{
  "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": "2026-09-30",
  "retrieved_at": "2026-09-30T01:45:37.914Z",
  "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": 2004,
  "rows": [
    {
      "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": 0.952862,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.2003321212121212,
      "avg_duration_per_task_ms": 124514.04545454546,
      "avg_output_tokens_per_task": 16571.29797979798,
      "last_run_timestamp": "2026-09-01T17:50:10.294Z"
    },
    {
      "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": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "OpenAI",
      "accuracy": 0.9461279999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.116425398989899,
      "avg_duration_per_task_ms": 44572.09595959596,
      "avg_output_tokens_per_task": 2176.570707070707,
      "last_run_timestamp": "2026-09-06T03:17:58.152Z"
    },
    {
      "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": 0.945286,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.2012271818181818,
      "avg_duration_per_task_ms": 134249.20707070708,
      "avg_output_tokens_per_task": 16644.886363636364,
      "last_run_timestamp": "2026-09-01T17:53:46.786Z"
    },
    {
      "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": 0.945286,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02981064646464646,
      "avg_duration_per_task_ms": 28082.027777777777,
      "avg_output_tokens_per_task": 7803.272727272727,
      "last_run_timestamp": "2026-09-01T13:28:27.192Z"
    },
    {
      "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": 0.944444,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.20185933333333334,
      "avg_duration_per_task_ms": 126427.48484848485,
      "avg_output_tokens_per_task": 16698.565656565657,
      "last_run_timestamp": "2026-09-01T17:51:04.719Z"
    },
    {
      "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": 0.944444,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.11967219191919193,
      "avg_duration_per_task_ms": 38197.17171717172,
      "avg_output_tokens_per_task": 2017.3383838383838,
      "last_run_timestamp": "2026-09-26T10:54:50.352Z"
    },
    {
      "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": 0.9436415317919074,
      "accuracy_std_dev": 0.0020135000000000014,
      "run_count": 2,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.09933485549132949,
      "avg_duration_per_task_ms": 198116.95664739885,
      "avg_output_tokens_per_task": 1767.9046242774566,
      "last_run_timestamp": "2026-09-06T03:05:46.155Z"
    },
    {
      "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": 0.942761,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.030509156565656566,
      "avg_duration_per_task_ms": 68136.79292929293,
      "avg_output_tokens_per_task": 7993.409090909091,
      "last_run_timestamp": "2026-09-01T13:25:46.082Z"
    },
    {
      "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": "anthropic/claude-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Google",
      "accuracy": 0.9382716666666666,
      "accuracy_std_dev": 0.006918527315998818,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0262555,
      "avg_duration_per_task_ms": 19027.23400673401,
      "avg_output_tokens_per_task": 2243.60101010101,
      "last_run_timestamp": "2026-09-29T12:03:13.309Z"
    },
    {
      "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": 0.9377104999999999,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.30815713636363634,
      "avg_duration_per_task_ms": 160641.4873737374,
      "avg_output_tokens_per_task": 10142.272727272728,
      "last_run_timestamp": "2026-09-01T13:20:32.033Z"
    },
    {
      "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": 0.93771,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.4446258484848485,
      "avg_duration_per_task_ms": 116119.66666666667,
      "avg_output_tokens_per_task": 11825.727272727272,
      "last_run_timestamp": "2026-08-01T08:45:01.642Z"
    },
    {
      "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": 0.9351849999999999,
      "accuracy_std_dev": 0.006162785766193721,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1365803898989899,
      "avg_duration_per_task_ms": 176670.50606060607,
      "avg_output_tokens_per_task": 8896.30606060606,
      "last_run_timestamp": "2026-09-26T09:17:14.173Z"
    },
    {
      "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": 0.9335082598425197,
      "accuracy_std_dev": 0.011397999999999964,
      "run_count": 2,
      "total_questions": 381,
      "avg_cost_per_task_usd": 0.0213603280839895,
      "avg_duration_per_task_ms": 53235.17585301837,
      "avg_output_tokens_per_task": 7688.425196850394,
      "last_run_timestamp": "2026-09-01T13:29:18.783Z"
    },
    {
      "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": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.93266,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.12205839393939394,
      "avg_duration_per_task_ms": 394338.13636363635,
      "avg_output_tokens_per_task": 20077.80808080808,
      "last_run_timestamp": "2026-09-01T13:22:57.685Z"
    },
    {
      "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": 0.9324968387096775,
      "accuracy_std_dev": 0.005080239036370865,
      "run_count": 3,
      "total_questions": 558,
      "avg_cost_per_task_usd": 0.0715788476702509,
      "avg_duration_per_task_ms": 64320.11827956989,
      "avg_output_tokens_per_task": 11534.012544802867,
      "last_run_timestamp": "2026-09-01T17:51:06.380Z"
    },
    {
      "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": 0.930976,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.025067358585858585,
      "avg_duration_per_task_ms": 20515.378787878788,
      "avg_output_tokens_per_task": 2267.0252525252527,
      "last_run_timestamp": "2026-09-29T12:03:17.610Z"
    },
    {
      "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": 0.930976,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.32996825757575754,
      "avg_duration_per_task_ms": 166479.3383838384,
      "avg_output_tokens_per_task": 10868.631313131313,
      "last_run_timestamp": "2026-09-01T13:22:14.451Z"
    },
    {
      "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": 0.9299660000000001,
      "accuracy_std_dev": 0.009009433123121587,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.32134650707070705,
      "avg_duration_per_task_ms": 152235.298989899,
      "avg_output_tokens_per_task": 9994.927272727273,
      "last_run_timestamp": "2026-09-01T13:24:49.600Z"
    },
    {
      "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": 0.929885,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 145,
      "avg_cost_per_task_usd": 0.09349187586206896,
      "avg_duration_per_task_ms": 204934.02068965518,
      "avg_output_tokens_per_task": 6095.813793103448,
      "last_run_timestamp": "2026-09-29T21:23:16.334Z"
    },
    {
      "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": 0.9294545733558178,
      "accuracy_std_dev": 0.004334827511626685,
      "run_count": 3,
      "total_questions": 593,
      "avg_cost_per_task_usd": 0.02681522428330523,
      "avg_duration_per_task_ms": 26517.369308600337,
      "avg_output_tokens_per_task": 2356.256323777403,
      "last_run_timestamp": "2026-09-23T15:41:43.462Z"
    },
    {
      "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": 0.9292930000000001,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14407252272727272,
      "avg_duration_per_task_ms": 58692.59848484849,
      "avg_output_tokens_per_task": 15884.012626262625,
