_tokens_per_task": 16260.81224489796,
      "last_run_timestamp": "2026-09-26T08:06:43.933Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Azure",
      "accuracy": 0.7207406666666667,
      "accuracy_std_dev": 0.027477417091131398,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.3540747222222222,
      "avg_duration_per_task_ms": 237714.0088888889,
      "avg_output_tokens_per_task": 10830.737777777778,
      "last_run_timestamp": "2026-10-03T08:11:26.721Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "Azure",
      "accuracy": 0.7205384141414142,
      "accuracy_std_dev": 0.051112796809483045,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 1.6761988181818182,
      "avg_duration_per_task_ms": 313695.66161616164,
      "avg_output_tokens_per_task": 23404.97474747475,
      "last_run_timestamp": "2026-10-01T16:01:13.647Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Baidu",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03553636,
      "avg_duration_per_task_ms": 263015.66,
      "avg_output_tokens_per_task": 14384.46,
      "last_run_timestamp": "2026-07-18T16:00:09.315Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Io Net",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.12671924,
      "avg_duration_per_task_ms": 566714.78,
      "avg_output_tokens_per_task": 16714.02,
      "last_run_timestamp": "2026-09-28T18:00:05.892Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Together",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.021081956479416224,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.03251169,
      "avg_duration_per_task_ms": 105065.11,
      "avg_output_tokens_per_task": 22739.925,
      "last_run_timestamp": "2026-10-03T09:03:15.742Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.03231762505197435,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.36211610499999997,
      "avg_duration_per_task_ms": 150104.23,
      "avg_output_tokens_per_task": 23123.4,
      "last_run_timestamp": "2026-10-01T19:04:55.721Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "Novita",
      "accuracy": 0.72,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00317077,
      "avg_duration_per_task_ms": 128719.23,
      "avg_output_tokens_per_task": 15118.03,
      "last_run_timestamp": "2026-10-01T08:00:27.045Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "BaseTen",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.15154258,
      "avg_duration_per_task_ms": 105755.46,
      "avg_output_tokens_per_task": 13133.32,
      "last_run_timestamp": "2026-08-04T03:12:01.476Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.7-max-20260520",
      "model_slug": "qwen/qwen3.7-max",
      "model_name": "Qwen: Qwen3.7 Max",
      "provider_name": "AtlasCloud",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.402901,
      "avg_duration_per_task_ms": 196016.4,
      "avg_output_tokens_per_task": 16256.52,
      "last_run_timestamp": "2026-09-27T00:00:03.103Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Parasail",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.028463379999999996,
      "avg_duration_per_task_ms": 583244.64,
      "avg_output_tokens_per_task": 43866.54,
      "last_run_timestamp": "2026-08-08T08:27:52.152Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Crusoe",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.037352619999999996,
      "avg_duration_per_task_ms": 206767.2,
      "avg_output_tokens_per_task": 11957.14,
      "last_run_timestamp": "2026-09-26T08:10:43.574Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.7199996666666667,
      "accuracy_std_dev": 0.009428090415820642,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09322211333333333,
      "avg_duration_per_task_ms": 149908.11333333334,
      "avg_output_tokens_per_task": 49346.81333333333,
      "last_run_timestamp": "2026-10-01T19:04:52.649Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Venice",
      "accuracy": 0.719620806020067,
      "accuracy_std_dev": 0.02809879417737037,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.05551290969899665,
      "avg_duration_per_task_ms": 968941.5852842809,
      "avg_output_tokens_per_task": 14053.80602006689,
      "last_run_timestamp": "2026-10-04T00:00:34.891Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Together",
      "accuracy": 0.719409382278481,
      "accuracy_std_dev": 0.05128756107661487,
      "run_count": 8,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.06101479240506329,
      "avg_duration_per_task_ms": 936708.475949367,
      "avg_output_tokens_per_task": 64773.12151898734,
      "last_run_timestamp": "2026-10-03T16:09:19.692Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Alibaba",
      "accuracy": 0.7193332,
      "accuracy_std_dev": 0.03143937799893632,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.045152332,
      "avg_duration_per_task_ms": 326196.928,
      "avg_output_tokens_per_task": 13427.836,
      "last_run_timestamp": "2026-10-04T00:00:34.898Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Relace",
      "accuracy": 0.7193332,
      "accuracy_std_dev": 0.02332387046268266,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.011747176000000002,
      "avg_duration_per_task_ms": 276612.664,
      "avg_output_tokens_per_task": 17492.112,
      "last_run_timestamp": "2026-10-03T08:34:32.707Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "Novita",
      "accuracy": 0.7192392617449664,
      "accuracy_std_dev": 0.027973240079682514,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.06318058053691275,
      "avg_duration_per_task_ms": 493682.4899328859,
      "avg_output_tokens_per_task": 14513.312080536913,
      "last_run_timestamp": "2026-08-29T16:33:37.924Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "auto-routing",
      "accuracy": 0.7191666250000001,
      "accuracy_std_dev": 0.02633896432539395,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.300338065,
      "avg_duration_per_task_ms": 406405.855,
      "avg_output_tokens_per_task": 9208.345,
      "last_run_timestamp": "2026-10-03T08:11:53.141Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.7191665,
      "accuracy_std_dev": 0.008291511728870679,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.046859950000000004,
      "avg_duration_per_task_ms": 203049.32,
      "avg_output_tokens_per_task": 32440.35,
      "last_run_timestamp": "2026-10-03T08:38:29.322Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Venice",
      "accuracy": 0.7188888333333333,
      "accuracy_std_dev": 0.026504201858175022,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3197561666666667,
      "avg_duration_per_task_ms": 193449.66,
      "avg_output_tokens_per_task": 11623.543333333333,
      "last_run_timestamp": "2026-10-03T08:06:43.539Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Phala",
      "accuracy": 0.7188888333333333,
      "accuracy_std_dev": 0.03297943429581869,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.11799081000000002,
      "avg_duration_per_task_ms": 673829.7433333333,
      "avg_output_tokens_per_task": 29526.56,
      "last_run_timestamp": "2026-10-03T08:37:23.062Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7188551919191919,
      "accuracy_std_dev": 0.027525589315898408,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.29904142424242425,
      "avg_duration_per_task_ms": 154139.7323232323,
      "avg_output_tokens_per_task": 5036.570707070707,
      "last_run_timestamp": "2026-10-01T19:05:12.160Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Morph",
      "accuracy": 0.718667,
      "accuracy_std_dev": 0.015999999999999966,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.012890876,
      "avg_duration_per_task_ms": 931505.22,
      "avg_output_tokens_per_task": 16944.504,
      "last_run_timestamp": "2026-10-01T09:55:07.922Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Decart",
      "accuracy": 0.7184344343434343,
      "accuracy_std_dev": 0.09707188521882894,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07242429797979798,
      "avg_duration_per_task_ms": 168379.42676767678,
      "avg_output_tokens_per_task": 14277.45202020202,
      "last_run_timestamp": "2026-10-03T16:16:30.145Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Venice",
      "accuracy": 0.71833325,
      "accuracy_std_dev": 0.022298399296530245,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.039712805,
      "avg_duration_per_task_ms": 138975.945,
      "avg_output_tokens_per_task": 15594.74,
      "last_run_timestamp": "2026-10-03T08:38:07.682Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "StreamLake",
      "accuracy": 0.7181818181818181,
      "accuracy_std_dev": 0.03494188642672075,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.03435519272727272,
      "avg_duration_per_task_ms": 183390.90545454546,
      "avg_output_tokens_per_task": 9125.847272727273,
      "last_run_timestamp": "2026-10-03T16:08:50.063Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "StreamLake",
      "accuracy": 0.7180952857142857,
      "accuracy_std_dev": 0.033752526067220005,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.13201746285714283,
      "avg_duration_per_task_ms": 330154.5142857143,
      "avg_output_tokens_per_task": 10313.082857142857,
      "last_run_timestamp": "2026-08-30T00:01:09.490Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Groq",
      "accuracy": 0.7177649598853869,
      "accuracy_std_dev": 0.02734697248773434,
      "run_count": 7,
      "total_questions": 349,
      "avg_cost_per_task_usd": 0.13113841547277938,
      "avg_duration_per_task_ms": 74329.58739255014,
      "avg_output_tokens_per_task": 10317.49570200573,
      "last_run_timestamp": "2026-08-29T17:11:32.461Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Together",
      "accuracy": 0.7173333000000001,
      "accuracy_std_dev": 0.03878137353691847,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.102133462,
      "avg_duration_per_task_ms": 118936.744,
      "avg_output_tokens_per_task": 9103.426,
      "last_run_timestamp": "2026-10-03T16:12:20.041Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Azure",
      "accuracy": 0.7170371111111111,
      "accuracy_std_dev": 0.016956299841680886,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.12728777111111111,
      "avg_duration_per_task_ms": 101323.44222222222,
      "avg_output_tokens_per_task": 3885.56,
      "last_run_timestamp": "2026-10-01T19:04:52.651Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Relace",
      "accuracy": 0.716667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.015077799999999999,
      "avg_duration_per_task_ms": 476878.72,
      "avg_output_tokens_per_task": 26097.54,
      "last_run_timestamp": "2026-09-28T09:00:02.615Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "GMICloud",
      "accuracy": 0.716666625,
      "accuracy_std_dev": 0.019148607435904436,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0257953375,
      "avg_duration_per_task_ms": 222790.8875,
      "avg_output_tokens_per_task": 10050.0125,
      "last_run_timestamp": "2026-10-03T16:12:13.501Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "auto-routing",
      "accuracy": 0.716666625,
      "accuracy_std_dev": 0.02309402881123982,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.06049543999999999,
      "avg_duration_per_task_ms": 304582.9575,
      "avg_output_tokens_per_task": 9615.66,
      "last_run_timestamp": "2026-10-04T00:00:28.270Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.7166665000000001,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06700715,
      "avg_duration_per_task_ms": 992682.5,
      "avg_output_tokens_per_task": 40930.55,
      "last_run_timestamp": "2026-10-03T08:22:39.765Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Minimax",
      "accuracy": 0.7164444666666666,
      "accuracy_std_dev": 0.014780646674471161,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.03797310133333333,
      "avg_duration_per_task_ms": 188817.49466666667,
      "avg_output_tokens_per_task": 9923.881333333333,
      "last_run_timestamp": "2026-10-03T16:15:22.425Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7155646528925619,
      "accuracy_std_dev": 0.07922982153103719,
      "run_count": 5,
      "total_questions": 242,
      "avg_cost_per_task_usd": 0.11734726446280992,
      "avg_duration_per_task_ms": 254637.0785123967,
      "avg_output_tokens_per_task": 42006.561983471074,
      "last_run_timestamp": "2026-08-29T17:18:52.408Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7155558333333333,
      "accuracy_std_dev": 0.0289741860421575,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04852732,
      "avg_duration_per_task_ms": 969869.3766666667,
      "avg_output_tokens_per_task": 8596.123333333333,
      "last_run_timestamp": "2026-08-29T16:47:05.512Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "OpenInference",
      "accuracy": 0.7155556666666666,
      "accuracy_std_dev": 0.040855070557058414,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012743326666666667,
      "avg_duration_per_task_ms": 1224143.2066666668,
      "avg_output_tokens_per_task": 18669.653333333332,
      "last_run_timestamp": "2026-08-29T09:20:05.864Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Relace",
      "accuracy": 0.7155553333333334,
      "accuracy_std_dev": 0.029979285029648195,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.13937678,
      "avg_duration_per_task_ms": 263819.3466666667,
      "avg_output_tokens_per_task": 10311.9,
      "last_run_timestamp": "2026-10-03T16:18:50.927Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Mara",
      "accuracy": 0.715488212121212,
      "accuracy_std_dev": 0.013403360380889571,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.04391666666666666,
      "avg_duration_per_task_ms": 58944.35858585859,
      "avg_output_tokens_per_task": 8110.530303030303,
      "last_run_timestamp": "2026-08-08T08:31:07.742Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "CoreWeave",
      "accuracy": 0.715238,
      "accuracy_std_dev": 0.036811758695907394,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.0628034857142857,
      "avg_duration_per_task_ms": 75005.29714285715,
      "avg_output_tokens_per_task": 9080.965714285714,
      "last_run_timestamp": "2026-09-26T08:24:48.840Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "OpenAI",
      "accuracy": 0.715,
      "accuracy_std_dev": 0.029202378139459823,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.09407665500000001,
      "avg_duration_per_task_ms": 83159.84,
      "avg_output_tokens_per_task": 3660.855,
      "last_run_timestamp": "2026-10-02T23:10:36.500Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.7148146666666667,
      "accuracy_std_dev": 0.031075936024161484,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.09082762666666666,
      "avg_duration_per_task_ms": 156907.88888888888,
      "avg_output_tokens_per_task": 48272.70222222222,
      "last_run_timestamp": "2026-10-01T19:06:49.017Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "OpenInference",
      "accuracy": 0.714166625,
      "accuracy_std_dev": 0.0311691275781722,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0320175475,
      "avg_duration_per_task_ms": 366948.7475,
      "avg_output_tokens_per_task": 5303.03,
      "last_run_timestamp": "2026-10-03T08:46:44.356Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7139999,
      "accuracy_std_dev": 0.1569839316620972,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.059771752,
      "avg_duration_per_task_ms": 154448.396,
      "avg_output_tokens_per_task": 18970.404,
      "last_run_timestamp": "2026-10-03T16:07:44.277Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "DeepInfra",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.006318460000000001,
      "avg_duration_per_task_ms": 234030.92,
      "avg_output_tokens_per_task": 14773.36,
      "last_run_timestamp": "2026-10-01T08:05:57.971Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "GMICloud",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.09868048,
      "avg_duration_per_task_ms": 318461,
      "avg_output_tokens_per_task": 11794.24,
      "last_run_timestamp": "2026-08-29T17:18:09.782Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-flash-20260826",
      "model_slug": "qwen/qwen3.8-flash",
      "model_name": "Qwen: Qwen3.8 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01665106,
      "avg_duration_per_task_ms": 327560.6,
      "avg_output_tokens_per_task": 17227.24,
      "last_run_timestamp": "2026-10-01T08:00:41.459Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Novita",
      "accuracy": 0.7129095028901734,
      "accuracy_std_dev": 0.03295967347643888,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.05974209248554914,
      "avg_duration_per_task_ms": 481297.3583815029,
      "avg_output_tokens_per_task": 14073.28612716763,
      "last_run_timestamp": "2026-08-30T00:01:09.483Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.8-flash-20260902",
      "model_slug": "google/gemini-3.8-flash",
      "model_name": "Google: Gemini 3.8 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.712766,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 47,
      "avg_cost_per_task_usd": 0.21022357446808512,
      "avg_duration_per_task_ms": 261220.25531914894,
      "avg_output_tokens_per_task": 27609.17021276596,
      "last_run_timestamp": "2026-10-01T08:02:55.088Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7126667,
      "accuracy_std_dev": 0.028666690698613956,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.00856998,
      "avg_duration_per_task_ms": 314663.846,
      "avg_output_tokens_per_task": 14195.494,
      "last_run_timestamp": "2026-10-03T08:22:39.768Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.712622768382353,
      "accuracy_std_dev": 0.07553839113396205,
      "run_count": 11,
      "total_questions": 544,
      "avg_cost_per_task_usd": 0.0459969705882353,
      "avg_duration_per_task_ms": 260039.11948529413,
      "avg_output_tokens_per_task": 26119.363970588234,
      "last_run_timestamp": "2026-10-03T16:03:45.633Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Cloudflare",
      "accuracy": 0.7121428571428572,
      "accuracy_std_dev": 0.03208647153253476,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.13476882285714287,
      "avg_duration_per_task_ms": 139447.84285714285,
      "avg_output_tokens_per_task": 8900.078571428572,
      "last_run_timestamp": "2026-10-03T16:12:20.071Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.712121101010101,
      "accuracy_std_dev": 0.029111223892169177,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.12218062626262625,
      "avg_duration_per_task_ms": 537818.2474747475,
      "avg_output_tokens_per_task": 55422.06060606061,
      "last_run_timestamp": "2026-10-02T23:27:09.304Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.71166675,
      "accuracy_std_dev": 0.019075733568790994,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07509325,
      "avg_duration_per_task_ms": 322731.32,
      "avg_output_tokens_per_task": 33483.435,
      "last_run_timestamp": "2026-10-02T23:12:53.530Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Fireworks",
      "accuracy": 0.7113343768844221,
      "accuracy_std_dev": 0.031502220499778535,
      "run_count": 12,
      "total_questions": 597,
      "avg_cost_per_task_usd": 0.40403888609715255,
      "avg_duration_per_task_ms": 235745.01172529312,
      "avg_output_tokens_per_task": 9241.058626465661,
      "last_run_timestamp": "2026-10-03T08:37:32.473Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Parasail",
      "accuracy": 0.7113334,
      "accuracy_std_dev": 0.032048543736650496,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.113843728,
      "avg_duration_per_task_ms": 97247.482,
      "avg_output_tokens_per_task": 9868.37,
      "last_run_timestamp": "2026-10-03T16:12:20.063Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "GMICloud",
      "accuracy": 0.7113332,
      "accuracy_std_dev": 0.04279658761116358,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.09086514600000001,
      "avg_duration_per_task_ms": 216121.818,
      "avg_output_tokens_per_task": 8339.134,
      "last_run_timestamp": "2026-10-03T16:13:56.534Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.7111113333333333,
      "accuracy_std_dev": 0.011331263428624743,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.07186812,
      "avg_duration_per_task_ms": 348295.26,
      "avg_output_tokens_per_task": 33945.17333333333,
      "last_run_timestamp": "2026-10-01T16:00:47.532Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7104762857142857,
      "accuracy_std_dev": 0.03359069896229476,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.04053524,
      "avg_duration_per_task_ms": 377429.66285714286,
      "avg_output_tokens_per_task": 9822.331428571428,
      "last_run_timestamp": "2026-10-03T16:14:22.302Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Inceptron",
      "accuracy": 0.7103825300546448,
      "accuracy_std_dev": 0.03582559506394764,
      "run_count": 11,
      "total_questions": 549,
      "avg_cost_per_task_usd": 0.07252062477231329,
      "avg_duration_per_task_ms": 252602.8378870674,
      "avg_output_tokens_per_task": 8810.581056466302,
      "last_run_timestamp": "2026-10-03T16:18:22.330Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.7103705555555555,
      "accuracy_std_dev": 0.030809822424065814,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.03036840888888889,
      "avg_duration_per_task_ms": 105409.86666666667,
      "avg_output_tokens_per_task": 5076.06,
      "last_run_timestamp": "2026-10-01T19:05:16.550Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.71,
      "accuracy_std_dev": 0.01795033271279393,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.12391315,
      "avg_duration_per_task_ms": 76197.77,
      "avg_output_tokens_per_task": 4134.025,
      "last_run_timestamp": "2026-10-01T19:04:38.304Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Nebius",
      "accuracy": 0.71,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.32462002,
      "avg_duration_per_task_ms": 110744.1,
      "avg_output_tokens_per_task": 8803.62,
      "last_run_timestamp": "2026-10-06T00:00:08.724Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Venice",
      "accuracy": 0.7099996666666667,
      "accuracy_std_dev": 0.03144660377352204,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.24028676333333332,
      "avg_duration_per_task_ms": 296408.56666666665,
      "avg_output_tokens_per_task": 20949.47,
      "last_run_timestamp": "2026-10-03T08:39:24.637Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Decart",
      "accuracy": 0.7097435384615385,
      "accuracy_std_dev": 0.030226114342543606,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.11513952769230772,
      "avg_duration_per_task_ms": 89188.81846153847,
      "avg_output_tokens_per_task": 10509.030769230769,
      "last_run_timestamp": "2026-10-03T16:12:29.882Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Venice",
      "accuracy": 0.70916675,
      "accuracy_std_dev": 0.01730050176114842,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0730223075,
      "avg_duration_per_task_ms": 288375.1675,
      "avg_output_tokens_per_task": 11030.785,
      "last_run_timestamp": "2026-10-03T16:18:18.265Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "BaseTen",
      "accuracy": 0.7091114702886249,
      "accuracy_std_dev": 0.03584624085968895,
      "run_count": 24,
      "total_questions": 1178,
      "avg_cost_per_task_usd": 0.09626076570458404,
      "avg_duration_per_task_ms": 76008.29456706283,
      "avg_output_tokens_per_task": 8032.944821731749,
      "last_run_timestamp": "2026-10-03T16:16:10.215Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Io Net",
      "accuracy": 0.7088886666666667,
      "accuracy_std_dev": 0.01749773883169545,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.11639628666666667,
      "avg_duration_per_task_ms": 755791.72,
      "avg_output_tokens_per_task": 27174.273333333334,
      "last_run_timestamp": "2026-08-15T08:15:24.293Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.7088121666666666,
      "accuracy_std_dev": 0.09574853521360532,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.07333836494252874,
      "avg_duration_per_task_ms": 189816.08620689655,
      "avg_output_tokens_per_task": 23421.635057471263,
      "last_run_timestamp": "2026-10-03T16:12:22.576Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "auto-routing",
      "accuracy": 0.70833325,
      "accuracy_std_dev": 0.014043632407162335,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.026613565000000002,
      "avg_duration_per_task_ms": 223836.8925,
      "avg_output_tokens_per_task": 10131.55,
      "last_run_timestamp": "2026-10-03T16:10:17.811Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7079999,
      "accuracy_std_dev": 0.061665789544041345,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.29473372800000003,
      "avg_duration_per_task_ms": 421854.93,
      "avg_output_tokens_per_task": 9864.05,
      "last_run_timestamp": "2026-10-05T20:05:10.432Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-fin-20260827",
      "model_slug": "inclusionai/ling-3.0-flash-fin",
      "model_name": "inclusionAI: Ling 3.0 Flash Fin",
      "provider_name": "auto-routing",
      "accuracy": 0.707483,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 49,
      "avg_cost_per_task_usd": 0.004842469387755102,
      "avg_duration_per_task_ms": 161558.55102040817,
      "avg_output_tokens_per_task": 17024.020408163266,
      "last_run_timestamp": "2026-10-01T08:01:49.153Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Nebius",
      "accuracy": 0.706667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.3541182,
      "avg_duration_per_task_ms": 1646927.84,
      "avg_output_tokens_per_task": 24534.74,
      "last_run_timestamp": "2026-09-10T00:00:02.948Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "WandB",
      "accuracy": 0.706667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.08614948,
      "avg_duration_per_task_ms": 52133.28,
      "avg_output_tokens_per_task": 7649.3,
      "last_run_timestamp": "2026-07-18T08:04:46.262Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "auto-routing",
      "accuracy": 0.706667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.22511376,
      "avg_duration_per_task_ms": 2937291.8,
      "avg_output_tokens_per_task": 55713.12,
      "last_run_timestamp": "2026-10-03T08:06:57.007Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "ModelRun",
      "accuracy": 0.706667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.1838298,
      "avg_duration_per_task_ms": 329811.9,
      "avg_output_tokens_per_task": 9540.84,
      "last_run_timestamp": "2026-08-15T00:16:18.405Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-astra-pro-20260903",
      "model_slug": "openai/gpt-6-astra-pro",
      "model_name": "OpenAI: GPT-6 Astra Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.7066669999999999,
      "accuracy_std_dev": 0.019999999999999962,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 3.04321209,
      "avg_duration_per_task_ms": 317897.29,
      "avg_output_tokens_per_task": 13272.53,
      "last_run_timestamp": "2026-10-02T23:13:16.494Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "OpenAI",
      "accuracy": 0.7066666666666667,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.5994524266666668,
      "avg_duration_per_task_ms": 636541.5466666666,
      "avg_output_tokens_per_task": 54354.86,
      "last_run_timestamp": "2026-10-02T23:22:41.519Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "CoreWeave",
      "accuracy": 0.7066666363636364,
      "accuracy_std_dev": 0.03527669560198404,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.08133863090909091,
      "avg_duration_per_task_ms": 74950.49272727272,
      "avg_output_tokens_per_task": 8535.285454545454,
      "last_run_timestamp": "2026-10-03T16:09:05.688Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "auto-routing",
      "accuracy": 0.7066665000000001,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.054561849999999995,
      "avg_duration_per_task_ms": 281726.22,
      "avg_output_tokens_per_task": 8888.58,
      "last_run_timestamp": "2026-08-01T17:38:27.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Z.AI",
      "accuracy": 0.7058335,
      "accuracy_std_dev": 0.044526811524069405,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.069417755,
      "avg_duration_per_task_ms": 340857.2225,
      "avg_output_tokens_per_task": 10007.65,
      "last_run_timestamp": "2026-10-04T00:00:36.971Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "Azure",
      "accuracy": 0.7056452258064517,
      "accuracy_std_dev": 0.012405393094940622,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.6654760282258064,
      "avg_duration_per_task_ms": 686538.5927419355,
      "avg_output_tokens_per_task": 55521.342741935485,
      "last_run_timestamp": "2026-10-01T16:01:02.669Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Venice",
      "accuracy": 0.7051926733668341,
      "accuracy_std_dev": 0.031064655778060283,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.04506834422110553,
      "avg_duration_per_task_ms": 454167.2688442211,
      "avg_output_tokens_per_task": 9105.824120603014,
      "last_run_timestamp": "2026-10-03T16:26:46.014Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7049549256756757,
      "accuracy_std_dev": 0.09115185938022449,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.009106506756756757,
      "avg_duration_per_task_ms": 215462.97972972973,
      "avg_output_tokens_per_task": 3961.837837837838,
      "last_run_timestamp": "2026-10-05T20:02:53.923Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "auto-routing",
      "accuracy": 0.7039998000000001,
      "accuracy_std_dev": 0.028782815862246673,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.059492744,
      "avg_duration_per_task_ms": 531413.344,
      "avg_output_tokens_per_task": 14170.264,
      "last_run_timestamp": "2026-08-29T16:22:37.842Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Phala",
      "accuracy": 0.7036621748878924,
      "accuracy_std_dev": 0.10398250962233035,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.19077936547085203,
      "avg_duration_per_task_ms": 1129697.726457399,
      "avg_output_tokens_per_task": 51230.17264573991,
      "last_run_timestamp": "2026-10-03T16:07:05.218Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.7-plus-20260602",
      "model_slug": "qwen/qwen3.7-plus",
      "model_name": "Qwen: Qwen3.7 Plus",
      "provider_name": "auto-routing",
      "accuracy": 0.703333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05425128,
      "avg_duration_per_task_ms": 396980.46,
      "avg_output_tokens_per_task": 20673.9,
      "last_run_timestamp": "2026-10-01T08:05:20.277Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.7022223333333333,
      "accuracy_std_dev": 0.011330936570096727,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.13357544,
      "avg_duration_per_task_ms": 105586.74666666667,
      "avg_output_tokens_per_task": 5260.826666666667,
      "last_run_timestamp": "2026-10-02T23:07:17.250Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "Azure",
      "accuracy": 0.701111,
      "accuracy_std_dev": 0.011811220865487748,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.13736028333333333,
      "avg_duration_per_task_ms": 494538.36333333334,
      "avg_output_tokens_per_task": 57774.223333333335,
      "last_run_timestamp": "2026-10-01T19:05:16.547Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.7011108333333333,
      "accuracy_std_dev": 0.020608011147906124,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.051610086666666666,
      "avg_duration_per_task_ms": 398260.51,
      "avg_output_tokens_per_task": 9237.883333333333,
      "last_run_timestamp": "2026-10-03T16:12:07.463Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "auto-routing",
      "accuracy": 0.7000003333333334,
      "accuracy_std_dev": 0.009428090415820642,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.010915493333333333,
      "avg_duration_per_task_ms": 115952.82,
      "avg_output_tokens_per_task": 5345.16,
      "last_run_timestamp": "2026-10-01T19:04:36.191Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.7000003333333332,
      "accuracy_std_dev": 0.024944382578492925,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12734470666666667,
      "avg_duration_per_task_ms": 591806.9066666666,
      "avg_output_tokens_per_task": 57532.753333333334,
      "last_run_timestamp": "2026-10-01T19:04:42.411Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Morph",
      "accuracy": 0.70000025,
      "accuracy_std_dev": 0.03162288201109283,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.12919751,
      "avg_duration_per_task_ms": 506623.765,
      "avg_output_tokens_per_task": 10892.11,
      "last_run_timestamp": "2026-08-22T08:53:18.310Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "GLM 5.3 Flash",
      "provider_name": "digitalocean",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00357748,
      "avg_duration_per_task_ms": 48793.08,
      "avg_output_tokens_per_task": 1567.76,
      "last_run_timestamp": "2026-10-01T23:50:45.022Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "auto-routing",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0030046400000000003,
      "avg_duration_per_task_ms": 98848.12,
      "avg_output_tokens_per_task": 17113.74,
      "last_run_timestamp": "2026-10-01T08:00:19.901Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.12726106,
      "avg_duration_per_task_ms": 181226.44,
      "avg_output_tokens_per_task": 9838.56,
      "last_run_timestamp": "2026-10-04T03:00:16.405Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.2-20251211",
      "model_slug": "openai/gpt-5.2",
      "model_name": "OpenAI: GPT-5.2",
      "provider_name": "Azure",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.31292506,
      "avg_duration_per_task_ms": 261751.42,
      "avg_output_tokens_per_task": 14948.8,
      "last_run_timestamp": "2026-08-01T08:06:08.193Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "auto-routing",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.1012818,
      "avg_duration_per_task_ms": 144103.72,
      "avg_output_tokens_per_task": 3638.14,
      "last_run_timestamp": "2026-10-01T08:03:00.822Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-flash-20260826",
      "model_slug": "qwen/qwen3.8-flash",
      "model_name": "Qwen: Qwen3.8 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.019795359999999998,
      "avg_duration_per_task_ms": 396615.94,
      "avg_output_tokens_per_task": 23218.96,
      "last_run_timestamp": "2026-10-01T08:03:06.120Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01539542,
      "avg_duration_per_task_ms": 2335452.5,
      "avg_output_tokens_per_task": 25378.6,
      "last_run_timestamp": "2026-07-18T08:04:08.007Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Near AI",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0112321,
      "avg_duration_per_task_ms": 596563.34,
      "avg_output_tokens_per_task": 4097.46,
      "last_run_timestamp": "2026-09-12T09:21:58.569Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "auto-routing",
      "accuracy": 0.6998645487804879,
      "accuracy_std_dev": 0.05531710950158186,
      "run_count": 10,
      "total_questions": 492,
      "avg_cost_per_task_usd": 0.031559408536585364,
      "avg_duration_per_task_ms": 1019512.3414634146,
      "avg_output_tokens_per_task": 43385.81707317073,
      "last_run_timestamp": "2026-10-03T16:10:32.034Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "DeepInfra",
      "accuracy": 0.699494803030303,
      "accuracy_std_dev": 0.010638522711559854,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.021464873737373737,
      "avg_duration_per_task_ms": 737254.1363636364,
      "avg_output_tokens_per_task": 19210.790404040403,
      "last_run_timestamp": "2026-10-08T22:48:13.259Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Parasail",
      "accuracy": 0.6986666,
      "accuracy_std_dev": 0.007483243510671025,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.04888268,
      "avg_duration_per_task_ms": 227205.572,
      "avg_output_tokens_per_task": 6367.684,
      "last_run_timestamp": "2026-08-22T08:43:02.347Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Venice",
      "accuracy": 0.6985507420289855,
      "accuracy_std_dev": 0.027723151851321442,
      "run_count": 7,
      "total_questions": 345,
      "avg_cost_per_task_usd": 0.01719656231884058,
      "avg_duration_per_task_ms": 309641.62898550724,
      "avg_output_tokens_per_task": 9142.672463768116,
      "last_run_timestamp": "2026-08-29T17:28:21.827Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "StreamLake",
      "accuracy": 0.6977778333333333,
      "accuracy_std_dev": 0.026573791972647694,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04297996,
      "avg_duration_per_task_ms": 184061.30666666667,
      "avg_output_tokens_per_task": 8458.556666666667,
      "last_run_timestamp": "2026-08-29T17:03:30.001Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.6977778333333333,
      "accuracy_std_dev": 0.017915990515148434,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.21024033333333336,
      "avg_duration_per_task_ms": 335029.48333333334,
      "avg_output_tokens_per_task": 34166.46666666667,
      "last_run_timestamp": "2026-10-01T16:01:04.829Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Novita",
      "accuracy": 0.6974275704697986,
      "accuracy_std_dev": 0.019759375021746003,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.01971821476510067,
      "avg_duration_per_task_ms": 421936.07718120806,
      "avg_output_tokens_per_task": 9528.71476510067,
      "last_run_timestamp": "2026-10-03T16:05:51.293Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6972790204081633,
      "accuracy_std_dev": 0.04361100000000001,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.08154698979591836,
      "avg_duration_per_task_ms": 436567.05102040817,
      "avg_output_tokens_per_task": 13545.377551020409,
      "last_run_timestamp": "2026-08-01T17:38:10.889Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "Azure",
      "accuracy": 0.6966665000000001,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.62488186,
      "avg_duration_per_task_ms": 152081.62,
      "avg_output_tokens_per_task": 3927.45,
      "last_run_timestamp": "2026-10-01T08:05:41.499Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6966665000000001,
      "accuracy_std_dev": 0.023333500000000007,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.14710045000000002,
      "avg_duration_per_task_ms": 163276.98,
      "avg_output_tokens_per_task": 5053.91,
      "last_run_timestamp": "2026-10-01T08:02:59.984Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Phala",
      "accuracy": 0.6961111666666666,
      "accuracy_std_dev": 0.025416492410878035,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.19325967666666666,
      "avg_duration_per_task_ms": 1865316.5633333332,
      "avg_output_tokens_per_task": 13131.506666666666,
      "last_run_timestamp": "2026-10-04T00:00:37.132Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Reka",
      "accuracy": 0.6958335,
      "accuracy_std_dev": 0.03781341986980283,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.132227115,
      "avg_duration_per_task_ms": 205681.995,
      "avg_output_tokens_per_task": 9407.695,
      "last_run_timestamp": "2026-10-01T22:50:26.018Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "InferenceNet",
      "accuracy": 0.6933334999999999,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04339576,