      "last_run_timestamp": "2026-09-01T17:55:40.117Z"
    },
    {
      "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": 0.9292929999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02512639898989899,
      "avg_duration_per_task_ms": 21072.075757575756,
      "avg_output_tokens_per_task": 2272.929292929293,
      "last_run_timestamp": "2026-09-29T12:03:27.573Z"
    },
    {
      "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": 0.92887175,
      "accuracy_std_dev": 0.005239941859171739,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.06947331565656567,
      "avg_duration_per_task_ms": 67244.06060606061,
      "avg_output_tokens_per_task": 11635.200757575758,
      "last_run_timestamp": "2026-09-01T17:56:37.544Z"
    },
    {
      "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": 0.928451,
      "accuracy_std_dev": 0.004209000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.06683859343434344,
      "avg_duration_per_task_ms": 59600.07575757576,
      "avg_output_tokens_per_task": 11674.472222222223,
      "last_run_timestamp": "2026-09-01T17:57:17.668Z"
    },
    {
      "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": 0.9283648630849219,
      "accuracy_std_dev": 0.017830944261311557,
      "run_count": 4,
      "total_questions": 577,
      "avg_cost_per_task_usd": 0.09151112998266897,
      "avg_duration_per_task_ms": 252153.69497400348,
      "avg_output_tokens_per_task": 7044.975736568457,
      "last_run_timestamp": "2026-09-26T11:09:52.628Z"
    },
    {
      "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": 0.9279571938392187,
      "accuracy_std_dev": 0.00835201542708194,
      "run_count": 22,
      "total_questions": 3993,
      "avg_cost_per_task_usd": 0.15530522213874282,
      "avg_duration_per_task_ms": 142386.56699223642,
      "avg_output_tokens_per_task": 8728.15927873779,
      "last_run_timestamp": "2026-09-26T11:50:40.297Z"
    },
    {
      "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": 0.927609,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.026096929292929293,
      "avg_duration_per_task_ms": 19999.954545454544,
      "avg_output_tokens_per_task": 2369.09595959596,
      "last_run_timestamp": "2026-09-29T12:06:06.771Z"
    },
    {
      "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": 0.927609,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14183102777777779,
      "avg_duration_per_task_ms": 79740.97727272728,
      "avg_output_tokens_per_task": 15646.176767676769,
      "last_run_timestamp": "2026-09-01T18:06:42.346Z"
    },
    {
      "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": 0.92613625,
      "accuracy_std_dev": 0.007949293313716653,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.09945460101010102,
      "avg_duration_per_task_ms": 230595.33712121213,
      "avg_output_tokens_per_task": 9430.997474747475,
      "last_run_timestamp": "2026-09-26T08:53:13.109Z"
    },
    {
      "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": 0.9259257500000001,
      "accuracy_std_dev": 0.004610545920766882,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.1489979318181818,
      "avg_duration_per_task_ms": 84816.23358585859,
      "avg_output_tokens_per_task": 15665.375,
      "last_run_timestamp": "2026-09-01T17:51:30.647Z"
    },
    {
      "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": 0.925347078125,
      "accuracy_std_dev": 0.005810499999999996,
      "run_count": 2,
      "total_questions": 384,
      "avg_cost_per_task_usd": 0.3411057265625,
      "avg_duration_per_task_ms": 143482.140625,
      "avg_output_tokens_per_task": 10207.984375,
      "last_run_timestamp": "2026-09-01T13:21:58.114Z"
    },
    {
      "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": 0.9245122977099236,
      "accuracy_std_dev": 0.009051464284164208,
      "run_count": 4,
      "total_questions": 786,
      "avg_cost_per_task_usd": 0.12044641603053435,
      "avg_duration_per_task_ms": 355037.55470737914,
      "avg_output_tokens_per_task": 19323.84987277354,
      "last_run_timestamp": "2026-09-29T03:08:09.349Z"
    },
    {
      "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": 0.9242420000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.026419161616161615,
      "avg_duration_per_task_ms": 21367.934343434343,
      "avg_output_tokens_per_task": 2243.222222222222,
      "last_run_timestamp": "2026-09-23T15:27:41.867Z"
    },
    {
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.9239580000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 160,
      "avg_cost_per_task_usd": 0.0028846625000000002,
      "avg_duration_per_task_ms": 871679.5875,
      "avg_output_tokens_per_task": 9877.29375,
      "last_run_timestamp": "2026-09-27T19:46:05.023Z"
    },
    {
      "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": 0.922559,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.4976751313131313,
      "avg_duration_per_task_ms": 91843.67676767676,
      "avg_output_tokens_per_task": 12034.212121212122,
      "last_run_timestamp": "2026-08-01T08:41:32.647Z"
    },
    {
      "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": 0.9224121127739175,
      "accuracy_std_dev": 0.007223723392406446,
      "run_count": 10,
      "total_questions": 1871,
      "avg_cost_per_task_usd": 0.15054908498129344,
      "avg_duration_per_task_ms": 122412.57936932122,
      "avg_output_tokens_per_task": 9889.19080705505,
      "last_run_timestamp": "2026-09-26T08:55:05.759Z"
    },
    {
      "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": 0.9215687647058822,
      "accuracy_std_dev": 0.0038094199325123296,
      "run_count": 3,
      "total_questions": 561,
      "avg_cost_per_task_usd": 0.10569379500891266,
      "avg_duration_per_task_ms": 54609.55971479501,
      "avg_output_tokens_per_task": 3484.6114081996434,
      "last_run_timestamp": "2026-09-01T13:22:25.850Z"
    },
    {
      "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": 0.9212118,
      "accuracy_std_dev": 0.008675973568424486,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1201016404040404,
      "avg_duration_per_task_ms": 61346.69292929293,
      "avg_output_tokens_per_task": 3579.7959595959596,
      "last_run_timestamp": "2026-09-01T13:28:53.908Z"
    },
    {
      "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": 0.920875,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.4329253232323232,
      "avg_duration_per_task_ms": 113725.84848484848,
      "avg_output_tokens_per_task": 11493.873737373737,
      "last_run_timestamp": "2026-08-01T08:53:39.678Z"
    },
    {
      "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": "Decart",
      "accuracy": 0.920034,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.12737118686868687,
      "avg_duration_per_task_ms": 194690.12121212122,
      "avg_output_tokens_per_task": 8293.338383838383,
      "last_run_timestamp": "2026-09-29T12:27:43.930Z"
    },
    {
      "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": 0.9200335,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07599968055555556,
      "avg_duration_per_task_ms": 289463.85101010103,
      "avg_output_tokens_per_task": 24687.122474747473,
      "last_run_timestamp": "2026-09-26T09:34:25.072Z"
    },
    {
      "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": 0.919865,
      "accuracy_std_dev": 0.007939302840930074,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.16831854444444444,
      "avg_duration_per_task_ms": 345204.0070707071,
      "avg_output_tokens_per_task": 9582.316161616161,
      "last_run_timestamp": "2026-09-26T09:45:25.093Z"
    },
    {