      "avg_duration_per_task_ms": 106333.2,
      "avg_output_tokens_per_task": 10497.62,
      "last_run_timestamp": "2026-10-03T16:12:20.058Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.6933333333333334,
      "accuracy_std_dev": 0.014401903168987411,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03953364,
      "avg_duration_per_task_ms": 309829.13333333336,
      "avg_output_tokens_per_task": 32261.04,
      "last_run_timestamp": "2026-10-02T23:10:33.891Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.693333,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.08382193,
      "avg_duration_per_task_ms": 858486.8,
      "avg_output_tokens_per_task": 15204.12,
      "last_run_timestamp": "2026-08-01T17:38:16.661Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Ionstream",
      "accuracy": 0.693333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.015592,
      "avg_duration_per_task_ms": 1084747.56,
      "avg_output_tokens_per_task": 17745.48,
      "last_run_timestamp": "2026-07-31T03:44:20.825Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Morph",
      "accuracy": 0.693333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.060288180000000004,
      "avg_duration_per_task_ms": 350636.38,
      "avg_output_tokens_per_task": 10540.6,
      "last_run_timestamp": "2026-07-18T08:04:08.785Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "auto-routing",
      "accuracy": 0.693333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.12999802000000002,
      "avg_duration_per_task_ms": 96250.54,
      "avg_output_tokens_per_task": 4982.06,
      "last_run_timestamp": "2026-10-01T08:05:35.867Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "Ambient",
      "accuracy": 0.693333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02187568,
      "avg_duration_per_task_ms": 236451.76,
      "avg_output_tokens_per_task": 11760.74,
      "last_run_timestamp": "2026-07-18T08:04:03.248Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Novita",
      "accuracy": 0.6931599132947976,
      "accuracy_std_dev": 0.025748486861030352,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.033708676300578035,
      "avg_duration_per_task_ms": 745817.6300578035,
      "avg_output_tokens_per_task": 19785.147398843932,
      "last_run_timestamp": "2026-10-03T16:26:51.484Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.6929823609022557,
      "accuracy_std_dev": 0.026094683447675196,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.03234121052631579,
      "avg_duration_per_task_ms": 241069.54887218046,
      "avg_output_tokens_per_task": 9648.423558897242,
      "last_run_timestamp": "2026-08-29T16:45:30.973Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Mistral",
      "accuracy": 0.6929165,
      "accuracy_std_dev": 0.02385937097609657,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.10850378125,
      "avg_duration_per_task_ms": 77457.025,
      "avg_output_tokens_per_task": 9370.97875,
      "last_run_timestamp": "2026-10-03T16:16:14.856Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Io Net",
      "accuracy": 0.6927083333333334,
      "accuracy_std_dev": 0.006303999999999976,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.03757610416666667,
      "avg_duration_per_task_ms": 270206.4791666667,
      "avg_output_tokens_per_task": 15460,
      "last_run_timestamp": "2026-10-05T08:30:52.469Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6923811428571428,
      "accuracy_std_dev": 0.04719813017612506,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03026138,
      "avg_duration_per_task_ms": 311692.76857142855,
      "avg_output_tokens_per_task": 10090.888571428572,
      "last_run_timestamp": "2026-10-03T16:07:02.038Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Venice",
      "accuracy": 0.6923076113360324,
      "accuracy_std_dev": 0.004408577938519381,
      "run_count": 5,
      "total_questions": 247,
      "avg_cost_per_task_usd": 0.009435951417004049,
      "avg_duration_per_task_ms": 259361.54655870446,
      "avg_output_tokens_per_task": 9080.562753036438,
      "last_run_timestamp": "2026-08-23T00:00:19.944Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6914287142857143,
      "accuracy_std_dev": 0.13703043242466176,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03953962285714285,
      "avg_duration_per_task_ms": 274567.7942857143,
      "avg_output_tokens_per_task": 10410.422857142858,
      "last_run_timestamp": "2026-10-03T16:12:22.581Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "AtlasCloud",
      "accuracy": 0.690000125,
      "accuracy_std_dev": 0.05343730230709045,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0988822075,
      "avg_duration_per_task_ms": 253655.9675,
      "avg_output_tokens_per_task": 10551.77,
      "last_run_timestamp": "2026-10-03T16:11:17.903Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-astra-pro-20260903",
      "model_slug": "openai/gpt-6-astra-pro",
      "model_name": "OpenAI: GPT-6 Astra Pro",
      "provider_name": "Azure",
      "accuracy": 0.69,
      "accuracy_std_dev": 0.009999999999999953,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 2.99530482,
      "avg_duration_per_task_ms": 320001.13,
      "avg_output_tokens_per_task": 12475.35,
      "last_run_timestamp": "2026-10-01T08:05:20.373Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "auto-routing",
      "accuracy": 0.69,
      "accuracy_std_dev": 0.003332999999999975,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01726069,
      "avg_duration_per_task_ms": 298145.74,
      "avg_output_tokens_per_task": 10516.76,
      "last_run_timestamp": "2026-08-01T17:39:06.362Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "auto-routing",
      "accuracy": 0.6891892297297297,
      "accuracy_std_dev": 0.03170486152346707,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.6622049662162162,
      "avg_duration_per_task_ms": 777791.8783783783,
      "avg_output_tokens_per_task": 59355.4527027027,
      "last_run_timestamp": "2026-10-01T16:01:02.610Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6879166250000001,
      "accuracy_std_dev": 0.07218680699569953,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.05169213000000001,
      "avg_duration_per_task_ms": 304751.5525,
      "avg_output_tokens_per_task": 8046.7025,
      "last_run_timestamp": "2026-10-03T16:13:53.555Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Google",
      "accuracy": 0.6877776666666667,
      "accuracy_std_dev": 0.010304212256267955,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.11416634666666667,
      "avg_duration_per_task_ms": 162037.17333333334,
      "avg_output_tokens_per_task": 8889.293333333333,
      "last_run_timestamp": "2026-08-01T17:38:46.618Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Minimax",
      "accuracy": 0.6876191428571429,
      "accuracy_std_dev": 0.0358884483600765,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.03518511571428571,
      "avg_duration_per_task_ms": 165038.81714285715,
      "avg_output_tokens_per_task": 10124.521428571428,
      "last_run_timestamp": "2026-08-29T17:02:26.936Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.6872909230769231,
      "accuracy_std_dev": 0.021147758508094125,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.006207953177257525,
      "avg_duration_per_task_ms": 252164.59866220737,
      "avg_output_tokens_per_task": 9519.789297658863,
      "last_run_timestamp": "2026-08-30T00:01:09.439Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.686667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00219582,
      "avg_duration_per_task_ms": 78121.92,
      "avg_output_tokens_per_task": 11659.46,
      "last_run_timestamp": "2026-10-01T16:00:49.880Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Z.AI",
      "accuracy": 0.6866665000000001,
      "accuracy_std_dev": 0.006666499999999964,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02072933,
      "avg_duration_per_task_ms": 450665.84,
      "avg_output_tokens_per_task": 10600.05,
      "last_run_timestamp": "2026-08-01T17:38:51.399Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "auto-routing",
      "accuracy": 0.6866664285714287,
      "accuracy_std_dev": 0.041709532236501465,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.02076982,
      "avg_duration_per_task_ms": 414109.71142857143,
      "avg_output_tokens_per_task": 20509.702857142856,
      "last_run_timestamp": "2026-10-03T16:11:33.008Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.6850001666666666,
      "accuracy_std_dev": 0.0143694957278334,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04138883,
      "avg_duration_per_task_ms": 242662.93333333332,
      "avg_output_tokens_per_task": 30866.096666666668,
      "last_run_timestamp": "2026-10-01T08:05:57.699Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Friendli",
      "accuracy": 0.6833334999999999,
      "accuracy_std_dev": 0.07322715191682658,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.11095547600000001,
      "avg_duration_per_task_ms": 69234.286,
      "avg_output_tokens_per_task": 8865.102,
      "last_run_timestamp": "2026-10-03T16:27:27.989Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.6833333333333333,
      "accuracy_std_dev": 0.07779388246014092,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.00909372,
      "avg_duration_per_task_ms": 253755.2,
      "avg_output_tokens_per_task": 15941.146666666667,
      "last_run_timestamp": "2026-10-08T22:38:42.670Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "Azure",
      "accuracy": 0.6833333333333333,
      "accuracy_std_dev": 0.012619766977070334,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.10710260999999997,
      "avg_duration_per_task_ms": 135398.91,
      "avg_output_tokens_per_task": 4027.3166666666666,
      "last_run_timestamp": "2026-10-01T08:06:09.665Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "OpenInference",
      "accuracy": 0.68333325,
      "accuracy_std_dev": 0.06368309125024867,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.05480554,
      "avg_duration_per_task_ms": 1657352.435,
      "avg_output_tokens_per_task": 54132.445,
      "last_run_timestamp": "2026-10-03T08:39:18.140Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Novita",
      "accuracy": 0.6824557368421051,
      "accuracy_std_dev": 0.05773499999999998,
      "run_count": 2,
      "total_questions": 95,
      "avg_cost_per_task_usd": 0.19861939999999997,
      "avg_duration_per_task_ms": 328265.04210526316,
      "avg_output_tokens_per_task": 23629.65263157895,
      "last_run_timestamp": "2026-10-03T16:20:06.515Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Minimax",
      "accuracy": 0.6822223333333333,
      "accuracy_std_dev": 0.058584163368230795,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.02942621,
      "avg_duration_per_task_ms": 175208.89333333334,
      "avg_output_tokens_per_task": 10644.23,
      "last_run_timestamp": "2026-08-01T17:38:12.859Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6.1-sol-20260929",
      "model_slug": "openai/gpt-6.1-sol",
      "model_name": "OpenAI: GPT-6.1 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.6822223333333333,
      "accuracy_std_dev": 0.011330936570096727,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09746202,
      "avg_duration_per_task_ms": 132162.47333333333,
      "avg_output_tokens_per_task": 3507.6066666666666,
      "last_run_timestamp": "2026-10-02T23:10:34.290Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Cloudflare",
      "accuracy": 0.6820807398843931,
      "accuracy_std_dev": 0.018427811270036426,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.015085965317919076,
      "avg_duration_per_task_ms": 243924.5809248555,
      "avg_output_tokens_per_task": 11146.326589595375,
      "last_run_timestamp": "2026-08-30T00:01:09.469Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Novita",
      "accuracy": 0.6819047142857143,
      "accuracy_std_dev": 0.025376904886779485,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.007673354285714286,
      "avg_duration_per_task_ms": 259027.58,
      "avg_output_tokens_per_task": 11768.4,
      "last_run_timestamp": "2026-08-30T00:01:09.470Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6816665000000001,
      "accuracy_std_dev": 0.02327362599059285,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.013877145,
      "avg_duration_per_task_ms": 70751.9,
      "avg_output_tokens_per_task": 5304.49,
      "last_run_timestamp": "2026-10-01T19:04:52.652Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "Together",
      "accuracy": 0.6813334,
      "accuracy_std_dev": 0.0026667999999999913,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.08704178,
      "avg_duration_per_task_ms": 82752.556,
      "avg_output_tokens_per_task": 3576.016,
      "last_run_timestamp": "2026-10-01T16:01:04.859Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.6811111666666666,
      "accuracy_std_dev": 0.008088865998326882,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.00903582,
      "avg_duration_per_task_ms": 105469.92666666667,
      "avg_output_tokens_per_task": 8459.866666666667,
      "last_run_timestamp": "2026-10-01T08:05:22.516Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "auto-routing",
      "accuracy": 0.68,
      "accuracy_std_dev": 0.019999999999999962,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.33301777,
      "avg_duration_per_task_ms": 171524.99,
      "avg_output_tokens_per_task": 8948.71,
      "last_run_timestamp": "2026-09-01T12:41:24.616Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "auto-routing",
      "accuracy": 0.68,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.58024224,
      "avg_duration_per_task_ms": 141487.98,
      "avg_output_tokens_per_task": 3731.58,
      "last_run_timestamp": "2026-10-01T08:05:19.702Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.68,
      "accuracy_std_dev": 0.03810292105162893,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.00883898,
      "avg_duration_per_task_ms": 114396.12,
      "avg_output_tokens_per_task": 8841.28,
      "last_run_timestamp": "2026-10-02T23:08:44.329Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Novita",
      "accuracy": 0.68,
      "accuracy_std_dev": 0.01440159456449179,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.015966806666666666,
      "avg_duration_per_task_ms": 259752.2,
      "avg_output_tokens_per_task": 10038.846666666666,
      "last_run_timestamp": "2026-08-01T17:38:31.690Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "Fireworks",
      "accuracy": 0.6777776666666667,
      "accuracy_std_dev": 0.013698571612479227,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0340762,
      "avg_duration_per_task_ms": 244093.44,
      "avg_output_tokens_per_task": 27086.806666666667,
      "last_run_timestamp": "2026-09-19T08:01:53.393Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "DekaLLM",
      "accuracy": 0.6766665000000001,
      "accuracy_std_dev": 0.016666499999999973,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01554348,
      "avg_duration_per_task_ms": 297033.57,
      "avg_output_tokens_per_task": 18106.79,
      "last_run_timestamp": "2026-10-09T05:05:31.663Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "OpenAI",
      "accuracy": 0.6766665000000001,
      "accuracy_std_dev": 0.0033335000000000448,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.5967202500000001,
      "avg_duration_per_task_ms": 137724.15,
      "avg_output_tokens_per_task": 4015.19,
      "last_run_timestamp": "2026-10-02T23:06:01.512Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6766665000000001,
      "accuracy_std_dev": 0.0033335000000000448,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04451592,
      "avg_duration_per_task_ms": 466424.76,
      "avg_output_tokens_per_task": 10206.78,
      "last_run_timestamp": "2026-08-01T17:38:21.896Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.6761905714285714,
      "accuracy_std_dev": 0.01880810362474601,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.009157782857142858,
      "avg_duration_per_task_ms": 249382.73428571428,
      "avg_output_tokens_per_task": 10084.637142857142,
      "last_run_timestamp": "2026-08-30T00:01:02.161Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Nebius",
      "accuracy": 0.675675972972973,
      "accuracy_std_dev": 0.02284471405428887,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.04042408108108108,
      "avg_duration_per_task_ms": 518660.3783783784,
      "avg_output_tokens_per_task": 20990.527027027027,
      "last_run_timestamp": "2026-09-26T08:11:26.490Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "DeepInfra",
      "accuracy": 0.6755556666666667,
      "accuracy_std_dev": 0.006285236475706844,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.07120372000000001,
      "avg_duration_per_task_ms": 144319.38666666666,
      "avg_output_tokens_per_task": 2561.846666666667,
      "last_run_timestamp": "2026-10-08T22:25:52.178Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Tenstorrent",
      "accuracy": 0.675,
      "accuracy_std_dev": 0.011667000000000038,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.1079227,
      "avg_duration_per_task_ms": 145751.42,
      "avg_output_tokens_per_task": 18113.26,
      "last_run_timestamp": "2026-10-04T05:20:00.959Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6733334999999999,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.64995528,
      "avg_duration_per_task_ms": 85231.23,
      "avg_output_tokens_per_task": 3880.06,
      "last_run_timestamp": "2026-10-01T08:05:26.145Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6733333636363636,
      "accuracy_std_dev": 0.06381806526834732,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.06596764181818182,
      "avg_duration_per_task_ms": 779394.0927272728,
      "avg_output_tokens_per_task": 18514.336363636365,
      "last_run_timestamp": "2026-10-05T20:43:11.175Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.6733333333333333,
      "accuracy_std_dev": 0.02372680934208296,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.11501549333333332,
      "avg_duration_per_task_ms": 145439.36666666667,
      "avg_output_tokens_per_task": 56059.84,
      "last_run_timestamp": "2026-10-01T19:05:16.548Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Nebius",
      "accuracy": 0.6733333333333333,
      "accuracy_std_dev": 0.05192585882668568,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.22792725999999996,
      "avg_duration_per_task_ms": 692249.9666666667,
      "avg_output_tokens_per_task": 10865.973333333333,
      "last_run_timestamp": "2026-10-03T16:12:13.503Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DekaLLM",
      "accuracy": 0.6733330000000001,
      "accuracy_std_dev": 0.019999999999999962,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00884946,
      "avg_duration_per_task_ms": 853445.82,
      "avg_output_tokens_per_task": 18061.82,
      "last_run_timestamp": "2026-10-09T04:22:40.083Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.673333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.09000018000000001,
      "avg_duration_per_task_ms": 1674604.48,
      "avg_output_tokens_per_task": 17624.46,
      "last_run_timestamp": "2026-07-18T16:00:08.839Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.2-chat-20251211",
      "model_slug": "openai/gpt-5.2-chat",
      "model_name": "OpenAI: GPT-5.2 Chat",
      "provider_name": "auto-routing",
      "accuracy": 0.673333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.1116277,
      "avg_duration_per_task_ms": 79582.44,
      "avg_output_tokens_per_task": 4280.42,
      "last_run_timestamp": "2026-08-01T08:05:57.504Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "auto-routing",
      "accuracy": 0.673333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00928508,
      "avg_duration_per_task_ms": 121838.98,
      "avg_output_tokens_per_task": 8898.44,
      "last_run_timestamp": "2026-10-01T08:00:22.469Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "GMICloud",
      "accuracy": 0.672,
      "accuracy_std_dev": 0.04759099821604921,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.018393836,
      "avg_duration_per_task_ms": 255084.928,
      "avg_output_tokens_per_task": 8922.8,
      "last_run_timestamp": "2026-08-29T16:33:39.721Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "NextBit",
      "accuracy": 0.672,
      "accuracy_std_dev": 0.038157690357253,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.01826553,
      "avg_duration_per_task_ms": 344635.034,
      "avg_output_tokens_per_task": 19832.934,
      "last_run_timestamp": "2026-10-07T19:11:59.694Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "auto-routing",
      "accuracy": 0.672,
      "accuracy_std_dev": 0.028409779224766934,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.07962264,
      "avg_duration_per_task_ms": 80557.024,
      "avg_output_tokens_per_task": 3531.236,
      "last_run_timestamp": "2026-10-01T16:00:45.064Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.671111,
      "accuracy_std_dev": 0.034282673881196074,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.1170691888888889,
      "avg_duration_per_task_ms": 122970.47333333333,
      "avg_output_tokens_per_task": 12647.033333333333,
      "last_run_timestamp": "2026-10-03T08:12:32.364Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "CoreWeave",
      "accuracy": 0.67083325,
      "accuracy_std_dev": 0.019346895388860186,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.017078424999999998,
      "avg_duration_per_task_ms": 126227.985,
      "avg_output_tokens_per_task": 21265.575,
      "last_run_timestamp": "2026-10-03T08:15:42.485Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.6706668,
      "accuracy_std_dev": 0.015548970672041286,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.142799016,
      "avg_duration_per_task_ms": 115124.684,
      "avg_output_tokens_per_task": 5068.812,
      "last_run_timestamp": "2026-10-01T08:04:38.837Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Venice",
      "accuracy": 0.6700225369127517,
      "accuracy_std_dev": 0.06602622409757607,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.02446618791946309,
      "avg_duration_per_task_ms": 477944.1677852349,
      "avg_output_tokens_per_task": 16475.68120805369,
      "last_run_timestamp": "2026-10-03T16:16:54.867Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Venice",
      "accuracy": 0.67,
      "accuracy_std_dev": 0.08041555612411072,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.07839768750000001,
      "avg_duration_per_task_ms": 277010.8,
      "avg_output_tokens_per_task": 25559.2775,
      "last_run_timestamp": "2026-10-03T16:03:45.631Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6692307307692308,
      "accuracy_std_dev": 0.031054260799641166,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.12451571153846153,
      "avg_duration_per_task_ms": 119043.02846153846,
      "avg_output_tokens_per_task": 12553.901538461538,
      "last_run_timestamp": "2026-10-03T08:32:52.484Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Cloudflare",
      "accuracy": 0.669204852791878,
      "accuracy_std_dev": 0.0421484109961455,
      "run_count": 8,
      "total_questions": 394,
      "avg_cost_per_task_usd": 0.023811266497461927,
      "avg_duration_per_task_ms": 266573.3654822335,
      "avg_output_tokens_per_task": 17577.88578680203,
      "last_run_timestamp": "2026-10-03T16:10:49.374Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Friendli",
      "accuracy": 0.6685714285714286,
      "accuracy_std_dev": 0.01875964453559928,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.029877088571428574,
      "avg_duration_per_task_ms": 242438.32571428572,
      "avg_output_tokens_per_task": 11057.357142857143,
      "last_run_timestamp": "2026-08-29T16:14:58.314Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Novita",
      "accuracy": 0.668333375,
      "accuracy_std_dev": 0.05299363562480287,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.02315143,
      "avg_duration_per_task_ms": 206155.715,
      "avg_output_tokens_per_task": 9305.305,
      "last_run_timestamp": "2026-10-03T16:13:50.327Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "DeepInfra",
      "accuracy": 0.668,
      "accuracy_std_dev": 0.03357230340027326,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.021775528000000002,
      "avg_duration_per_task_ms": 221637.98,
      "avg_output_tokens_per_task": 29545.616,
      "last_run_timestamp": "2026-10-08T22:26:21.249Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.666667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.038762819999999996,
      "avg_duration_per_task_ms": 314594.08,
      "avg_output_tokens_per_task": 31952.48,
      "last_run_timestamp": "2026-10-01T08:00:24.890Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Parasail",
      "accuracy": 0.6658330000000001,
      "accuracy_std_dev": 0.014790199457749011,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.02195716,
      "avg_duration_per_task_ms": 162018.975,
      "avg_output_tokens_per_task": 16412.775,
      "last_run_timestamp": "2026-10-03T16:06:47.455Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "BaseTen",
      "accuracy": 0.6655251643835617,
      "accuracy_std_dev": 0.050919510401766024,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.0655772808219178,
      "avg_duration_per_task_ms": 50910.90410958904,
      "avg_output_tokens_per_task": 3110.958904109589,
      "last_run_timestamp": "2026-09-12T08:11:26.919Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Friendli",
      "accuracy": 0.664762,
      "accuracy_std_dev": 0.09453067457406919,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08868562,
      "avg_duration_per_task_ms": 89242.03714285714,
      "avg_output_tokens_per_task": 8290.754285714285,
      "last_run_timestamp": "2026-10-03T16:14:25.374Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.6644296040268456,
      "accuracy_std_dev": 0.09950169970519207,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.07663720134228189,
      "avg_duration_per_task_ms": 896985.9765100671,
      "avg_output_tokens_per_task": 47829.18120805369,
      "last_run_timestamp": "2026-10-03T16:13:53.353Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "Google",
      "accuracy": 0.6622221666666667,
      "accuracy_std_dev": 0.0198761101108564,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.31710583,
      "avg_duration_per_task_ms": 192864.10333333333,
      "avg_output_tokens_per_task": 8814.376666666667,
      "last_run_timestamp": "2026-09-01T13:04:20.137Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.6603704444444444,
      "accuracy_std_dev": 0.02063470181843689,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.11393844222222221,
      "avg_duration_per_task_ms": 124032.99777777778,
      "avg_output_tokens_per_task": 12195,
      "last_run_timestamp": "2026-10-03T08:07:11.726Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.6600001428571428,
      "accuracy_std_dev": 0.024688651721727035,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.02290525142857143,
      "avg_duration_per_task_ms": 311367.0028571429,
      "avg_output_tokens_per_task": 10406.782857142856,
      "last_run_timestamp": "2026-08-29T16:11:34.887Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "WandB",
      "accuracy": 0.66,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00861218,
      "avg_duration_per_task_ms": 379136.4,
      "avg_output_tokens_per_task": 24201,
      "last_run_timestamp": "2026-07-18T16:00:10.533Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Mancer 2",
      "accuracy": 0.659524,
      "accuracy_std_dev": 0.0937819106087553,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.058310568571428564,
      "avg_duration_per_task_ms": 524817.9057142857,
      "avg_output_tokens_per_task": 17039.157142857144,
      "last_run_timestamp": "2026-10-03T16:12:24.719Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6588888333333334,
      "accuracy_std_dev": 0.02354406690595791,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.022377286666666666,
      "avg_duration_per_task_ms": 254720.16666666666,
      "avg_output_tokens_per_task": 9788.826666666666,
      "last_run_timestamp": "2026-08-29T16:50:17.446Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6580951428571429,
      "accuracy_std_dev": 0.018418057788319512,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.026633637142857144,
      "avg_duration_per_task_ms": 383567.4857142857,
      "avg_output_tokens_per_task": 10552.454285714286,
      "last_run_timestamp": "2026-08-29T16:36:42.164Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Google",
      "accuracy": 0.6560493333333334,
      "accuracy_std_dev": 0.029495577726286477,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.12169699629629627,
      "avg_duration_per_task_ms": 120224.82148148149,
      "avg_output_tokens_per_task": 12512.395555555555,
      "last_run_timestamp": "2026-10-03T08:13:54.011Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "auto-routing",
      "accuracy": 0.6557788844221106,
      "accuracy_std_dev": 0.01812503680513505,
      "run_count": 4,
      "total_questions": 199,
      "avg_cost_per_task_usd": 0.017334296482412057,
      "avg_duration_per_task_ms": 354124.29145728645,
      "avg_output_tokens_per_task": 22535.824120603014,
      "last_run_timestamp": "2026-10-03T08:21:59.532Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Phala",
      "accuracy": 0.6555554666666666,
      "accuracy_std_dev": 0.010407070153866542,
      "run_count": 4,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.014745425641025642,
      "avg_duration_per_task_ms": 278015.8512820513,
      "avg_output_tokens_per_task": 20205.384615384617,
      "last_run_timestamp": "2026-10-03T08:06:43.564Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-fin-20260827",
      "model_slug": "inclusionai/ling-3.0-flash-fin",
      "model_name": "inclusionAI: Ling 3.0 Flash Fin",
      "provider_name": "Novita",
      "accuracy": 0.6533335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00488981,
      "avg_duration_per_task_ms": 111736.2,
      "avg_output_tokens_per_task": 17377.29,
      "last_run_timestamp": "2026-10-01T08:04:38.584Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "auto-routing",
      "accuracy": 0.6533333333333333,
      "accuracy_std_dev": 0.023726809342082957,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03210012,
      "avg_duration_per_task_ms": 143919.89333333334,
      "avg_output_tokens_per_task": 17353.273333333334,
      "last_run_timestamp": "2026-10-01T19:04:28.503Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.653333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00923216,
      "avg_duration_per_task_ms": 145753.66,
      "avg_output_tokens_per_task": 8192.38,
      "last_run_timestamp": "2026-09-23T06:00:08.549Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.65166675,
      "accuracy_std_dev": 0.012801703448662586,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.031589305,
      "avg_duration_per_task_ms": 139715.76,
      "avg_output_tokens_per_task": 17049.6,
      "last_run_timestamp": "2026-10-02T23:09:54.584Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "Together",
      "accuracy": 0.65166675,
      "accuracy_std_dev": 0.04506177549972369,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.04474220500000001,
      "avg_duration_per_task_ms": 212693.005,
      "avg_output_tokens_per_task": 3236.785,
      "last_run_timestamp": "2026-09-12T08:12:16.067Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "StreamLake",
      "accuracy": 0.6495237142857143,
      "accuracy_std_dev": 0.031240416581794835,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.021353940000000002,
      "avg_duration_per_task_ms": 237904.40285714285,
      "avg_output_tokens_per_task": 10393.82,
      "last_run_timestamp": "2026-08-29T16:17:15.093Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Azure",
      "accuracy": 0.6481482222222222,
      "accuracy_std_dev": 0.02470134810436142,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.11660464888888887,
      "avg_duration_per_task_ms": 145744.04888888888,
      "avg_output_tokens_per_task": 12598.522222222222,
      "last_run_timestamp": "2026-10-03T08:13:37.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "CoreWeave",
      "accuracy": 0.6481482020202021,
      "accuracy_std_dev": 0.0054059166891471805,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.009114631313131312,
      "avg_duration_per_task_ms": 604721.6666666666,
      "avg_output_tokens_per_task": 19746.29292929293,
      "last_run_timestamp": "2026-08-22T16:10:30.348Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "Azure",
      "accuracy": 0.6477778333333334,
      "accuracy_std_dev": 0.03430091421724631,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09916395333333333,
      "avg_duration_per_task_ms": 153090.92333333334,
      "avg_output_tokens_per_task": 12859.136666666667,
      "last_run_timestamp": "2026-10-01T19:04:42.436Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Novita",
      "accuracy": 0.6476190000000001,
      "accuracy_std_dev": 0.02121044324316276,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.02434872285714286,
      "avg_duration_per_task_ms": 154447.84,
      "avg_output_tokens_per_task": 10538.351428571428,
      "last_run_timestamp": "2026-08-29T16:49:08.574Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "AkashML",
      "accuracy": 0.647474909090909,
      "accuracy_std_dev": 0.02520056863029086,
      "run_count": 10,
      "total_questions": 495,
      "avg_cost_per_task_usd": 0.011397434343434345,
      "avg_duration_per_task_ms": 317877.1191919192,
      "avg_output_tokens_per_task": 16792.08282828283,
      "last_run_timestamp": "2026-10-09T16:52:01.244Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6474037386934673,
      "accuracy_std_dev": 0.11098825656657328,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.08922564070351757,
      "avg_duration_per_task_ms": 216302.63819095478,
      "avg_output_tokens_per_task": 23678.577889447235,
      "last_run_timestamp": "2026-10-03T16:12:24.719Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Inceptron",
      "accuracy": 0.6473332,
      "accuracy_std_dev": 0.027840964224681578,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.022036860000000002,
      "avg_duration_per_task_ms": 301895.372,
      "avg_output_tokens_per_task": 10556.636,
      "last_run_timestamp": "2026-08-15T14:15:19.430Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DigitalOcean",
      "accuracy": 0.646566095477387,
      "accuracy_std_dev": 0.12986162562313927,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.04628975628140704,
      "avg_duration_per_task_ms": 2127959.7989949747,
      "avg_output_tokens_per_task": 71164.3944723618,
      "last_run_timestamp": "2026-10-03T08:42:41.257Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Io Net",
      "accuracy": 0.646,
      "accuracy_std_dev": 0.03453813994991622,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.014338291999999999,
      "avg_duration_per_task_ms": 1162755.396,
      "avg_output_tokens_per_task": 23357.7,
      "last_run_timestamp": "2026-10-05T08:23:35.777Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Krea",
      "accuracy": 0.6455556666666666,
      "accuracy_std_dev": 0.09964144484545016,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.11177372666666668,
      "avg_duration_per_task_ms": 710817.72,
      "avg_output_tokens_per_task": 87845.5,
      "last_run_timestamp": "2026-09-29T19:45:03.655Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6454080918367348,
      "accuracy_std_dev": 0.11180720901616487,
      "run_count": 8,
      "total_questions": 392,
      "avg_cost_per_task_usd": 0.3985840484693877,
      "avg_duration_per_task_ms": 2129131.2704081633,
      "avg_output_tokens_per_task": 67490.7780612245,
      "last_run_timestamp": "2026-10-05T20:21:41.559Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "Azure",
      "accuracy": 0.6433335,
      "accuracy_std_dev": 0.014272415770172448,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.031060726666666667,
      "avg_duration_per_task_ms": 229987.72666666665,
      "avg_output_tokens_per_task": 16358.803333333333,
      "last_run_timestamp": "2026-10-01T19:05:16.543Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "OpenInference",
      "accuracy": 0.6433333333333333,
      "accuracy_std_dev": 0.14455548467015086,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.10531282,
      "avg_duration_per_task_ms": 1453976.54,
      "avg_output_tokens_per_task": 18878.353333333333,
      "last_run_timestamp": "2026-08-08T08:21:31.992Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Groq",
      "accuracy": 0.642083375,
      "accuracy_std_dev": 0.03004338703000005,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.033152745,
      "avg_duration_per_task_ms": 82992.1225,
      "avg_output_tokens_per_task": 19582.065,
      "last_run_timestamp": "2026-10-03T16:20:43.625Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Darkbloom",
      "accuracy": 0.6415409547738694,
      "accuracy_std_dev": 0.019239053075710364,
      "run_count": 4,
      "total_questions": 199,
      "avg_cost_per_task_usd": 0.012467824120603016,
      "avg_duration_per_task_ms": 963027.175879397,
      "avg_output_tokens_per_task": 22096.447236180906,
      "last_run_timestamp": "2026-10-03T16:01:59.425Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.64,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.29234208,
      "avg_duration_per_task_ms": 227202.07,
      "avg_output_tokens_per_task": 8702.01,
      "last_run_timestamp": "2026-09-01T13:00:51.629Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "ibm-granite/granite-4.2-8b-20260831",
      "model_slug": "ibm-granite/granite-4.2-8b",
      "model_name": "IBM: Granite 4.2 8B",
      "provider_name": "DeepInfra",
      "accuracy": 0.64,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01684376,
      "avg_duration_per_task_ms": 1178898.4,
      "avg_output_tokens_per_task": 43776.78,
      "last_run_timestamp": "2026-10-09T00:39:20.903Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "auto-routing",
      "accuracy": 0.6390477142857143,
      "accuracy_std_dev": 0.02969234538065464,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.018472702857142857,
      "avg_duration_per_task_ms": 329519.3171428571,
      "avg_output_tokens_per_task": 20296.345714285715,
      "last_run_timestamp": "2026-10-03T16:05:02.801Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Venice",
      "accuracy": 0.6383332500000001,
      "accuracy_std_dev": 0.028431218171008805,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.216643625,
      "avg_duration_per_task_ms": 1122861.59,
      "avg_output_tokens_per_task": 30036.18,
      "last_run_timestamp": "2026-08-22T16:17:16.779Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "inclusionai/ling-3.0-flash-fin-20260827",
      "model_slug": "inclusionai/ling-3.0-flash-fin",
      "model_name": "inclusionAI: Ling 3.0 Flash Fin",
      "provider_name": "DeepInfra",
      "accuracy": 0.638298,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 47,
      "avg_cost_per_task_usd": 0.020538510638297872,
      "avg_duration_per_task_ms": 686659.4680851063,
      "avg_output_tokens_per_task": 47094.31914893617,
      "last_run_timestamp": "2026-10-09T00:23:16.829Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.6377776666666667,
      "accuracy_std_dev": 0.04228062043326973,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08944130666666666,
      "avg_duration_per_task_ms": 118157.82666666666,