      "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": 0.9192883359550563,
      "accuracy_std_dev": 0.004632004101898027,
      "run_count": 9,
      "total_questions": 1780,
      "avg_cost_per_task_usd": 0.19511707022471914,
      "avg_duration_per_task_ms": 338842.3713483146,
      "avg_output_tokens_per_task": 44188.06966292135,
      "last_run_timestamp": "2026-09-26T08:56:52.320Z"
    },
    {
      "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": 0.9191915,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07126612626262625,
      "avg_duration_per_task_ms": 64525.257575757576,
      "avg_output_tokens_per_task": 3520.8156565656564,
      "last_run_timestamp": "2026-09-01T13:21:08.334Z"
    },
    {
      "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": 0.9188843954372623,
      "accuracy_std_dev": 0.012617923846160294,
      "run_count": 8,
      "total_questions": 1578,
      "avg_cost_per_task_usd": 0.13335884410646387,
      "avg_duration_per_task_ms": 335309.8307984791,
      "avg_output_tokens_per_task": 10503.390367553866,
      "last_run_timestamp": "2026-09-26T08:52:39.134Z"
    },
    {
      "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": 0.9186867999999999,
      "accuracy_std_dev": 0.0068919500984844775,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.13311589696969697,
      "avg_duration_per_task_ms": 546514.5777777778,
      "avg_output_tokens_per_task": 36924.677777777775,
      "last_run_timestamp": "2026-09-26T09:34:49.008Z"
    },
    {
      "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": 0.9179851756329115,
      "accuracy_std_dev": 0.013506707135345767,
      "run_count": 10,
      "total_questions": 1896,
      "avg_cost_per_task_usd": 0.12114161814345992,
      "avg_duration_per_task_ms": 120275.13765822785,
      "avg_output_tokens_per_task": 8135.254219409283,
      "last_run_timestamp": "2026-09-26T09:59:38.434Z"
    },
    {
      "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": 0.9175826670976084,
      "accuracy_std_dev": 0.014195212289272039,
      "run_count": 8,
      "total_questions": 1547,
      "avg_cost_per_task_usd": 0.008065983839689723,
      "avg_duration_per_task_ms": 489574.17323852616,
      "avg_output_tokens_per_task": 20043.881060116353,
      "last_run_timestamp": "2026-09-26T09:23:12.184Z"
    },
    {
      "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": "OpenAI",
      "accuracy": 0.917508,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.025625949494949497,
      "avg_duration_per_task_ms": 37548.47474747475,
      "avg_output_tokens_per_task": 2404.0757575757575,
      "last_run_timestamp": "2026-09-23T15:32:55.695Z"
    },
    {
      "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": 0.9166664999999999,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.026023550505050508,
      "avg_duration_per_task_ms": 20014.84595959596,
      "avg_output_tokens_per_task": 2239.1363636363635,
      "last_run_timestamp": "2026-09-29T12:03:59.394Z"
    },
    {
      "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": 0.9164982,
      "accuracy_std_dev": 0.005280767383629021,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.09286130606060607,
      "avg_duration_per_task_ms": 636505.3202020202,
      "avg_output_tokens_per_task": 40298.016161616164,
      "last_run_timestamp": "2026-09-26T09:33:50.566Z"
    },
    {
      "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": 0.914983,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07349445707070708,
      "avg_duration_per_task_ms": 64386.71464646464,
      "avg_output_tokens_per_task": 3569.0353535353534,
      "last_run_timestamp": "2026-09-01T13:22:37.815Z"
    },
    {
      "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": 0.9147023333333334,
      "accuracy_std_dev": 0.011850896290529652,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.1591313367003367,
      "avg_duration_per_task_ms": 538081.3703703703,
      "avg_output_tokens_per_task": 39867.37205387205,
      "last_run_timestamp": "2026-09-26T09:14:59.717Z"
    },
    {
      "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": 0.9139881818181818,
      "accuracy_std_dev": 0.01312802668698594,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.032382866391184574,
      "avg_duration_per_task_ms": 222132.83562901744,
      "avg_output_tokens_per_task": 26800.957759412304,
      "last_run_timestamp": "2026-09-26T16:05:15.213Z"
    },
    {
      "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": 0.9139752900807381,
      "accuracy_std_dev": 0.006586842457168369,
      "run_count": 9,
      "total_questions": 1734,
      "avg_cost_per_task_usd": 0.14062165570934257,
      "avg_duration_per_task_ms": 294457.8246828143,
      "avg_output_tokens_per_task": 9226.012687427912,
      "last_run_timestamp": "2026-09-26T09:18:47.498Z"
    },
    {
      "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": 0.9127381666666666,
      "accuracy_std_dev": 0.0041711655212369175,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.27201272474747473,
      "avg_duration_per_task_ms": 357401.7491582492,
      "avg_output_tokens_per_task": 38560.36363636364,
      "last_run_timestamp": "2026-09-26T09:19:10.819Z"
    },
    {
      "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": 0.9124577500000001,
      "accuracy_std_dev": 0.006628136630117118,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0378646856060606,
      "avg_duration_per_task_ms": 130990.08964646465,
      "avg_output_tokens_per_task": 28457.30808080808,
      "last_run_timestamp": "2026-09-26T08:54:48.831Z"
    },
    {
      "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": 0.9118049954065226,
      "accuracy_std_dev": 0.030664109703108882,
      "run_count": 11,
      "total_questions": 2177,
      "avg_cost_per_task_usd": 0.018931710151584753,
      "avg_duration_per_task_ms": 61807.37344970142,
      "avg_output_tokens_per_task": 3220.4230592558565,
      "last_run_timestamp": "2026-09-28T14:41:42.105Z"
    },
    {
      "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": 0.9107743333333334,
      "accuracy_std_dev": 0.004956316261992276,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03197664478114478,
      "avg_duration_per_task_ms": 194795.97138047137,
      "avg_output_tokens_per_task": 26470.969696969696,
      "last_run_timestamp": "2026-09-26T08:52:40.621Z"
    },
    {
      "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": 0.9104936666666668,
      "accuracy_std_dev": 0.05077924822427192,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.10505623232323233,
      "avg_duration_per_task_ms": 246993.5,
      "avg_output_tokens_per_task": 9288.036195286195,
      "last_run_timestamp": "2026-09-26T09:35:57.204Z"
    },
    {
      "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": 0.9102129999999999,
      "accuracy_std_dev": 0.002861334420627323,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0912948888888889,
      "avg_duration_per_task_ms": 155548.53703703705,
      "avg_output_tokens_per_task": 27869.387205387204,
      "last_run_timestamp": "2026-09-01T13:31:50.869Z"
    },
    {
      "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": 0.9096159477085782,
      "accuracy_std_dev": 0.03256740651027582,
      "run_count": 18,
      "total_questions": 3404,
      "avg_cost_per_task_usd": 0.1496695376028202,
      "avg_duration_per_task_ms": 473661.7364864865,
      "avg_output_tokens_per_task": 8839.687720329024,
      "last_run_timestamp": "2026-09-26T11:50:25.798Z"
    },
    {
      "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": 0.9091866447534767,
      "accuracy_std_dev": 0.0045401682512876225,