      "avg_output_tokens_per_task": 12612.513333333334,
      "last_run_timestamp": "2026-10-01T19:04:42.468Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Mara",
      "accuracy": 0.637681,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 46,
      "avg_cost_per_task_usd": 0.04568002173913044,
      "avg_duration_per_task_ms": 70889.17391304347,
      "avg_output_tokens_per_task": 8168.934782608696,
      "last_run_timestamp": "2026-08-01T17:41:39.392Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Parasail",
      "accuracy": 0.6346803232323233,
      "accuracy_std_dev": 0.01547599999999999,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.019095151515151516,
      "avg_duration_per_task_ms": 172977.9898989899,
      "avg_output_tokens_per_task": 9245.858585858587,
      "last_run_timestamp": "2026-08-01T17:40:52.342Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6336735020408163,
      "accuracy_std_dev": 0.10052381662193295,
      "run_count": 10,
      "total_questions": 490,
      "avg_cost_per_task_usd": 0.02281844081632653,
      "avg_duration_per_task_ms": 906837.293877551,
      "avg_output_tokens_per_task": 30461.944897959183,
      "last_run_timestamp": "2026-10-03T16:08:50.061Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Google",
      "accuracy": 0.632496338912134,
      "accuracy_std_dev": 0.08745539404199149,
      "run_count": 5,
      "total_questions": 239,
      "avg_cost_per_task_usd": 0.1367811171548117,
      "avg_duration_per_task_ms": 350756.1715481171,
      "avg_output_tokens_per_task": 10337.066945606695,
      "last_run_timestamp": "2026-08-29T16:44:17.688Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6314284285714286,
      "accuracy_std_dev": 0.018418242461965675,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.028592208571428572,
      "avg_duration_per_task_ms": 225611.1542857143,
      "avg_output_tokens_per_task": 10242.23142857143,
      "last_run_timestamp": "2026-08-29T16:13:54.380Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "auto-routing",
      "accuracy": 0.631111,
      "accuracy_std_dev": 0.013698950130089036,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02504141333333333,
      "avg_duration_per_task_ms": 168170.98,
      "avg_output_tokens_per_task": 11245.873333333333,
      "last_run_timestamp": "2026-08-01T17:38:10.364Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DeepInfra",
      "accuracy": 0.6300762106598985,
      "accuracy_std_dev": 0.03222693627952934,
      "run_count": 16,
      "total_questions": 788,
      "avg_cost_per_task_usd": 0.021151480964467007,
      "avg_duration_per_task_ms": 298993.4543147208,
      "avg_output_tokens_per_task": 18695.17766497462,
      "last_run_timestamp": "2026-10-09T00:49:56.319Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "CoreWeave",
      "accuracy": 0.63,
      "accuracy_std_dev": 0.023333000000000048,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07672618,
      "avg_duration_per_task_ms": 231628.06,
      "avg_output_tokens_per_task": 10580.58,
      "last_run_timestamp": "2026-08-01T17:40:48.165Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Parasail",
      "accuracy": 0.6278538630136987,
      "accuracy_std_dev": 0.07680767503894964,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.09331428767123287,
      "avg_duration_per_task_ms": 1104390.4520547944,
      "avg_output_tokens_per_task": 65652.6198630137,
      "last_run_timestamp": "2026-10-03T16:14:02.971Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.6266667499999999,
      "accuracy_std_dev": 0.012472414008021857,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.116286155,
      "avg_duration_per_task_ms": 669417.115,
      "avg_output_tokens_per_task": 30726.725,
      "last_run_timestamp": "2026-08-22T16:05:00.942Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Novita",
      "accuracy": 0.6266663333333334,
      "accuracy_std_dev": 0.018856180831641284,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02545802,
      "avg_duration_per_task_ms": 175220.38666666666,
      "avg_output_tokens_per_task": 11691.16,
      "last_run_timestamp": "2026-08-01T17:38:10.202Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Venice",
      "accuracy": 0.6234567925925926,
      "accuracy_std_dev": 0.06433938027058783,
      "run_count": 11,
      "total_questions": 540,
      "avg_cost_per_task_usd": 0.022850422222222222,
      "avg_duration_per_task_ms": 254929.7462962963,
      "avg_output_tokens_per_task": 24515.818518518518,
      "last_run_timestamp": "2026-10-03T16:12:24.693Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.6216667499999999,
      "accuracy_std_dev": 0.04997217837144482,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.08547625,
      "avg_duration_per_task_ms": 106037.93,
      "avg_output_tokens_per_task": 11996.815,
      "last_run_timestamp": "2026-10-02T23:10:27.235Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "GMICloud",
      "accuracy": 0.6211111666666667,
      "accuracy_std_dev": 0.18460855717654465,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.13787135666666667,
      "avg_duration_per_task_ms": 390996.2266666667,
      "avg_output_tokens_per_task": 8010.466666666666,
      "last_run_timestamp": "2026-10-03T16:12:26.818Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6211110000000001,
      "accuracy_std_dev": 0.041171085383636265,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09511168999999998,
      "avg_duration_per_task_ms": 55650.66,
      "avg_output_tokens_per_task": 3586.0333333333333,
      "last_run_timestamp": "2026-10-03T16:14:18.044Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Cerebras",
      "accuracy": 0.6204761428571429,
      "accuracy_std_dev": 0.025224357448582882,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.06417251142857144,
      "avg_duration_per_task_ms": 29559.602857142858,
      "avg_output_tokens_per_task": 20035.14285714286,
      "last_run_timestamp": "2026-10-03T16:18:39.939Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Mara",
      "accuracy": 0.6180555,
      "accuracy_std_dev": 0.00694450000000002,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.025814302083333334,
      "avg_duration_per_task_ms": 76875.0625,
      "avg_output_tokens_per_task": 17301.635416666668,
      "last_run_timestamp": "2026-08-29T16:33:38.021Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.6174497919463087,
      "accuracy_std_dev": 0.011290615681716505,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.11355312416107383,
      "avg_duration_per_task_ms": 671627.1577181208,
      "avg_output_tokens_per_task": 32151.05033557047,
      "last_run_timestamp": "2026-08-22T16:01:40.197Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6158248141414141,
      "accuracy_std_dev": 0.2285695922821756,
      "run_count": 10,
      "total_questions": 495,
      "avg_cost_per_task_usd": 0.020359361616161618,
      "avg_duration_per_task_ms": 340617.86262626265,
      "avg_output_tokens_per_task": 10409.925252525252,
      "last_run_timestamp": "2026-09-26T08:14:40.334Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "Novita",
      "accuracy": 0.6155553333333333,
      "accuracy_std_dev": 0.04228062043326968,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09065885333333334,
      "avg_duration_per_task_ms": 148176.83333333334,
      "avg_output_tokens_per_task": 3889.28,
      "last_run_timestamp": "2026-08-01T17:38:24.442Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Crusoe",
      "accuracy": 0.613333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01616616,
      "avg_duration_per_task_ms": 108573.58,
      "avg_output_tokens_per_task": 21305.1,
      "last_run_timestamp": "2026-10-03T16:13:54.745Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "NextBit",
      "accuracy": 0.613333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.07264287999999999,
      "avg_duration_per_task_ms": 1482714.84,
      "avg_output_tokens_per_task": 23767.86,
      "last_run_timestamp": "2026-07-18T16:00:08.935Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Google",
      "accuracy": 0.6114287142857142,
      "accuracy_std_dev": 0.0649995778815179,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.037125854285714285,
      "avg_duration_per_task_ms": 389444.8,
      "avg_output_tokens_per_task": 19927.58,
      "last_run_timestamp": "2026-10-03T16:12:10.894Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AkashML",
      "accuracy": 0.608,
      "accuracy_std_dev": 0.11488351975457575,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.058990216,
      "avg_duration_per_task_ms": 582295.388,
      "avg_output_tokens_per_task": 35784.112,
      "last_run_timestamp": "2026-08-15T14:08:35.086Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "DeepInfra",
      "accuracy": 0.6076188571428572,
      "accuracy_std_dev": 0.03274303997331158,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.046068719999999994,
      "avg_duration_per_task_ms": 191059.44,
      "avg_output_tokens_per_task": 3405.154285714286,
      "last_run_timestamp": "2026-10-08T22:20:40.622Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6-20251208",
      "model_slug": "z-ai/glm-4.6v",
      "model_name": "Z.ai: GLM 4.6V",
      "provider_name": "auto-routing",
      "accuracy": 0.606667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03212956,
      "avg_duration_per_task_ms": 335001.52,
      "avg_output_tokens_per_task": 8851.18,
      "last_run_timestamp": "2026-10-01T08:03:01.921Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "auto-routing",
      "accuracy": 0.6066668000000001,
      "accuracy_std_dev": 0.026666399999999996,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.046513716,
      "avg_duration_per_task_ms": 209556.196,
      "avg_output_tokens_per_task": 3673.824,
      "last_run_timestamp": "2026-10-01T16:00:32.452Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-pro",
      "model_slug": "google/gemini-2.5-pro",
      "model_name": "Google: Gemini 2.5 Pro",
      "provider_name": "Google",
      "accuracy": 0.6060269955357143,
      "accuracy_std_dev": 0.032348691330028045,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.19772456473214287,
      "avg_duration_per_task_ms": 188670.52455357142,
      "avg_output_tokens_per_task": 13047.183035714286,
      "last_run_timestamp": "2026-10-01T19:04:55.725Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "SambaNova",
      "accuracy": 0.605442137755102,
      "accuracy_std_dev": 0.02995845369815299,
      "run_count": 8,
      "total_questions": 392,
      "avg_cost_per_task_usd": 0.041257920918367344,
      "avg_duration_per_task_ms": 86489.80612244898,
      "avg_output_tokens_per_task": 24455.91581632653,
      "last_run_timestamp": "2026-10-03T16:12:29.878Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "StreamLake",
      "accuracy": 0.6033332222222223,
      "accuracy_std_dev": 0.027397264890973577,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.04748484000000001,
      "avg_duration_per_task_ms": 91610.96,
      "avg_output_tokens_per_task": 4621.971111111111,
      "last_run_timestamp": "2026-10-03T16:14:40.781Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Chutes",
      "accuracy": 0.603333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.030717460000000002,
      "avg_duration_per_task_ms": 422658.18,
      "avg_output_tokens_per_task": 11680.98,
      "last_run_timestamp": "2026-07-18T16:00:09.161Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Parasail",
      "accuracy": 0.5979899899497488,
      "accuracy_std_dev": 0.022095125194471644,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.020116060301507536,
      "avg_duration_per_task_ms": 100330.36180904522,
      "avg_output_tokens_per_task": 4353.20351758794,
      "last_run_timestamp": "2026-10-03T16:14:27.398Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-pro",
      "model_slug": "google/gemini-2.5-pro",
      "model_name": "Google: Gemini 2.5 Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.5933333333333334,
      "accuracy_std_dev": 0.02880349773127486,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.21649743333333332,
      "avg_duration_per_task_ms": 191120.76666666666,
      "avg_output_tokens_per_task": 14122.313333333334,
      "last_run_timestamp": "2026-10-01T19:04:42.461Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "BaseTen",
      "accuracy": 0.593333,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04552234,
      "avg_duration_per_task_ms": 32685.39,
      "avg_output_tokens_per_task": 3125.98,
      "last_run_timestamp": "2026-09-26T08:22:32.737Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6-20251208",
      "model_slug": "z-ai/glm-4.6v",
      "model_name": "Z.ai: GLM 4.6V",
      "provider_name": "Z.AI",
      "accuracy": 0.593333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01896666,
      "avg_duration_per_task_ms": 306972.98,
      "avg_output_tokens_per_task": 7436.64,
      "last_run_timestamp": "2026-10-01T08:02:59.889Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Novita",
      "accuracy": 0.5913978951612903,
      "accuracy_std_dev": 0.012203964545998971,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.156355625,
      "avg_duration_per_task_ms": 554104.3709677419,
      "avg_output_tokens_per_task": 27726.576612903227,
      "last_run_timestamp": "2026-08-22T16:20:42.106Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "Google",
      "accuracy": 0.5881481111111111,
      "accuracy_std_dev": 0.007554077566224961,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.029889999999999996,
      "avg_duration_per_task_ms": 78610.16,
      "avg_output_tokens_per_task": 8108.122222222222,
      "last_run_timestamp": "2026-10-01T19:05:46.198Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "ibm-granite/granite-4.2-8b-20260831",
      "model_slug": "ibm-granite/granite-4.2-8b",
      "model_name": "IBM: Granite 4.2 8B",
      "provider_name": "auto-routing",
      "accuracy": 0.583333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.015297799999999999,
      "avg_duration_per_task_ms": 833284.24,
      "avg_output_tokens_per_task": 38596.74,
      "last_run_timestamp": "2026-10-01T08:04:13.086Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.2-chat-20251211",
      "model_slug": "openai/gpt-5.2-chat",
      "model_name": "OpenAI: GPT-5.2 Chat",
      "provider_name": "OpenAI",
      "accuracy": 0.58,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.10846708,
      "avg_duration_per_task_ms": 79571.2,
      "avg_output_tokens_per_task": 4199.82,
      "last_run_timestamp": "2026-08-01T08:05:35.385Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Alibaba",
      "accuracy": 0.579037868556701,
      "accuracy_std_dev": 0.04024006578337068,
      "run_count": 8,
      "total_questions": 388,
      "avg_cost_per_task_usd": 0.07504983247422681,
      "avg_duration_per_task_ms": 90151.55927835051,
      "avg_output_tokens_per_task": 5664.682989690722,
      "last_run_timestamp": "2026-10-03T16:05:18.093Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "auto-routing",
      "accuracy": 0.572072,
      "accuracy_std_dev": 0.02796372233607127,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.09883952702702703,
      "avg_duration_per_task_ms": 105307.2027027027,
      "avg_output_tokens_per_task": 3660.560810810811,
      "last_run_timestamp": "2026-08-01T17:38:23.537Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.571111,
      "accuracy_std_dev": 0.008314764498569157,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.028662713333333336,
      "avg_duration_per_task_ms": 78035.85333333333,
      "avg_output_tokens_per_task": 8001.593333333333,
      "last_run_timestamp": "2026-10-01T19:04:40.340Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Novita",
      "accuracy": 0.5693332,
      "accuracy_std_dev": 0.03832036650346651,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.0105344,
      "avg_duration_per_task_ms": 254767.968,
      "avg_output_tokens_per_task": 16832.444,
      "last_run_timestamp": "2026-08-29T16:33:38.124Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Cerebras",
      "accuracy": 0.5648772953020134,
      "accuracy_std_dev": 0.0424873867714486,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.1486826308724832,
      "avg_duration_per_task_ms": 18002.88590604027,
      "avg_output_tokens_per_task": 2546.3288590604025,
      "last_run_timestamp": "2026-08-29T17:07:32.456Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "Novita",
      "accuracy": 0.5641890743243243,
      "accuracy_std_dev": 0.04318929456634928,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.18690275337837842,
      "avg_duration_per_task_ms": 1880221.320945946,
      "avg_output_tokens_per_task": 41218.97635135135,
      "last_run_timestamp": "2026-10-03T16:17:09.625Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-pro",
      "model_slug": "google/gemini-2.5-pro",
      "model_name": "Google: Gemini 2.5 Pro",
      "provider_name": "Google AI Studio",
      "accuracy": 0.5622223333333333,
      "accuracy_std_dev": 0.03456997912768952,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.23936231333333335,
      "avg_duration_per_task_ms": 165980.19333333333,
      "avg_output_tokens_per_task": 13303.673333333334,
      "last_run_timestamp": "2026-10-01T19:04:52.649Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "auto-routing",
      "accuracy": 0.5574712183908046,
      "accuracy_std_dev": 0.07593309995634047,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.05341043390804598,
      "avg_duration_per_task_ms": 469408.1063218391,
      "avg_output_tokens_per_task": 9846.341954022988,
      "last_run_timestamp": "2026-10-04T00:00:28.213Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Mancer 2",
      "accuracy": 0.5572390808080808,
      "accuracy_std_dev": 0.08974967405616299,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.04370982828282828,
      "avg_duration_per_task_ms": 3081206.1666666665,
      "avg_output_tokens_per_task": 93737.79292929293,
      "last_run_timestamp": "2026-10-03T16:33:53.219Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Darkbloom",
      "accuracy": 0.5521540748299321,
      "accuracy_std_dev": 0.03338584001572458,
      "run_count": 3,
      "total_questions": 147,
      "avg_cost_per_task_usd": 0.0026003809523809525,
      "avg_duration_per_task_ms": 390785.49659863947,
      "avg_output_tokens_per_task": 6412.292517006803,
      "last_run_timestamp": "2026-10-03T16:14:16.049Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.5511113333333334,
      "accuracy_std_dev": 0.006285236475706844,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.035753513333333334,
      "avg_duration_per_task_ms": 61316.40666666667,
      "avg_output_tokens_per_task": 6659.566666666667,
      "last_run_timestamp": "2026-10-01T19:04:42.483Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "auto-routing",
      "accuracy": 0.54833325,
      "accuracy_std_dev": 0.09993065790305547,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0400913925,
      "avg_duration_per_task_ms": 97587.1625,
      "avg_output_tokens_per_task": 4289.61,
      "last_run_timestamp": "2026-10-03T16:07:50.329Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.5416667142857143,
      "accuracy_std_dev": 0.10219251042847255,
      "run_count": 7,
      "total_questions": 336,
      "avg_cost_per_task_usd": 0.06745207738095238,
      "avg_duration_per_task_ms": 446328.3333333333,
      "avg_output_tokens_per_task": 28269.869047619046,
      "last_run_timestamp": "2026-08-29T08:10:31.240Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.5409524285714286,
      "accuracy_std_dev": 0.098072099444173,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.027534851428571427,
      "avg_duration_per_task_ms": 272348.91142857145,
      "avg_output_tokens_per_task": 7453.248571428571,
      "last_run_timestamp": "2026-10-03T16:13:14.992Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "auto-routing",
      "accuracy": 0.5351473333333333,
      "accuracy_std_dev": 0.016949235571913614,
      "run_count": 6,
      "total_questions": 294,
      "avg_cost_per_task_usd": 0.025283034013605446,
      "avg_duration_per_task_ms": 1563224.0068027212,
      "avg_output_tokens_per_task": 123876.9387755102,
      "last_run_timestamp": "2026-10-03T16:10:46.805Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Nebius",
      "accuracy": 0.5333335,
      "accuracy_std_dev": 0.026666500000000037,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.10219375,
      "avg_duration_per_task_ms": 375443.12,
      "avg_output_tokens_per_task": 17495.06,
      "last_run_timestamp": "2026-08-15T14:16:11.776Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "WandB",
      "accuracy": 0.526667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02328844,
      "avg_duration_per_task_ms": 1276910.42,
      "avg_output_tokens_per_task": 153904.82,
      "last_run_timestamp": "2026-07-18T16:00:10.167Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Fireworks",
      "accuracy": 0.526667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03416076,
      "avg_duration_per_task_ms": 1845834.04,
      "avg_output_tokens_per_task": 105339.92,
      "last_run_timestamp": "2026-07-18T16:00:08.888Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "BaseTen",
      "accuracy": 0.523035243902439,
      "accuracy_std_dev": 0.26305710961614404,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.01812669918699187,
      "avg_duration_per_task_ms": 110026.72764227643,
      "avg_output_tokens_per_task": 18887.439024390245,
      "last_run_timestamp": "2026-08-29T17:06:47.726Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "Azure",
      "accuracy": 0.5221089591836734,
      "accuracy_std_dev": 0.023889008586586408,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.13644156632653062,
      "avg_duration_per_task_ms": 73125.92346938775,
      "avg_output_tokens_per_task": 2553.7397959183672,
      "last_run_timestamp": "2026-10-01T19:05:18.003Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.5201006834170854,
      "accuracy_std_dev": 0.01697055860112964,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.06560986683417085,
      "avg_duration_per_task_ms": 298265.93216080405,
      "avg_output_tokens_per_task": 4016.8693467336684,
      "last_run_timestamp": "2026-10-04T00:00:34.892Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.517500125,
      "accuracy_std_dev": 0.024930426803594342,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0282700725,
      "avg_duration_per_task_ms": 444815,
      "avg_output_tokens_per_task": 58343.125,
      "last_run_timestamp": "2026-10-03T16:09:19.689Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.5156599127516779,
      "accuracy_std_dev": 0.03449671802746843,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.024214489932885906,
      "avg_duration_per_task_ms": 190286.47651006712,
      "avg_output_tokens_per_task": 48937.10738255033,
      "last_run_timestamp": "2026-10-01T19:05:00.102Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6-20251208",
      "model_slug": "z-ai/glm-4.6v",
      "model_name": "Z.ai: GLM 4.6V",
      "provider_name": "Novita",
      "accuracy": 0.513333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03238348,
      "avg_duration_per_task_ms": 271942.88,
      "avg_output_tokens_per_task": 7084.96,
      "last_run_timestamp": "2026-10-01T08:05:50.069Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "DeepInfra",
      "accuracy": 0.5104164999999999,
      "accuracy_std_dev": 0.003472500000000045,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.024820281250000003,
      "avg_duration_per_task_ms": 480070.4583333333,
      "avg_output_tokens_per_task": 9578.8125,
      "last_run_timestamp": "2026-08-08T16:03:43.735Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "OpenAI",
      "accuracy": 0.5083332500000001,
      "accuracy_std_dev": 0.026404802805313633,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.12635733,
      "avg_duration_per_task_ms": 53878.245,
      "avg_output_tokens_per_task": 2800.47,
      "last_run_timestamp": "2026-10-02T23:07:04.603Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.5045660136986302,
      "accuracy_std_dev": 0.02732875245760204,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.05861860273972602,
      "avg_duration_per_task_ms": 2642554.2671232875,
      "avg_output_tokens_per_task": 40736.36301369863,
      "last_run_timestamp": "2026-08-15T14:12:30.435Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.5033670303030303,
      "accuracy_std_dev": 0.030340499999999992,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.01769851515151515,
      "avg_duration_per_task_ms": 135563.53535353535,
      "avg_output_tokens_per_task": 30988.575757575756,
      "last_run_timestamp": "2026-10-01T19:04:42.466Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.5033669747474748,
      "accuracy_std_dev": 0.027292446564651823,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.025276146464646466,
      "avg_duration_per_task_ms": 577739.3712121212,
      "avg_output_tokens_per_task": 53598.469696969696,
      "last_run_timestamp": "2026-09-26T16:00:25.689Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DeepInfra",
      "accuracy": 0.5013551951219513,
      "accuracy_std_dev": 0.01245431828082131,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.020900845528455284,
      "avg_duration_per_task_ms": 1272391.512195122,
      "avg_output_tokens_per_task": 112484.65853658537,
      "last_run_timestamp": "2026-08-29T17:16:15.676Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "auto-routing",
      "accuracy": 0.5011262027027026,
      "accuracy_std_dev": 0.034811519447191935,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.15286014527027028,
      "avg_duration_per_task_ms": 1793120.7567567567,
      "avg_output_tokens_per_task": 35316.18243243243,
      "last_run_timestamp": "2026-10-03T16:05:51.297Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Together",
      "accuracy": 0.49958325000000003,
      "accuracy_std_dev": 0.03216451093888264,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.026658079999999997,
      "avg_duration_per_task_ms": 280333.52,
      "avg_output_tokens_per_task": 19065.1925,
      "last_run_timestamp": "2026-10-03T16:15:51.386Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "CoreWeave",
      "accuracy": 0.49666650000000007,
      "accuracy_std_dev": 0.023333500000000007,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.23109727999999996,
      "avg_duration_per_task_ms": 80168.37,
      "avg_output_tokens_per_task": 2746.03,
      "last_run_timestamp": "2026-08-01T17:41:27.708Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Mancer 2",
      "accuracy": 0.493333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.14235776,
      "avg_duration_per_task_ms": 443941.74,
      "avg_output_tokens_per_task": 8833.66,
      "last_run_timestamp": "2026-08-06T00:00:02.908Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Parasail",
      "accuracy": 0.489899005050505,
      "accuracy_std_dev": 0.009452416119701875,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.014275666666666666,
      "avg_duration_per_task_ms": 117584.4797979798,
      "avg_output_tokens_per_task": 3141.5757575757575,
      "last_run_timestamp": "2026-10-03T16:05:51.297Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "SiliconFlow",
      "accuracy": 0.48888899999999996,
      "accuracy_std_dev": 0.01872462563043652,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06696013,
      "avg_duration_per_task_ms": 315550.43333333335,
      "avg_output_tokens_per_task": 4089.9266666666667,
      "last_run_timestamp": "2026-08-29T16:26:59.974Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "Nebius",
      "accuracy": 0.48833333333333334,
      "accuracy_std_dev": 0.018831396165859726,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03323320666666667,
      "avg_duration_per_task_ms": 665027.6466666666,
      "avg_output_tokens_per_task": 61694.416666666664,
      "last_run_timestamp": "2026-10-03T16:04:32.703Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "Crusoe",
      "accuracy": 0.4852150080645161,
      "accuracy_std_dev": 0.023420064714684302,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.028619552419354842,
      "avg_duration_per_task_ms": 386610.2439516129,
      "avg_output_tokens_per_task": 66355.3125,
      "last_run_timestamp": "2026-10-03T16:14:03.053Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Nebius",
      "accuracy": 0.4842343648648649,
      "accuracy_std_dev": 0.02401659838297026,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.023849905405405405,
      "avg_duration_per_task_ms": 836568.3851351351,
      "avg_output_tokens_per_task": 22266.58108108108,
      "last_run_timestamp": "2026-08-29T16:13:54.382Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-3-12b-it",
      "model_slug": "google/gemma-3-12b-it",
      "model_name": "Google: Gemma 3 12B",
      "provider_name": "DeepInfra",
      "accuracy": 0.48,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.011001320000000002,
      "avg_duration_per_task_ms": 80205.64,
      "avg_output_tokens_per_task": 3762.88,
      "last_run_timestamp": "2026-10-08T22:08:03.352Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Google",
      "accuracy": 0.48,
      "accuracy_std_dev": 0.0241687453956551,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.034828557142857146,
      "avg_duration_per_task_ms": 43808.188571428575,
      "avg_output_tokens_per_task": 3663.8714285714286,
      "last_run_timestamp": "2026-10-03T16:17:11.217Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.48,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.07899766,
      "avg_duration_per_task_ms": 132704.5,
      "avg_output_tokens_per_task": 3161.38,
      "last_run_timestamp": "2026-10-04T03:00:16.401Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Crusoe",
      "accuracy": 0.47916700000000007,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.04362447916666667,
      "avg_duration_per_task_ms": 66953.94791666667,
      "avg_output_tokens_per_task": 2828.96875,
      "last_run_timestamp": "2026-08-15T14:08:14.681Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "auto-routing",
      "accuracy": 0.4777776666666667,
      "accuracy_std_dev": 0.013698571612479242,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12510532,
      "avg_duration_per_task_ms": 53067.566666666666,
      "avg_output_tokens_per_task": 2709.2933333333335,
      "last_run_timestamp": "2026-10-01T19:04:26.703Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4776285234899329,
      "accuracy_std_dev": 0.010657414972163225,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.1647015033557047,
      "avg_duration_per_task_ms": 2471417.852348993,
      "avg_output_tokens_per_task": 44160.27181208054,
      "last_run_timestamp": "2026-10-03T16:12:22.579Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.47583349999999996,
      "accuracy_std_dev": 0.03650536870790377,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.02416941,
      "avg_duration_per_task_ms": 59754.7125,
      "avg_output_tokens_per_task": 3775.0125,
      "last_run_timestamp": "2026-10-03T16:10:49.378Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.47333325000000004,
      "accuracy_std_dev": 0.025603927651583064,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0251383825,
      "avg_duration_per_task_ms": 126301.635,
      "avg_output_tokens_per_task": 3791.835,
      "last_run_timestamp": "2026-10-03T16:13:56.619Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.3-chat-20260303",
      "model_slug": "openai/gpt-5.3-chat",
      "model_name": "OpenAI: GPT-5.3 Chat",
      "provider_name": "OpenAI",
      "accuracy": 0.473333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.09545477999999999,
      "avg_duration_per_task_ms": 59370.42,
      "avg_output_tokens_per_task": 2482.38,
      "last_run_timestamp": "2026-08-01T08:06:06.635Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Alibaba",
      "accuracy": 0.473333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03429832,
      "avg_duration_per_task_ms": 458175.58,
      "avg_output_tokens_per_task": 37835.14,
      "last_run_timestamp": "2026-07-18T16:00:08.521Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Parasail",
      "accuracy": 0.4727888571428572,
      "accuracy_std_dev": 0.053216013468691904,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.026510076530612242,
      "avg_duration_per_task_ms": 2225913.1275510206,
      "avg_output_tokens_per_task": 135590.98469387754,
      "last_run_timestamp": "2026-10-03T16:09:05.687Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "WandB",
      "accuracy": 0.472222,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.02306027083333333,
      "avg_duration_per_task_ms": 25624.729166666668,
      "avg_output_tokens_per_task": 1174.2916666666667,
      "last_run_timestamp": "2026-07-18T16:00:08.906Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.4721422449567723,
      "accuracy_std_dev": 0.09390832636273526,
      "run_count": 14,
      "total_questions": 694,
      "avg_cost_per_task_usd": 0.04128501008645533,
      "avg_duration_per_task_ms": 719831.6239193083,
      "avg_output_tokens_per_task": 196500.71037463978,
      "last_run_timestamp": "2026-10-03T16:19:24.321Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "auto-routing",
      "accuracy": 0.47047614285714284,
      "accuracy_std_dev": 0.020426164806293848,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.021544591428571434,
      "avg_duration_per_task_ms": 680725.9685714286,
      "avg_output_tokens_per_task": 21489.04857142857,
      "last_run_timestamp": "2026-08-29T16:16:50.456Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.46952399999999994,
      "accuracy_std_dev": 0.034523096873170074,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.018877348571428573,
      "avg_duration_per_task_ms": 90973.66571428571,
      "avg_output_tokens_per_task": 3606.62,
      "last_run_timestamp": "2026-10-03T16:12:20.062Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "auto-routing",
      "accuracy": 0.46833325000000003,
      "accuracy_std_dev": 0.03295620410389368,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.09710940000000001,
      "avg_duration_per_task_ms": 99117.8225,
      "avg_output_tokens_per_task": 2749.945,
      "last_run_timestamp": "2026-10-03T16:07:50.332Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Groq",
      "accuracy": 0.4682815259515572,
      "accuracy_std_dev": 0.039807100596362034,
      "run_count": 6,
      "total_questions": 289,
      "avg_cost_per_task_usd": 0.04939322491349481,
      "avg_duration_per_task_ms": 229004.3321799308,
      "avg_output_tokens_per_task": 98239.9169550173,
      "last_run_timestamp": "2026-10-03T16:14:31.002Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "auto-routing",
      "accuracy": 0.4662164729729729,
      "accuracy_std_dev": 0.03892660093903227,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.04219703378378378,
      "avg_duration_per_task_ms": 711257.6756756756,
      "avg_output_tokens_per_task": 98782.22297297297,
      "last_run_timestamp": "2026-10-01T19:04:24.509Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.465926,
      "accuracy_std_dev": 0.03230940398363575,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.022285648888888884,
      "avg_duration_per_task_ms": 222200.71555555557,
      "avg_output_tokens_per_task": 3598.428888888889,
      "last_run_timestamp": "2026-10-08T22:47:27.753Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "DigitalOcean",
      "accuracy": 0.46444433333333335,
      "accuracy_std_dev": 0.008314853589945066,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.033371,
      "avg_duration_per_task_ms": 305705.87333333335,
      "avg_output_tokens_per_task": 3594.2266666666665,
      "last_run_timestamp": "2026-08-15T08:15:42.180Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "Azure",
      "accuracy": 0.463333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.26064434,
      "avg_duration_per_task_ms": 31467.4,
      "avg_output_tokens_per_task": 1819.56,
      "last_run_timestamp": "2026-08-01T08:06:06.230Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "Azure",
      "accuracy": 0.4608501677852349,
      "accuracy_std_dev": 0.034994868451198195,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.04671855033557047,
      "avg_duration_per_task_ms": 767600.1241610738,
      "avg_output_tokens_per_task": 101728.0033557047,
      "last_run_timestamp": "2026-10-01T19:06:01.717Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.46068351282051284,
      "accuracy_std_dev": 0.04225414673955088,
      "run_count": 4,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.01900871794871795,
      "avg_duration_per_task_ms": 170098.1794871795,
      "avg_output_tokens_per_task": 35854.89743589744,
      "last_run_timestamp": "2026-10-01T19:06:10.500Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "FakeProvider",
      "accuracy": 0.46,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": null,
      "avg_duration_per_task_ms": 158124.84,
      "avg_output_tokens_per_task": 4515,
      "last_run_timestamp": "2026-10-01T19:06:51.244Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.3-chat-20260303",
      "model_slug": "openai/gpt-5.3-chat",
      "model_name": "OpenAI: GPT-5.3 Chat",
      "provider_name": "auto-routing",
      "accuracy": 0.46,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.09873096,
      "avg_duration_per_task_ms": 65244.1,
      "avg_output_tokens_per_task": 2748.04,
      "last_run_timestamp": "2026-08-01T08:05:59.441Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Nebius",
      "accuracy": 0.45958325,
      "accuracy_std_dev": 0.040771748628032865,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.052828784999999996,
      "avg_duration_per_task_ms": 158699.07,
      "avg_output_tokens_per_task": 3758.8575,
      "last_run_timestamp": "2026-10-03T16:12:33.495Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.45777799999999996,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.00622334,
      "avg_duration_per_task_ms": 39170.17333333333,
      "avg_output_tokens_per_task": 1174.36,
      "last_run_timestamp": "2026-08-15T08:16:13.926Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "StreamLake",
      "accuracy": 0.4577777777777778,
      "accuracy_std_dev": 0.04455537267959145,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05422914222222222,
      "avg_duration_per_task_ms": 126062.82666666666,
      "avg_output_tokens_per_task": 3580.0733333333333,
      "last_run_timestamp": "2026-10-03T16:19:54.632Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Google",
      "accuracy": 0.4570707575757576,
      "accuracy_std_dev": 0.044267226719521297,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.05378886363636363,
      "avg_duration_per_task_ms": 73384.54040404041,
      "avg_output_tokens_per_task": 3408.1742424242425,
      "last_run_timestamp": "2026-10-03T16:07:05.221Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Google",
      "accuracy": 0.4560182083333333,
      "accuracy_std_dev": 0.022912149746368182,
      "run_count": 3,
      "total_questions": 144,
      "avg_cost_per_task_usd": 0.08057918055555555,
      "avg_duration_per_task_ms": 52530.756944444445,
      "avg_output_tokens_per_task": 3473.4930555555557,
      "last_run_timestamp": "2026-08-15T14:15:58.674Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.4511785656565656,
      "accuracy_std_dev": 0.03515039216708685,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.04521708585858586,
      "avg_duration_per_task_ms": 779859.6717171717,