      "run_count": 8,
      "total_questions": 1582,
      "avg_cost_per_task_usd": 0.029416860935524654,
      "avg_duration_per_task_ms": 453238.0587863464,
      "avg_output_tokens_per_task": 30418.960809102402,
      "last_run_timestamp": "2026-09-26T16:21:44.172Z"
    },
    {
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.9090566691616767,
      "accuracy_std_dev": 0.011036068624917007,
      "run_count": 7,
      "total_questions": 1336,
      "avg_cost_per_task_usd": 0.012388138473053893,
      "avg_duration_per_task_ms": 304706.8982035928,
      "avg_output_tokens_per_task": 16929.82260479042,
      "last_run_timestamp": "2026-09-26T08:52:35.552Z"
    },
    {
      "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": 0.9087729427112877,
      "accuracy_std_dev": 0.01605553275354013,
      "run_count": 9,
      "total_questions": 1763,
      "avg_cost_per_task_usd": 0.188728044242768,
      "avg_duration_per_task_ms": 517068.7197958026,
      "avg_output_tokens_per_task": 12546.453204764606,
      "last_run_timestamp": "2026-09-26T08:17:01.669Z"
    },
    {
      "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": 0.9086699999999999,
      "accuracy_std_dev": 0.0024899934738870384,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.035269609848484845,
      "avg_duration_per_task_ms": 184713.31186868687,
      "avg_output_tokens_per_task": 29215.651515151516,
      "last_run_timestamp": "2026-09-26T09:27:25.556Z"
    },
    {
      "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": 0.908249,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.1657261111111111,
      "avg_duration_per_task_ms": 467040.702020202,
      "avg_output_tokens_per_task": 43784.85353535353,
      "last_run_timestamp": "2026-09-28T06:30:03.588Z"
    },
    {
      "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": 0.9074072500000001,
      "accuracy_std_dev": 0.004610545920766882,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.030798316919191916,
      "avg_duration_per_task_ms": 142600.18308080808,
      "avg_output_tokens_per_task": 28269.325757575756,
      "last_run_timestamp": "2026-09-26T09:35:41.692Z"
    },
    {
      "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": 0.9073550816901409,
      "accuracy_std_dev": 0.02204460352471743,
      "run_count": 11,
      "total_questions": 2130,
      "avg_cost_per_task_usd": 0.03663173568075117,
      "avg_duration_per_task_ms": 263398.2685446009,
      "avg_output_tokens_per_task": 30725.946948356806,
      "last_run_timestamp": "2026-09-26T16:06:12.156Z"
    },
    {
      "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": 0.9069262857142857,
      "accuracy_std_dev": 0.01966827796016641,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.13709101803751803,
      "avg_duration_per_task_ms": 571292.8715728716,
      "avg_output_tokens_per_task": 38969.107503607505,
      "last_run_timestamp": "2026-09-26T08:52:35.549Z"
    },
    {
      "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": 0.905724,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.054357919191919186,
      "avg_duration_per_task_ms": 31472.4797979798,
      "avg_output_tokens_per_task": 2509.6414141414143,
      "last_run_timestamp": "2026-09-26T08:54:32.056Z"
    },
    {
      "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": 0.9057234999999999,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03042029292929293,
      "avg_duration_per_task_ms": 203300.4116161616,
      "avg_output_tokens_per_task": 29610.87373737374,
      "last_run_timestamp": "2026-09-26T10:10:49.311Z"
    },
    {
      "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": 0.9054903294117648,
      "accuracy_std_dev": 0.002950830881105991,
      "run_count": 3,
      "total_questions": 425,
      "avg_cost_per_task_usd": 0.0986494188235294,
      "avg_duration_per_task_ms": 590399.5082352941,
      "avg_output_tokens_per_task": 26203.308235294116,
      "last_run_timestamp": "2026-09-05T08:31:24.281Z"
    },
    {
      "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": 0.9053029638429752,
      "accuracy_std_dev": 0.01776711826408548,
      "run_count": 10,
      "total_questions": 1936,
      "avg_cost_per_task_usd": 0.03572506508264463,
      "avg_duration_per_task_ms": 311302.8930785124,
      "avg_output_tokens_per_task": 29769.41993801653,
      "last_run_timestamp": "2026-09-26T16:32:47.455Z"
    },
    {
      "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": 0.9051626666666666,
      "accuracy_std_dev": 0.00649574344389378,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.028482462962962966,
      "avg_duration_per_task_ms": 244979.07912457912,
      "avg_output_tokens_per_task": 29233.41919191919,
      "last_run_timestamp": "2026-09-26T08:52:19.672Z"
    },
    {
      "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": 0.9041170761421321,
      "accuracy_std_dev": 0.007233580779177685,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.013811796954314722,
      "avg_duration_per_task_ms": 592278.3874788494,
      "avg_output_tokens_per_task": 28886.412859560067,
      "last_run_timestamp": "2026-09-26T09:15:49.652Z"
    },
    {
      "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": 0.9040400000000001,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02929719696969697,
      "avg_duration_per_task_ms": 146549.648989899,
      "avg_output_tokens_per_task": 27342.593434343435,
      "last_run_timestamp": "2026-09-01T13:36:23.129Z"
    },
    {
      "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": 0.9037085815423515,
      "accuracy_std_dev": 0.00877916433950291,
      "run_count": 4,
      "total_questions": 791,
      "avg_cost_per_task_usd": 0.033949044247787614,
      "avg_duration_per_task_ms": 139050.61061946902,
      "avg_output_tokens_per_task": 28840.025284450065,
      "last_run_timestamp": "2026-09-26T08:51:27.499Z"
    },
    {
      "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": 0.903601419310345,
      "accuracy_std_dev": 0.02045915087022377,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.03446190114942529,
      "avg_duration_per_task_ms": 316983.98942528735,
      "avg_output_tokens_per_task": 28739.95632183908,
      "last_run_timestamp": "2026-09-26T16:08:07.059Z"
    },
    {
      "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": 0.9031985,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14676221212121213,
      "avg_duration_per_task_ms": 106397.32070707071,
      "avg_output_tokens_per_task": 9655.194444444445,
      "last_run_timestamp": "2026-09-01T17:54:59.196Z"
    },
    {
      "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": 0.9031985,
      "accuracy_std_dev": 0.004208500000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.043649184343434344,
      "avg_duration_per_task_ms": 141329.08080808082,
      "avg_output_tokens_per_task": 28078.323232323233,
      "last_run_timestamp": "2026-09-01T13:21:17.284Z"
    },
    {
      "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": 0.902848790174002,
      "accuracy_std_dev": 0.012183292931305548,
      "run_count": 10,
      "total_questions": 1954,
      "avg_cost_per_task_usd": 0.030339729273285567,
      "avg_duration_per_task_ms": 232833.84339815762,
      "avg_output_tokens_per_task": 25157.37871033777,
      "last_run_timestamp": "2026-09-26T16:18:02.009Z"
    },
    {
      "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": 0.90256725,
      "accuracy_std_dev": 0.031891435832171314,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.042016968434343424,
      "avg_duration_per_task_ms": 260202.41792929292,