      "avg_output_tokens_per_task": 105433.07575757576,
      "last_run_timestamp": "2026-10-02T23:22:06.591Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "DeepInfra",
      "accuracy": 0.45095228571428575,
      "accuracy_std_dev": 0.02741314960126298,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.020993765714285715,
      "avg_duration_per_task_ms": 566234.7285714286,
      "avg_output_tokens_per_task": 21154.594285714287,
      "last_run_timestamp": "2026-10-09T00:21:42.483Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.45,
      "accuracy_std_dev": 0.01784692109767581,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03456686,
      "avg_duration_per_task_ms": 476526.7733333333,
      "avg_output_tokens_per_task": 23360.6,
      "last_run_timestamp": "2026-10-08T23:09:09.727Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.448889,
      "accuracy_std_dev": 0.0031423825355930196,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08242190666666667,
      "avg_duration_per_task_ms": 81216.82666666666,
      "avg_output_tokens_per_task": 4095.7466666666664,
      "last_run_timestamp": "2026-08-15T08:11:06.981Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Novita",
      "accuracy": 0.44750012499999997,
      "accuracy_std_dev": 0.06501608615073176,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.16146158749999998,
      "avg_duration_per_task_ms": 120680.0125,
      "avg_output_tokens_per_task": 3281.1225,
      "last_run_timestamp": "2026-10-03T16:19:14.095Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Venice",
      "accuracy": 0.445416625,
      "accuracy_std_dev": 0.03036532482099237,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0408187325,
      "avg_duration_per_task_ms": 287335.82,
      "avg_output_tokens_per_task": 3635.3075,
      "last_run_timestamp": "2026-10-03T16:17:30.535Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.4444443333333334,
      "accuracy_std_dev": 0.030952277087298266,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03472188,
      "avg_duration_per_task_ms": 283560.81333333335,
      "avg_output_tokens_per_task": 21592.06,
      "last_run_timestamp": "2026-08-01T17:38:29.532Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.4441665,
      "accuracy_std_dev": 0.023555566693459103,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.03439918,
      "avg_duration_per_task_ms": 113438.9075,
      "avg_output_tokens_per_task": 3146.3175,
      "last_run_timestamp": "2026-10-03T16:18:20.377Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.44333325,
      "accuracy_std_dev": 0.01914841159698371,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.025700495,
      "avg_duration_per_task_ms": 61023.305,
      "avg_output_tokens_per_task": 2441.375,
      "last_run_timestamp": "2026-10-02T23:07:02.872Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "DigitalOcean",
      "accuracy": 0.44186027325581395,
      "accuracy_std_dev": 0.0336023614676719,
      "run_count": 7,
      "total_questions": 344,
      "avg_cost_per_task_usd": 0.05302273255813953,
      "avg_duration_per_task_ms": 228343.90697674418,
      "avg_output_tokens_per_task": 2116.357558139535,
      "last_run_timestamp": "2026-10-05T20:09:54.588Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4416665,
      "accuracy_std_dev": 0.017873086533948188,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.104832345,
      "avg_duration_per_task_ms": 890298.59,
      "avg_output_tokens_per_task": 21334.155,
      "last_run_timestamp": "2026-08-29T16:10:09.206Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "OpenAI",
      "accuracy": 0.44,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.7291186999999999,
      "avg_duration_per_task_ms": 35507.54,
      "avg_output_tokens_per_task": 3272.36,
      "last_run_timestamp": "2026-08-01T08:05:57.702Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "WandB",
      "accuracy": 0.44,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.2288897,
      "avg_duration_per_task_ms": 83836.68,
      "avg_output_tokens_per_task": 2734.56,
      "last_run_timestamp": "2026-07-18T08:04:08.872Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Google",
      "accuracy": 0.43999974999999997,
      "accuracy_std_dev": 0.047081383440671115,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.050863820000000004,
      "avg_duration_per_task_ms": 37018.795,
      "avg_output_tokens_per_task": 2487.295,
      "last_run_timestamp": "2026-10-03T16:12:29.903Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Venice",
      "accuracy": 0.43869733333333333,
      "accuracy_std_dev": 0.023050038409813497,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.01657819540229885,
      "avg_duration_per_task_ms": 200645.3132183908,
      "avg_output_tokens_per_task": 2785.6264367816093,
      "last_run_timestamp": "2026-10-03T16:14:29.926Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Io Net",
      "accuracy": 0.4383965367088608,
      "accuracy_std_dev": 0.03960836597472182,
      "run_count": 8,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.05240046329113924,
      "avg_duration_per_task_ms": 422067.2658227848,
      "avg_output_tokens_per_task": 20533.427848101266,
      "last_run_timestamp": "2026-09-19T16:00:30.864Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "auto-routing",
      "accuracy": 0.43777799999999994,
      "accuracy_std_dev": 0.03861798340151904,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.19450383999999998,
      "avg_duration_per_task_ms": 50798.87333333334,
      "avg_output_tokens_per_task": 2047.5333333333333,
      "last_run_timestamp": "2026-10-01T16:01:13.623Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.4377776666666667,
      "accuracy_std_dev": 0.011330936570096718,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02595394666666667,
      "avg_duration_per_task_ms": 66273.3,
      "avg_output_tokens_per_task": 2407.2933333333335,
      "last_run_timestamp": "2026-10-01T19:04:38.295Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-2024-08-06",
      "model_slug": "openai/gpt-4o-2024-08-06",
      "model_name": "OpenAI: GPT-4o (2024-08-06)",
      "provider_name": "auto-routing",
      "accuracy": 0.4377776666666667,
      "accuracy_std_dev": 0.016629618082071397,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.21387673999999998,
      "avg_duration_per_task_ms": 47267.2,
      "avg_output_tokens_per_task": 2224.56,
      "last_run_timestamp": "2026-10-01T16:00:49.880Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "DeepInfra",
      "accuracy": 0.4372223333333333,
      "accuracy_std_dev": 0.025270150254576823,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.027693089999999997,
      "avg_duration_per_task_ms": 78315.11,
      "avg_output_tokens_per_task": 2793.7166666666667,
      "last_run_timestamp": "2026-10-03T16:12:20.034Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "mistralai/mistral-small-2603",
      "model_slug": "mistralai/mistral-small-2603",
      "model_name": "Mistral: Mistral Small 4",
      "provider_name": "auto-routing",
      "accuracy": 0.43666650000000007,
      "accuracy_std_dev": 0.023333500000000007,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.014376880000000002,
      "avg_duration_per_task_ms": 93606.58,
      "avg_output_tokens_per_task": 10159.34,
      "last_run_timestamp": "2026-09-01T13:02:23.335Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "auto-routing",
      "accuracy": 0.4362416308724832,
      "accuracy_std_dev": 0.02031037140436711,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.0464321610738255,
      "avg_duration_per_task_ms": 136873.4798657718,
      "avg_output_tokens_per_task": 2090.986577181208,
      "last_run_timestamp": "2026-08-29T16:33:29.885Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "auto-routing",
      "accuracy": 0.435,
      "accuracy_std_dev": 0.020749805890658335,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.074401825,
      "avg_duration_per_task_ms": 778012.545,
      "avg_output_tokens_per_task": 20313.29,
      "last_run_timestamp": "2026-08-29T16:11:46.632Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "DeepInfra",
      "accuracy": 0.43434354545454545,
      "accuracy_std_dev": 0.022037645087440712,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.06906736868686869,
      "avg_duration_per_task_ms": 323353.22222222225,
      "avg_output_tokens_per_task": 3378.3585858585857,
      "last_run_timestamp": "2026-10-08T22:54:33.845Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Alibaba",
      "accuracy": 0.43416675,
      "accuracy_std_dev": 0.04222137919570487,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.400907955,
      "avg_duration_per_task_ms": 107994.6975,
      "avg_output_tokens_per_task": 3151.7075,
      "last_run_timestamp": "2026-10-03T16:14:53.189Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "OpenAI",
      "accuracy": 0.43416662500000003,
      "accuracy_std_dev": 0.01613729505630899,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.2170375225,
      "avg_duration_per_task_ms": 46329.7025,
      "avg_output_tokens_per_task": 2265.4175,
      "last_run_timestamp": "2026-10-02T23:10:05.271Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "DeepInfra",
      "accuracy": 0.4337431393442623,
      "accuracy_std_dev": 0.021677717946315293,
      "run_count": 5,
      "total_questions": 244,
      "avg_cost_per_task_usd": 0.032824139344262294,
      "avg_duration_per_task_ms": 478154.40573770495,
      "avg_output_tokens_per_task": 4694.745901639344,
      "last_run_timestamp": "2026-08-22T16:13:43.431Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.4311113333333333,
      "accuracy_std_dev": 0.006285236475706817,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.023601506666666664,
      "avg_duration_per_task_ms": 216917.58666666667,
      "avg_output_tokens_per_task": 3146.213333333333,
      "last_run_timestamp": "2026-08-15T08:02:48.904Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "auto-routing",
      "accuracy": 0.431111,
      "accuracy_std_dev": 0.013698950130089036,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.029769013333333334,
      "avg_duration_per_task_ms": 630152.1266666667,
      "avg_output_tokens_per_task": 18760.18666666667,
      "last_run_timestamp": "2026-10-01T16:00:49.878Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-4-maverick-17b-128e-instruct",
      "model_slug": "meta-llama/llama-4-maverick",
      "model_name": "Meta: Llama 4 Maverick",
      "provider_name": "Parasail",
      "accuracy": 0.43089426016260163,
      "accuracy_std_dev": 0.03483555910043644,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.04202976422764227,
      "avg_duration_per_task_ms": 58240.52032520325,
      "avg_output_tokens_per_task": 2134.682926829268,
      "last_run_timestamp": "2026-08-29T16:52:55.844Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "DeepInfra",
      "accuracy": 0.43083325,
      "accuracy_std_dev": 0.031213896924086562,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.029063365,
      "avg_duration_per_task_ms": 761962.965,
      "avg_output_tokens_per_task": 19001.86,
      "last_run_timestamp": "2026-10-08T22:41:10.664Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-2024-08-06",
      "model_slug": "openai/gpt-4o-2024-08-06",
      "model_name": "OpenAI: GPT-4o (2024-08-06)",
      "provider_name": "OpenAI",
      "accuracy": 0.43,
      "accuracy_std_dev": 0.0033330000000000026,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.24811575000000002,
      "avg_duration_per_task_ms": 46287.18,
      "avg_output_tokens_per_task": 2565.8,
      "last_run_timestamp": "2026-10-02T23:09:02.629Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "DeepInfra",
      "accuracy": 0.42787744843049325,
      "accuracy_std_dev": 0.02999847077995481,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.024826468609865468,
      "avg_duration_per_task_ms": 337438.9708520179,
      "avg_output_tokens_per_task": 3556.2600896860986,
      "last_run_timestamp": "2026-10-09T00:15:21.434Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.42666699999999996,
      "accuracy_std_dev": 0.04000000000000001,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.031438360000000005,
      "avg_duration_per_task_ms": 468393.89,
      "avg_output_tokens_per_task": 22584.78,
      "last_run_timestamp": "2026-08-01T17:38:32.120Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Novita",
      "accuracy": 0.425833375,
      "accuracy_std_dev": 0.013616710018737092,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.024625232500000004,
      "avg_duration_per_task_ms": 228841.615,
      "avg_output_tokens_per_task": 3561.265,
      "last_run_timestamp": "2026-10-03T16:13:22.758Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "AtlasCloud",
      "accuracy": 0.4247621428571428,
      "accuracy_std_dev": 0.04058525139022997,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.05475465714285714,
      "avg_duration_per_task_ms": 141536.79714285713,
      "avg_output_tokens_per_task": 3538.397142857143,
      "last_run_timestamp": "2026-08-29T16:12:10.467Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.42144822950819677,
      "accuracy_std_dev": 0.025118027628776907,
      "run_count": 10,
      "total_questions": 488,
      "avg_cost_per_task_usd": 0.050889221311475406,
      "avg_duration_per_task_ms": 182960.04508196723,
      "avg_output_tokens_per_task": 4075.3032786885246,
      "last_run_timestamp": "2026-10-03T08:12:06.170Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "StreamLake",
      "accuracy": 0.4166667045454545,
      "accuracy_std_dev": 0.029461897667544087,
      "run_count": 9,
      "total_questions": 440,
      "avg_cost_per_task_usd": 0.011195077272727273,
      "avg_duration_per_task_ms": 112204.60909090909,
      "avg_output_tokens_per_task": 3271.315909090909,
      "last_run_timestamp": "2026-10-03T16:17:21.424Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "CoreWeave",
      "accuracy": 0.4166667027027027,
      "accuracy_std_dev": 0.06529892179218752,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.23566864189189185,
      "avg_duration_per_task_ms": 251017.3918918919,
      "avg_output_tokens_per_task": 9861.692567567568,
      "last_run_timestamp": "2026-10-04T00:00:34.897Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.4166666,
      "accuracy_std_dev": 0.024435281744590136,
      "run_count": 8,
      "total_questions": 390,
      "avg_cost_per_task_usd": 0.03287333333333334,
      "avg_duration_per_task_ms": 742207.0153846154,
      "avg_output_tokens_per_task": 7710.807692307692,
      "last_run_timestamp": "2026-10-09T01:20:30.611Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "Alibaba",
      "accuracy": 0.413333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06003468,
      "avg_duration_per_task_ms": 392895.94,
      "avg_output_tokens_per_task": 17058.96,
      "last_run_timestamp": "2026-08-01T08:06:29.769Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.4125,
      "accuracy_std_dev": 0.12760346564455058,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.023525742499999995,
      "avg_duration_per_task_ms": 178134.34,
      "avg_output_tokens_per_task": 3278.42,
      "last_run_timestamp": "2026-10-03T16:09:19.691Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "mistralai/mistral-small-2603",
      "model_slug": "mistralai/mistral-small-2603",
      "model_name": "Mistral: Mistral Small 4",
      "provider_name": "Mistral",
      "accuracy": 0.41142871428571426,
      "accuracy_std_dev": 0.022737753675420126,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.014496525714285714,
      "avg_duration_per_task_ms": 78866.04571428572,
      "avg_output_tokens_per_task": 9061.385714285714,
      "last_run_timestamp": "2026-09-17T02:10:06.482Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Nebius",
      "accuracy": 0.408045959770115,
      "accuracy_std_dev": 0.16546342883327816,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.14008307183908045,
      "avg_duration_per_task_ms": 345147.8563218391,
      "avg_output_tokens_per_task": 46487.37356321839,
      "last_run_timestamp": "2026-10-03T16:22:23.573Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "Azure",
      "accuracy": 0.40777783333333334,
      "accuracy_std_dev": 0.01781244190087991,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.025491406666666664,
      "avg_duration_per_task_ms": 73232.91333333333,
      "avg_output_tokens_per_task": 2860.7966666666666,
      "last_run_timestamp": "2026-10-01T19:05:00.133Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.406667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06754938,
      "avg_duration_per_task_ms": 114529.78,
      "avg_output_tokens_per_task": 3789.7,
      "last_run_timestamp": "2026-10-04T03:00:16.402Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.40541675,
      "accuracy_std_dev": 0.021209148643627825,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.06368001,
      "avg_duration_per_task_ms": 101217.8175,
      "avg_output_tokens_per_task": 2945.865,
      "last_run_timestamp": "2026-10-03T16:12:33.500Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "StreamLake",
      "accuracy": 0.4053599698492462,
      "accuracy_std_dev": 0.008101744842933528,
      "run_count": 4,
      "total_questions": 199,
      "avg_cost_per_task_usd": 0.06672347236180905,
      "avg_duration_per_task_ms": 212067.50753768845,
      "avg_output_tokens_per_task": 8080.718592964824,
      "last_run_timestamp": "2026-10-02T15:44:48.231Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.40333325000000003,
      "accuracy_std_dev": 0.034960136029304874,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.05923553749999999,
      "avg_duration_per_task_ms": 320941.4325,
      "avg_output_tokens_per_task": 7834.6225,
      "last_run_timestamp": "2026-10-03T16:05:51.296Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "SiliconFlow",
      "accuracy": 0.402500125,
      "accuracy_std_dev": 0.02559013346212511,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.026921807500000002,
      "avg_duration_per_task_ms": 293558.7475,
      "avg_output_tokens_per_task": 4184.715,
      "last_run_timestamp": "2026-10-03T16:13:56.972Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Venice",
      "accuracy": 0.400833375,
      "accuracy_std_dev": 0.029332849192234545,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0642667,
      "avg_duration_per_task_ms": 297648.6875,
      "avg_output_tokens_per_task": 11410.885,
      "last_run_timestamp": "2026-10-03T16:12:29.905Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.40083312500000007,
      "accuracy_std_dev": 0.02459377320805766,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.07675446250000001,
      "avg_duration_per_task_ms": 122147.305,
      "avg_output_tokens_per_task": 3115.835,
      "last_run_timestamp": "2026-10-03T16:12:07.495Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "auto-routing",
      "accuracy": 0.4,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 1.1315643,
      "avg_duration_per_task_ms": 71996.84,
      "avg_output_tokens_per_task": 3249.82,
      "last_run_timestamp": "2026-08-01T08:06:02.002Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4,
      "accuracy_std_dev": 0.019626323972325204,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.020620846666666668,
      "avg_duration_per_task_ms": 194089.08666666667,
      "avg_output_tokens_per_task": 3830.9066666666668,
      "last_run_timestamp": "2026-08-15T08:02:44.458Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.398333375,
      "accuracy_std_dev": 0.05074458509766707,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.048334209999999996,
      "avg_duration_per_task_ms": 87716.1125,
      "avg_output_tokens_per_task": 3723.8325,