      "avg_output_tokens_per_task": 31470.176136363636,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "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": 0.9024125477178423,
      "accuracy_std_dev": 0.026053497227534046,
      "run_count": 11,
      "total_questions": 2169,
      "avg_cost_per_task_usd": 0.04220562056247118,
      "avg_duration_per_task_ms": 254266.19502074688,
      "avg_output_tokens_per_task": 31038.156754264637,
      "last_run_timestamp": "2026-09-26T16:06:21.027Z"
    },
    {
      "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": 0.9023565,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.053309800505050506,
      "avg_duration_per_task_ms": 26495.010101010103,
      "avg_output_tokens_per_task": 2451.373737373737,
      "last_run_timestamp": "2026-09-26T08:59:34.279Z"
    },
    {
      "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": 0.9013467199999999,
      "accuracy_std_dev": 0.005345045073860445,
      "run_count": 25,
      "total_questions": 4950,
      "avg_cost_per_task_usd": 0.21570517737373737,
      "avg_duration_per_task_ms": 94985.5094949495,
      "avg_output_tokens_per_task": 7862.312121212121,
      "last_run_timestamp": "2026-09-26T08:55:06.809Z"
    },
    {
      "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": 0.901199375,
      "accuracy_std_dev": 0.005700202593713214,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.2040564772727273,
      "avg_duration_per_task_ms": 98886.75126262626,
      "avg_output_tokens_per_task": 7886.399621212121,
      "last_run_timestamp": "2026-09-26T09:36:04.311Z"
    },
    {
      "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": 0.9006730000000001,
      "accuracy_std_dev": 0.007008969420202443,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1542171708754209,
      "avg_duration_per_task_ms": 127676.21296296296,
      "avg_output_tokens_per_task": 9507.647306397306,
      "last_run_timestamp": "2026-09-01T17:57:22.186Z"
    },
    {
      "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": 0.900673,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.03341542424242424,
      "avg_duration_per_task_ms": 161489.9797979798,
      "avg_output_tokens_per_task": 27673.237373737375,
      "last_run_timestamp": "2026-09-28T18:09:11.509Z"
    },
    {
      "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": 0.9006724572127138,
      "accuracy_std_dev": 0.009841934649752557,
      "run_count": 10,
      "total_questions": 1636,
      "avg_cost_per_task_usd": 0.02167028850855746,
      "avg_duration_per_task_ms": 209667.61246943765,
      "avg_output_tokens_per_task": 17928.619193154034,
      "last_run_timestamp": "2026-09-26T17:07:03.598Z"
    },
    {
      "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": 0.9006104712571067,
      "accuracy_std_dev": 0.058973902356121696,
      "run_count": 8,
      "total_questions": 1583,
      "avg_cost_per_task_usd": 0.14550911939355657,
      "avg_duration_per_task_ms": 301854.2046746684,
      "avg_output_tokens_per_task": 10187.655716993051,
      "last_run_timestamp": "2026-09-26T10:02:09.445Z"
    },
    {
      "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": 0.9004862222222222,
      "accuracy_std_dev": 0.044346110676956096,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.14017589562289562,
      "avg_duration_per_task_ms": 217013.79966329967,
      "avg_output_tokens_per_task": 9267.047699214367,
      "last_run_timestamp": "2026-09-26T09:14:25.270Z"
    },
    {
      "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": 0.9003924999999999,
      "accuracy_std_dev": 0.01706955011855908,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.02751066750841751,
      "avg_duration_per_task_ms": 211685.4351851852,
      "avg_output_tokens_per_task": 25746.074915824916,
      "last_run_timestamp": "2026-09-26T08:50:52.330Z"
    },
    {
      "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": 0.900252375,
      "accuracy_std_dev": 0.024011889023656058,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.15187610795454545,
      "avg_duration_per_task_ms": 607070.2234848485,
      "avg_output_tokens_per_task": 38443.12626262626,
      "last_run_timestamp": "2026-09-26T09:23:06.229Z"
    },
    {
      "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": 0.8999515714285714,
      "accuracy_std_dev": 0.004917020144563225,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.23012572438672438,
      "avg_duration_per_task_ms": 111331.89033189033,
      "avg_output_tokens_per_task": 9001.132756132756,
      "last_run_timestamp": "2026-09-26T08:54:49.538Z"
    },
    {
      "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": 0.8998315,
      "accuracy_std_dev": 0.007575499999999957,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13947029797979799,
      "avg_duration_per_task_ms": 103780.42929292929,
      "avg_output_tokens_per_task": 9170.424242424242,
      "last_run_timestamp": "2026-09-01T17:56:45.430Z"
    },
    {
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "NextBit",
      "accuracy": 0.8995508333333332,
      "accuracy_std_dev": 0.004524373673657892,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.031289763468013466,
      "avg_duration_per_task_ms": 215945.1616161616,
      "avg_output_tokens_per_task": 26542.147306397306,
      "last_run_timestamp": "2026-09-26T09:19:06.104Z"
    },
    {
      "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": 0.8993638888888887,
      "accuracy_std_dev": 0.005599097039194869,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.009368053310886646,
      "avg_duration_per_task_ms": 186510.51515151514,
      "avg_output_tokens_per_task": 26068.07014590348,
      "last_run_timestamp": "2026-09-26T08:51:16.648Z"
    },
    {
      "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": 0.8993387243502051,
      "accuracy_std_dev": 0.026386926357751014,
      "run_count": 8,
      "total_questions": 1462,
      "avg_cost_per_task_usd": 0.027840691518467853,
      "avg_duration_per_task_ms": 740625.9062927497,
      "avg_output_tokens_per_task": 24883.40560875513,
      "last_run_timestamp": "2026-09-12T16:17:57.063Z"
    },
    {
      "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": 0.8989897499999999,
      "accuracy_std_dev": 0.007241239789393777,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.12451013636363638,
      "avg_duration_per_task_ms": 29637.002525252527,
      "avg_output_tokens_per_task": 2284.373737373737,
      "last_run_timestamp": "2026-09-26T08:53:11.291Z"
    },
    {
      "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": 0.8987091666666667,
      "accuracy_std_dev": 0.006255284576970791,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.05856475252525253,
      "avg_duration_per_task_ms": 31561.38973063973,
      "avg_output_tokens_per_task": 2534.496632996633,
      "last_run_timestamp": "2026-09-26T09:18:55.607Z"
    },
    {
      "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": 0.8986660000000002,
      "accuracy_std_dev": 0.006701794386580333,
      "run_count": 26,
      "total_questions": 5148,
      "avg_cost_per_task_usd": 0.22345247746697744,
      "avg_duration_per_task_ms": 95310.8921911422,
      "avg_output_tokens_per_task": 8135.330031080031,
      "last_run_timestamp": "2026-09-26T09:16:27.499Z"
    },
    {
      "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": 0.8978795217391303,
      "accuracy_std_dev": 0.02199333682676437,
      "run_count": 7,
      "total_questions": 1242,
      "avg_cost_per_task_usd": 0.15500312399355878,
      "avg_duration_per_task_ms": 574496.192431562,
      "avg_output_tokens_per_task": 37507.40821256039,
      "last_run_timestamp": "2026-09-26T09:25:58.145Z"
    },
    {
      "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": 0.8978675,
      "accuracy_std_dev": 0.004085377899697062,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.057460788720538725,
      "avg_duration_per_task_ms": 25029.281986531987,
      "avg_output_tokens_per_task": 2489.7180134680134,
      "last_run_timestamp": "2026-09-26T09:20:45.216Z"
    },
    {
      "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": 0.897348959390863,
      "accuracy_std_dev": 0.009933136206097672,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.021320504230118442,
      "avg_duration_per_task_ms": 601175.5499153977,
      "avg_output_tokens_per_task": 17639.832487309646,
      "last_run_timestamp": "2026-09-26T09:15:29.044Z"
    },
    {
      "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": 0.8973064285714285,
      "accuracy_std_dev": 0.009132616193107635,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.19710293722943723,
      "avg_duration_per_task_ms": 95577.20057720058,
      "avg_output_tokens_per_task": 7683.466089466089,
      "last_run_timestamp": "2026-09-26T09:05:17.040Z"
    },
    {
      "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": 0.8973061999999999,
      "accuracy_std_dev": 0.012997001706547546,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.03622809090909091,
      "avg_duration_per_task_ms": 162801.12121212122,
      "avg_output_tokens_per_task": 27180.768686868687,
      "last_run_timestamp": "2026-09-26T08:53:11.955Z"
    },
    {
      "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": 0.8970957500000001,
      "accuracy_std_dev": 0.013079356862533413,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.20200678535353536,
      "avg_duration_per_task_ms": 97703.5946969697,
      "avg_output_tokens_per_task": 7866.200757575758,
      "last_run_timestamp": "2026-09-26T08:52:41.027Z"
    },
    {
      "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": 0.8967507774936062,
      "accuracy_std_dev": 0.0073637026710600875,
      "run_count": 18,
      "total_questions": 3519,
      "avg_cost_per_task_usd": 0.12523853253765274,
      "avg_duration_per_task_ms": 58154.57402671214,
      "avg_output_tokens_per_task": 4750.331628303496,
      "last_run_timestamp": "2026-09-26T08:53:17.186Z"
    },
    {
      "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": 0.8964645000000001,
      "accuracy_std_dev": 0.007575499999999957,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03375818181818182,
      "avg_duration_per_task_ms": 56405.70202020202,
      "avg_output_tokens_per_task": 3599.8333333333335,
      "last_run_timestamp": "2026-09-01T13:22:46.235Z"
    },
    {
      "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": 0.8961193926247291,
      "accuracy_std_dev": 0.016375418175062377,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.11571196312364425,
      "avg_duration_per_task_ms": 621953.1062906724,
      "avg_output_tokens_per_task": 36359.131597975414,
      "last_run_timestamp": "2026-09-26T09:20:04.760Z"
    },
    {
      "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": 0.8960958972972972,
      "accuracy_std_dev": 0.005877557334660528,
      "run_count": 3,
      "total_questions": 555,
      "avg_cost_per_task_usd": 0.1555473855855856,
      "avg_duration_per_task_ms": 83534.15135135135,
      "avg_output_tokens_per_task": 9299.446846846848,
      "last_run_timestamp": "2026-09-01T17:56:10.791Z"
    },
    {
      "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": 0.8959085241877256,
      "accuracy_std_dev": 0.035618009581258374,
      "run_count": 7,
      "total_questions": 1385,
      "avg_cost_per_task_usd": 0.11392027942238267,
      "avg_duration_per_task_ms": 563860.525631769,
      "avg_output_tokens_per_task": 38531.1083032491,
      "last_run_timestamp": "2026-09-26T09:10:17.436Z"
    },
    {
      "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": 0.895623,
      "accuracy_std_dev": 0.006733999999999962,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.016944118686868685,
      "avg_duration_per_task_ms": 304656.29545454547,
      "avg_output_tokens_per_task": 21473.98232323232,
      "last_run_timestamp": "2026-09-26T08:26:07.956Z"
    },
    {
      "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": 0.89562275,
      "accuracy_std_dev": 0.0051888143816771525,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.057555659090909095,
      "avg_duration_per_task_ms": 24846.121212121212,
      "avg_output_tokens_per_task": 2521.655303030303,
      "last_run_timestamp": "2026-09-26T09:36:25.776Z"
    },
    {
      "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": 0.8954697272727271,
      "accuracy_std_dev": 0.006025463303060488,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.20493744031221303,
      "avg_duration_per_task_ms": 100289.02800734619,
      "avg_output_tokens_per_task": 7987.1299357208445,
      "last_run_timestamp": "2026-09-26T08:51:26.325Z"
    },
    {
      "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": 0.8953789065420559,
      "accuracy_std_dev": 0.010355017014034617,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.012003740654205609,
      "avg_duration_per_task_ms": 284653.5062305296,
      "avg_output_tokens_per_task": 27098.440809968848,
      "last_run_timestamp": "2026-09-26T08:56:10.606Z"
    },
    {
      "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": 0.8953486936299293,
      "accuracy_std_dev": 0.019178869929169434,
      "run_count": 5,
      "total_questions": 989,
      "avg_cost_per_task_usd": 0.03527696764408493,
      "avg_duration_per_task_ms": 158313.13549039434,
      "avg_output_tokens_per_task": 26466.745197168857,
      "last_run_timestamp": "2026-09-26T09:36:01.171Z"
    },
    {
      "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": 0.8950614444444446,
      "accuracy_std_dev": 0.007975755054623836,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.12023811111111112,
      "avg_duration_per_task_ms": 57861.70202020202,
      "avg_output_tokens_per_task": 4595.748597081931,
      "last_run_timestamp": "2026-09-26T08:56:21.122Z"
    },
    {
      "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": 0.8949365645569619,
      "accuracy_std_dev": 0.017816727566741866,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.021203940084388185,
      "avg_duration_per_task_ms": 226768.21476793248,
      "avg_output_tokens_per_task": 26685.461603375526,
      "last_run_timestamp": "2026-09-26T08:50:36.105Z"
    },
    {
      "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": 0.8946420862998922,
      "accuracy_std_dev": 0.007735019467331668,
      "run_count": 5,
      "total_questions": 927,
      "avg_cost_per_task_usd": 0.006440504854368933,
      "avg_duration_per_task_ms": 428399.7303128371,
      "avg_output_tokens_per_task": 13363.50377562028,
      "last_run_timestamp": "2026-09-26T08:57:06.874Z"
    },
    {
      "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": 0.8939478279932548,
      "accuracy_std_dev": 0.009294310315828362,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.22570986565486229,
      "avg_duration_per_task_ms": 108477.89263631254,
      "avg_output_tokens_per_task": 8804.630691399663,
      "last_run_timestamp": "2026-09-26T08:50:43.725Z"
    },
    {
      "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": 0.8934859242520689,
      "accuracy_std_dev": 0.016678015536459194,
      "run_count": 8,
      "total_questions": 1571,
      "avg_cost_per_task_usd": 0.11211757733927435,
      "avg_duration_per_task_ms": 393101.5945257798,