      "last_run_timestamp": "2026-10-03T16:05:51.296Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Ionstream",
      "accuracy": 0.39726041780821914,
      "accuracy_std_dev": 0.26615553243962864,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.011427219178082192,
      "avg_duration_per_task_ms": 86701.55479452055,
      "avg_output_tokens_per_task": 2538.931506849315,
      "last_run_timestamp": "2026-08-08T16:03:49.463Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.396667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0528661,
      "avg_duration_per_task_ms": 135118.22,
      "avg_output_tokens_per_task": 7825.32,
      "last_run_timestamp": "2026-07-25T08:07:19.515Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.39521795358649786,
      "accuracy_std_dev": 0.015940999687598014,
      "run_count": 5,
      "total_questions": 237,
      "avg_cost_per_task_usd": 0.009386248945147679,
      "avg_duration_per_task_ms": 65821.63713080168,
      "avg_output_tokens_per_task": 1124.5696202531647,
      "last_run_timestamp": "2026-10-03T16:15:51.302Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "auto-routing",
      "accuracy": 0.3944446666666666,
      "accuracy_std_dev": 0.015947545189011245,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.07351260000000001,
      "avg_duration_per_task_ms": 297595.3333333333,
      "avg_output_tokens_per_task": 7950.953333333333,
      "last_run_timestamp": "2026-10-01T19:04:42.463Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.39249987500000005,
      "accuracy_std_dev": 0.016811927398706402,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0369896575,
      "avg_duration_per_task_ms": 191395.045,
      "avg_output_tokens_per_task": 3738.2225,
      "last_run_timestamp": "2026-10-03T16:08:50.061Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.3912555560538117,
      "accuracy_std_dev": 0.029116472984579173,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.059476125560538114,
      "avg_duration_per_task_ms": 375172.7399103139,
      "avg_output_tokens_per_task": 3892.973094170404,
      "last_run_timestamp": "2026-10-08T23:21:09.489Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "CoreWeave",
      "accuracy": 0.39,
      "accuracy_std_dev": 0.014782396941407474,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.028387520000000006,
      "avg_duration_per_task_ms": 78569.82333333333,
      "avg_output_tokens_per_task": 4258.516666666666,
      "last_run_timestamp": "2026-08-29T16:13:12.633Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.389166625,
      "accuracy_std_dev": 0.02355558880572453,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.025453972499999998,
      "avg_duration_per_task_ms": 227372.8075,
      "avg_output_tokens_per_task": 4499.2425,
      "last_run_timestamp": "2026-10-03T16:03:16.197Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Ambient",
      "accuracy": 0.386667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.014510460000000001,
      "avg_duration_per_task_ms": 64332.54,
      "avg_output_tokens_per_task": 4106.96,
      "last_run_timestamp": "2026-10-09T15:37:38.473Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.386667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0592903,
      "avg_duration_per_task_ms": 203536.16,
      "avg_output_tokens_per_task": 5170.96,
      "last_run_timestamp": "2026-10-04T03:00:16.402Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Nebius",
      "accuracy": 0.38666671428571425,
      "accuracy_std_dev": 0.008728715610219052,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.014610905714285714,
      "avg_duration_per_task_ms": 90807.17428571428,
      "avg_output_tokens_per_task": 601.44,
      "last_run_timestamp": "2026-08-29T17:06:47.617Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "AkashML",
      "accuracy": 0.37924317696160265,
      "accuracy_std_dev": 0.017662856537252282,
      "run_count": 12,
      "total_questions": 599,
      "avg_cost_per_task_usd": 0.012139260434056762,
      "avg_duration_per_task_ms": 36090.01335559265,
      "avg_output_tokens_per_task": 625.2086811352253,
      "last_run_timestamp": "2026-10-09T15:32:27.707Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.37583325,
      "accuracy_std_dev": 0.03264919388265964,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.051693865,
      "avg_duration_per_task_ms": 72470.795,
      "avg_output_tokens_per_task": 4363.9325,
      "last_run_timestamp": "2026-10-03T16:05:34.166Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "WandB",
      "accuracy": 0.373333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.07157664,
      "avg_duration_per_task_ms": 16604.76,
      "avg_output_tokens_per_task": 863.76,
      "last_run_timestamp": "2026-07-18T16:00:09.280Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "WandB",
      "accuracy": 0.373333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.030392839999999997,
      "avg_duration_per_task_ms": 86392.54,
      "avg_output_tokens_per_task": 5907.74,
      "last_run_timestamp": "2026-07-18T08:03:14.950Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Nebius",
      "accuracy": 0.3714284285714286,
      "accuracy_std_dev": 0.017715366073046326,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03264426571428572,
      "avg_duration_per_task_ms": 156269.61714285714,
      "avg_output_tokens_per_task": 4433.125714285714,
      "last_run_timestamp": "2026-10-03T16:07:02.084Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Chutes",
      "accuracy": 0.37063948837209304,
      "accuracy_std_dev": 0.0721313728137764,
      "run_count": 7,
      "total_questions": 344,
      "avg_cost_per_task_usd": 0.004010601744186046,
      "avg_duration_per_task_ms": 293679.6831395349,
      "avg_output_tokens_per_task": 4202.110465116279,
      "last_run_timestamp": "2026-10-03T16:11:33.007Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "GMICloud",
      "accuracy": 0.36833325,
      "accuracy_std_dev": 0.17853565236861657,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.015193219999999999,
      "avg_duration_per_task_ms": 127682.09,
      "avg_output_tokens_per_task": 2917.505,
      "last_run_timestamp": "2026-10-03T16:12:29.879Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "Azure",
      "accuracy": 0.36666699999999997,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.80181,
      "avg_duration_per_task_ms": 63430.68,
      "avg_output_tokens_per_task": 3064.6,
      "last_run_timestamp": "2026-08-01T08:06:07.633Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "CoreWeave",
      "accuracy": 0.36476185714285714,
      "accuracy_std_dev": 0.029753169004943562,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07516260571428572,
      "avg_duration_per_task_ms": 18915.674285714285,
      "avg_output_tokens_per_task": 875.96,
      "last_run_timestamp": "2026-10-03T16:05:34.168Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Parasail",
      "accuracy": 0.36416687499999995,
      "accuracy_std_dev": 0.07292984785641181,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0056979475000000016,
      "avg_duration_per_task_ms": 133944.15,
      "avg_output_tokens_per_task": 3653.925,
      "last_run_timestamp": "2026-10-03T16:07:05.240Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.36,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.041000339999999996,
      "avg_duration_per_task_ms": 322345.3,
      "avg_output_tokens_per_task": 4212.74,
      "last_run_timestamp": "2026-10-04T03:00:10.237Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.3567604417670683,
      "accuracy_std_dev": 0.024567950668299542,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.0795283875502008,
      "avg_duration_per_task_ms": 211758.7168674699,
      "avg_output_tokens_per_task": 7341.082329317269,
      "last_run_timestamp": "2026-10-03T08:12:20.019Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.3566663333333333,
      "accuracy_std_dev": 0.016996731711975962,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.043029019999999994,
      "avg_duration_per_task_ms": 111012.42666666667,
      "avg_output_tokens_per_task": 4511.693333333334,
      "last_run_timestamp": "2026-08-01T17:38:36.457Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.35592184615384614,
      "accuracy_std_dev": 0.02760835600602419,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.07156286996336997,
      "avg_duration_per_task_ms": 377710.5695970696,
      "avg_output_tokens_per_task": 7880.789377289378,
      "last_run_timestamp": "2026-10-08T23:34:52.306Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "Novita",
      "accuracy": 0.34058,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 46,
      "avg_cost_per_task_usd": 0.11353834782608695,
      "avg_duration_per_task_ms": 397172.5,
      "avg_output_tokens_per_task": 14067.434782608696,
      "last_run_timestamp": "2026-08-01T08:05:57.857Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "CoreWeave",
      "accuracy": 0.3367119324324324,
      "accuracy_std_dev": 0.23078482961913535,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.014007192567567568,
      "avg_duration_per_task_ms": 1521898.9797297297,
      "avg_output_tokens_per_task": 94037.88851351352,
      "last_run_timestamp": "2026-10-03T16:12:24.696Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.33666633333333335,
      "accuracy_std_dev": 0.004714045207910321,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.029078933333333334,
      "avg_duration_per_task_ms": 79568.04666666666,
      "avg_output_tokens_per_task": 3568.733333333333,
      "last_run_timestamp": "2026-08-01T17:38:37.779Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.3358587171717172,
      "accuracy_std_dev": 0.05280919547472296,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02850434343434343,
      "avg_duration_per_task_ms": 105990.88888888889,
      "avg_output_tokens_per_task": 1237.1767676767677,
      "last_run_timestamp": "2026-10-03T16:14:36.877Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.333333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.051354699999999996,
      "avg_duration_per_task_ms": 358714.54,
      "avg_output_tokens_per_task": 69285.14,
      "last_run_timestamp": "2026-10-04T03:00:10.240Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.3296970909090909,
      "accuracy_std_dev": 0.023845300316279384,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.08440126545454546,
      "avg_duration_per_task_ms": 151545.54545454544,
      "avg_output_tokens_per_task": 8175.638181818182,
      "last_run_timestamp": "2026-10-03T08:23:09.856Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.326667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02872976,
      "avg_duration_per_task_ms": 133695.86,
      "avg_output_tokens_per_task": 3408.56,
      "last_run_timestamp": "2026-10-04T03:00:16.402Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Venice",
      "accuracy": 0.3206666,
      "accuracy_std_dev": 0.3296116389532384,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.019227021999999996,
      "avg_duration_per_task_ms": 58892.988,
      "avg_output_tokens_per_task": 10442.698,
      "last_run_timestamp": "2026-10-03T16:12:07.499Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.313333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.15186704,
      "avg_duration_per_task_ms": 32635.52,
      "avg_output_tokens_per_task": 1294.54,
      "last_run_timestamp": "2026-10-04T03:00:16.406Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.30896692982456136,
      "accuracy_std_dev": 0.03129504337294144,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.01647059649122807,
      "avg_duration_per_task_ms": 365014.59064327483,
      "avg_output_tokens_per_task": 4454.833333333333,
      "last_run_timestamp": "2026-08-29T16:43:42.449Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Together",
      "accuracy": 0.30416662499999997,
      "accuracy_std_dev": 0.03799677343846942,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.16497433250000001,
      "avg_duration_per_task_ms": 125643.485,
      "avg_output_tokens_per_task": 1850.835,
      "last_run_timestamp": "2026-10-03T16:16:35.584Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-mini",
      "model_slug": "openai/gpt-4o-mini",
      "model_name": "OpenAI: GPT-4o-mini",
      "provider_name": "auto-routing",
      "accuracy": 0.2844443333333333,
      "accuracy_std_dev": 0.031894788459698065,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02081186,
      "avg_duration_per_task_ms": 62184.12666666666,
      "avg_output_tokens_per_task": 3804.7533333333336,
      "last_run_timestamp": "2026-10-01T19:04:58.075Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-mini",
      "model_slug": "openai/gpt-4o-mini",
      "model_name": "OpenAI: GPT-4o-mini",
      "provider_name": "OpenAI",
      "accuracy": 0.28000025,
      "accuracy_std_dev": 0.027080189561513406,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.02535222,
      "avg_duration_per_task_ms": 67904.845,
      "avg_output_tokens_per_task": 4023.05,
      "last_run_timestamp": "2026-10-02T23:13:47.090Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "SiliconFlow",
      "accuracy": 0.2797619910714286,
      "accuracy_std_dev": 0.1899431271985133,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.15921619642857143,
      "avg_duration_per_task_ms": 942346.96875,
      "avg_output_tokens_per_task": 6098.834821428572,
      "last_run_timestamp": "2026-10-03T08:09:23.857Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen-2.5-72b-instruct",
      "model_slug": "qwen/qwen-2.5-72b-instruct",
      "model_name": "Qwen2.5 72B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.2622223333333333,
      "accuracy_std_dev": 0.01911642045177101,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08337879999999999,
      "avg_duration_per_task_ms": 576772.3666666667,
      "avg_output_tokens_per_task": 5220.566666666667,
      "last_run_timestamp": "2026-10-01T16:00:45.063Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "auto-routing",
      "accuracy": 0.260417,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.027361125,
      "avg_duration_per_task_ms": 394399.5416666667,
      "avg_output_tokens_per_task": 4477.166666666667,
      "last_run_timestamp": "2026-07-18T08:04:09.749Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "Mistral",
      "accuracy": 0.25,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02790544,
      "avg_duration_per_task_ms": 464213.12,
      "avg_output_tokens_per_task": 5425.66,
      "last_run_timestamp": "2026-07-18T08:04:25.389Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen-2.5-72b-instruct",
      "model_slug": "qwen/qwen-2.5-72b-instruct",
      "model_name": "Qwen2.5 72B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.25,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0713485,
      "avg_duration_per_task_ms": 660516.14,
      "avg_output_tokens_per_task": 6662.1,
      "last_run_timestamp": "2026-09-01T12:59:14.290Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen-2.5-72b-instruct",
      "model_slug": "qwen/qwen-2.5-72b-instruct",
      "model_name": "Qwen2.5 72B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.24200919178082192,
      "accuracy_std_dev": 0.027381797668443095,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.07592458219178082,
      "avg_duration_per_task_ms": 576526.0342465753,
      "avg_output_tokens_per_task": 5262.068493150685,
      "last_run_timestamp": "2026-10-01T16:00:49.856Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Novita",
      "accuracy": 0.2325,
      "accuracy_std_dev": 0.06851085072271398,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.013581069999999997,
      "avg_duration_per_task_ms": 21748.2225,
      "avg_output_tokens_per_task": 1125.92,
      "last_run_timestamp": "2026-10-03T16:13:42.255Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.19666649999999997,
      "accuracy_std_dev": 0.023804971471312466,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.00425299,
      "avg_duration_per_task_ms": 15962.445,
      "avg_output_tokens_per_task": 1579.045,
      "last_run_timestamp": "2026-10-03T16:07:44.178Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "AtlasCloud",
      "accuracy": 0.18791944295302013,
      "accuracy_std_dev": 0.10708233389741122,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.20449519798657714,
      "avg_duration_per_task_ms": 608397.9966442953,
      "avg_output_tokens_per_task": 10862.70134228188,
      "last_run_timestamp": "2026-08-29T16:22:37.844Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.18766967073170732,
      "accuracy_std_dev": 0.018396157170452745,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.011055426829268291,
      "avg_duration_per_task_ms": 138140.34552845528,
      "avg_output_tokens_per_task": 5390.59756097561,
      "last_run_timestamp": "2026-10-01T19:04:52.650Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.186667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01693508,
      "avg_duration_per_task_ms": 26941.4,
      "avg_output_tokens_per_task": 1041.58,
      "last_run_timestamp": "2026-10-04T03:00:16.405Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "AtlasCloud",
      "accuracy": 0.17999985714285713,
      "accuracy_std_dev": 0.12173356543044388,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.013454491428571429,
      "avg_duration_per_task_ms": 61907.58571428571,
      "avg_output_tokens_per_task": 1212.5742857142857,
      "last_run_timestamp": "2026-08-29T16:17:08.780Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen-2.5-7b-instruct",
      "model_slug": "qwen/qwen-2.5-7b-instruct",
      "model_name": "Qwen: Qwen2.5 7B Instruct",
      "provider_name": "Phala",
      "accuracy": 0.166667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.01877496,
      "avg_duration_per_task_ms": 63137.36,
      "avg_output_tokens_per_task": 2797.32,
      "last_run_timestamp": "2026-08-01T08:06:16.826Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen-2.5-7b-instruct",
      "model_slug": "qwen/qwen-2.5-7b-instruct",
      "model_name": "Qwen: Qwen2.5 7B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.166667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.016978299999999998,
      "avg_duration_per_task_ms": 57790.52,
      "avg_output_tokens_per_task": 2492.08,
      "last_run_timestamp": "2026-08-01T08:06:07.057Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-nano-2025-04-14",
      "model_slug": "openai/gpt-4.1-nano",
      "model_name": "OpenAI: GPT-4.1 Nano",
      "provider_name": "Azure",
      "accuracy": 0.13333333333333333,
      "accuracy_std_dev": 0.021430473285072877,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.007961373333333334,
      "avg_duration_per_task_ms": 80728.30333333333,
      "avg_output_tokens_per_task": 2469.17,
      "last_run_timestamp": "2026-10-01T19:04:58.093Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-nano-2025-04-14",
      "model_slug": "openai/gpt-4.1-nano",
      "model_name": "OpenAI: GPT-4.1 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.11833325,
      "accuracy_std_dev": 0.018484159885899604,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.007158584999999999,
      "avg_duration_per_task_ms": 55526.67,
      "avg_output_tokens_per_task": 2353.445,
      "last_run_timestamp": "2026-10-02T23:07:59.112Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4.1-nano-2025-04-14",
      "model_slug": "openai/gpt-4.1-nano",
      "model_name": "OpenAI: GPT-4.1 Nano",
      "provider_name": "auto-routing",
      "accuracy": 0.108889,
      "accuracy_std_dev": 0.008314764498569193,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.006775626666666668,
      "avg_duration_per_task_ms": 60620.74,
      "avg_output_tokens_per_task": 2327.0066666666667,
      "last_run_timestamp": "2026-10-01T19:04:36.194Z"
    }
  ]
}