      "avg_output_tokens_per_task": 7662.472310630172,
      "last_run_timestamp": "2026-09-26T09:51:40.881Z"
    },
    {
      "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": 0.8931287407407407,
      "accuracy_std_dev": 0.007059243819500409,
      "run_count": 27,
      "total_questions": 5346,
      "avg_cost_per_task_usd": 0.2046095759446315,
      "avg_duration_per_task_ms": 71087.34867190423,
      "avg_output_tokens_per_task": 7454.418069584736,
      "last_run_timestamp": "2026-09-26T09:18:54.068Z"
    },
    {
      "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": 0.8930976666666668,
      "accuracy_std_dev": 0.010182556173290759,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.05509197895622895,
      "avg_duration_per_task_ms": 49604.97474747475,
      "avg_output_tokens_per_task": 3808.635521885522,
      "last_run_timestamp": "2026-09-01T13:23:50.897Z"
    },
    {
      "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": 0.8930975,
      "accuracy_std_dev": 0.0025254999999999583,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.054297787878787876,
      "avg_duration_per_task_ms": 31463.27525252525,
      "avg_output_tokens_per_task": 2503.689393939394,
      "last_run_timestamp": "2026-09-26T08:53:18.330Z"
    },
    {
      "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-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Novita",
      "accuracy": 0.8920512903225806,
      "accuracy_std_dev": 0.019911726970242915,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.13115581593927894,
      "avg_duration_per_task_ms": 517762.9652118912,
      "avg_output_tokens_per_task": 37601.20493358634,
      "last_run_timestamp": "2026-09-26T09:00:46.358Z"
    },
    {
      "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": 0.8918440425531915,
      "accuracy_std_dev": 0.0082731599766957,
      "run_count": 3,
      "total_questions": 564,
      "avg_cost_per_task_usd": 0.050205347517730495,
      "avg_duration_per_task_ms": 44870.67021276596,
      "avg_output_tokens_per_task": 3654.1524822695037,
      "last_run_timestamp": "2026-09-01T13:21:31.062Z"
    },
    {
      "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": "ModelRun",
      "accuracy": 0.8914139999999999,
      "accuracy_std_dev": 0.023312104420236266,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.086096711489899,
      "avg_duration_per_task_ms": 172501.05744949495,
      "avg_output_tokens_per_task": 29063.86047979798,
      "last_run_timestamp": "2026-09-26T16:08:44.458Z"
    },
    {
      "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": 0.8909932500000002,
      "accuracy_std_dev": 0.010197164639619184,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.00789814393939394,
      "avg_duration_per_task_ms": 268244.18434343435,
      "avg_output_tokens_per_task": 27958.415404040403,
      "last_run_timestamp": "2026-08-08T10:06:09.471Z"
    },
    {
      "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": 0.8905724999999999,
      "accuracy_std_dev": 0.011784499999999976,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.027024376262626264,
      "avg_duration_per_task_ms": 278293.2196969697,
      "avg_output_tokens_per_task": 26273.641414141413,
      "last_run_timestamp": "2026-09-19T09:00:31.402Z"
    },
    {
      "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": 0.890572,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.21863568181818183,
      "avg_duration_per_task_ms": 106364.56818181818,
      "avg_output_tokens_per_task": 8520.704545454546,
      "last_run_timestamp": "2026-09-01T13:30:25.730Z"
    },
    {
      "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": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.1261595202020202,
      "avg_duration_per_task_ms": 738082.0303030303,
      "avg_output_tokens_per_task": 50200.232323232325,
      "last_run_timestamp": "2026-09-23T06:12:17.757Z"
    },
    {
      "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": 0.8905389900990098,
      "accuracy_std_dev": 0.007335499999999995,
      "run_count": 2,
      "total_questions": 303,
      "avg_cost_per_task_usd": 0.01631713201320132,
      "avg_duration_per_task_ms": 272632.9339933993,
      "avg_output_tokens_per_task": 22974.214521452144,
      "last_run_timestamp": "2026-09-26T09:25:09.058Z"
    },
    {
      "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": 0.8897369109414758,
      "accuracy_std_dev": 0.022583806568842405,
      "run_count": 6,
      "total_questions": 1179,
      "avg_cost_per_task_usd": 0.1931469830364716,
      "avg_duration_per_task_ms": 1259659.2798982188,
      "avg_output_tokens_per_task": 56717.179813401184,
      "last_run_timestamp": "2026-08-29T18:18:56.485Z"
    },
    {
      "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": 0.8897304999999999,
      "accuracy_std_dev": 0.005892499999999967,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0318945303030303,
      "avg_duration_per_task_ms": 52439.469696969696,
      "avg_output_tokens_per_task": 3386.712121212121,
      "last_run_timestamp": "2026-09-01T13:38:25.421Z"
    },
    {
      "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": 0.8897192682926828,
      "accuracy_std_dev": 0.012292154908071387,
      "run_count": 13,
      "total_questions": 2542,
      "avg_cost_per_task_usd": 0.0036178985051140834,
      "avg_duration_per_task_ms": 277074.0971675846,
      "avg_output_tokens_per_task": 19808.066876475215,
      "last_run_timestamp": "2026-09-26T09:01:36.155Z"
    },
    {
      "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": 0.888889,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00832349494949495,
      "avg_duration_per_task_ms": 86259.9494949495,
      "avg_output_tokens_per_task": 13087.217171717171,
      "last_run_timestamp": "2026-09-24T09:11:45.985Z"
    },
    {
      "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": 0.8888888430379747,
      "accuracy_std_dev": 0.009622693519820156,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.21254848649789027,
      "avg_duration_per_task_ms": 102777.66371308017,
      "avg_output_tokens_per_task": 8273.845147679325,
      "last_run_timestamp": "2026-09-26T09:15:06.377Z"
    },
    {
      "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": 0.8887592307692308,
      "accuracy_std_dev": 0.013361790161926698,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.012804753690753689,
      "avg_duration_per_task_ms": 213849.3986013986,
      "avg_output_tokens_per_task": 24454.132867132867,
      "last_run_timestamp": "2026-09-26T08:57:03.139Z"
    },
    {
      "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": 0.888552,
      "accuracy_std_dev": 0.03117712682079604,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1316316707070707,
      "avg_duration_per_task_ms": 39126.25959595959,
      "avg_output_tokens_per_task": 2423.617171717172,
      "last_run_timestamp": "2026-09-26T08:53:20.595Z"
    },
    {
      "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": 0.8884639560229446,
      "accuracy_std_dev": 0.012263269453369714,
      "run_count": 11,
      "total_questions": 2092,
      "avg_cost_per_task_usd": 0.09539477390057362,
      "avg_duration_per_task_ms": 534326.7394837476,
      "avg_output_tokens_per_task": 29086.82982791587,
      "last_run_timestamp": "2026-09-26T08:56:57.645Z"
    },
    {
      "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": 0.8883868300395255,
      "accuracy_std_dev": 0.008563760488459709,
      "run_count": 27,
      "total_questions": 5313,
      "avg_cost_per_task_usd": 0.1336053077357425,
      "avg_duration_per_task_ms": 59533.9862601167,
      "avg_output_tokens_per_task": 4780.512516469038,
      "last_run_timestamp": "2026-09-26T09:19:13.119Z"
    },
    {
      "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": 0.8876862857142855,
      "accuracy_std_dev": 0.018697250017157118,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.13258516161616163,
      "avg_duration_per_task_ms": 538569.6168831169,
      "avg_output_tokens_per_task": 33498.70923520924,
      "last_run_timestamp": "2026-09-26T09:29:09.038Z"
    },
    {
      "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": 0.8875662777777777,
      "accuracy_std_dev": 0.03225232181080149,
      "run_count": 4,
      "total_questions": 756,
      "avg_cost_per_task_usd": 0.0045816917989418,
      "avg_duration_per_task_ms": 604811.5727513228,
      "avg_output_tokens_per_task": 27586.727513227514,
      "last_run_timestamp": "2026-09-26T08:58:10.498Z"
    },
    {
      "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": 0.8872987222222223,
      "accuracy_std_dev": 0.006285923036662124,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.24028202581369246,
      "avg_duration_per_task_ms": 84935.12850729517,
      "avg_output_tokens_per_task": 8938.721380471381,
      "last_run_timestamp": "2026-09-26T08:53:09.256Z"
    },
    {
      "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": 0.8872055000000001,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20363688131313135,
      "avg_duration_per_task_ms": 99426.9191919192,
      "avg_output_tokens_per_task": 7920.752525252526,
      "last_run_timestamp": "2026-09-01T13:21:53.444Z"
    },
    {
      "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": 0.887153,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 192,
      "avg_cost_per_task_usd": 0.112435484375,
      "avg_duration_per_task_ms": 652216.9166666666,
      "avg_output_tokens_per_task": 32066.067708333332,
      "last_run_timestamp": "2026-09-29T02:34:51.989Z"
    },
    {
      "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": "Azure",
      "accuracy": 0.8868242601351352,
      "accuracy_std_dev": 0.009599924258960477,
      "run_count": 9,
      "total_questions": 1776,
      "avg_cost_per_task_usd": 0.20454106644144146,
      "avg_duration_per_task_ms": 100059.57263513513,
      "avg_output_tokens_per_task": 7955.086148648648,
      "last_run_timestamp": "2026-09-26T08:50:54.262Z"
    },
    {
      "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": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.8866439999999999,
      "accuracy_std_dev": 0.012472582491208459,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.008770978114478115,
      "avg_duration_per_task_ms": 61645.40740740741,
      "avg_output_tokens_per_task": 12897.501683501683,
      "last_run_timestamp": "2026-09-24T09:11:57.282Z"
    },
    {
      "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": 0.8864601803278688,
      "accuracy_std_dev": 0.006211138077858382,
      "run_count": 3,
      "total_questions": 549,
      "avg_cost_per_task_usd": 0.009764420765027322,
      "avg_duration_per_task_ms": 50630.12750455373,
      "avg_output_tokens_per_task": 7246.790528233151,
      "last_run_timestamp": "2026-09-01T17:51:11.087Z"
    },
    {
      "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": 0.8864530691399664,
      "accuracy_std_dev": 0.01029973730112676,
      "run_count": 6,
      "total_questions": 1186,
      "avg_cost_per_task_usd": 0.023206354974704888,
      "avg_duration_per_task_ms": 81414.45531197301,
      "avg_output_tokens_per_task": 7880.962900505902,
      "last_run_timestamp": "2026-09-01T17:55:01.420Z"
    },
    {
      "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": 0.8863635000000001,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0069983762626262625,
      "avg_duration_per_task_ms": 635651.6565656565,
      "avg_output_tokens_per_task": 21974.18181818182,
      "last_run_timestamp": "2026-09-26T10:42:27.487Z"
    },
    {
      "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": 0.8862431428571429,
      "accuracy_std_dev": 0.008624548475610979,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.016605583694083693,
      "avg_duration_per_task_ms": 131101.75468975469,
      "avg_output_tokens_per_task": 25668.465367965367,
      "last_run_timestamp": "2026-09-26T08:53:15.047Z"
    },
    {
      "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": 0.8860826666666668,
      "accuracy_std_dev": 0.004199386807089298,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0278726632996633,
      "avg_duration_per_task_ms": 157091.45622895623,
      "avg_output_tokens_per_task": 23053.37037037037,
      "last_run_timestamp": "2026-09-26T09:47:15.375Z"
    },
    {
      "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": 0.8860007604485219,
      "accuracy_std_dev": 0.0277911335645022,
      "run_count": 5,
      "total_questions": 981,
      "avg_cost_per_task_usd": 0.04559343119266055,
      "avg_duration_per_task_ms": 698776.9031600407,
      "avg_output_tokens_per_task": 26722.172273190623,
      "last_run_timestamp": "2026-08-22T10:03:34.994Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Cloudflare",
      "accuracy": 0.8849736038095237,
      "accuracy_std_dev": 0.03742100947662959,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.1364461853968254,
      "avg_duration_per_task_ms": 794770.7498412698,
      "avg_output_tokens_per_task": 34155.479365079365,
      "last_run_timestamp": "2026-09-26T10:09:19.111Z"
    },
    {
      "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": 0.8849202417582418,
      "accuracy_std_dev": 0.01938538920596302,
      "run_count": 3,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.04324962454212454,
      "avg_duration_per_task_ms": 205056.15750915752,
      "avg_output_tokens_per_task": 33761.00366300366,
      "last_run_timestamp": "2026-09-26T10:02:16.696Z"
    },
    {
      "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": 0.88468,
      "accuracy_std_dev": 0.005892000000000008,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.009158123737373737,
      "avg_duration_per_task_ms": 142387.94696969696,
      "avg_output_tokens_per_task": 32237.108585858587,
      "last_run_timestamp": "2026-09-26T09:33:13.989Z"
    },
    {
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Relace",
      "accuracy": 0.8841129772439951,
      "accuracy_std_dev": 0.0033388888402431233,
      "run_count": 4,
      "total_questions": 791,
      "avg_cost_per_task_usd": 0.011443802781289506,
      "avg_duration_per_task_ms": 364670.14664981037,
      "avg_output_tokens_per_task": 22004.87104930468,
      "last_run_timestamp": "2026-09-26T12:21:31.561Z"
    },
    {
      "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": 0.8839537178510998,
      "accuracy_std_dev": 0.013099282603342133,
      "run_count": 12,
      "total_questions": 2364,
      "avg_cost_per_task_usd": 0.00725531429780034,
      "avg_duration_per_task_ms": 489249.17766497465,
      "avg_output_tokens_per_task": 25997.4179357022,
      "last_run_timestamp": "2026-09-26T08:56:41.992Z"
    },
    {
      "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": 0.883838,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.03956430303030303,
      "avg_duration_per_task_ms": 264270.07575757575,
      "avg_output_tokens_per_task": 34221.91414141414,
      "last_run_timestamp": "2026-09-28T18:58:44.942Z"
    },
    {
      "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": 0.883838,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.11161540404040404,
      "avg_duration_per_task_ms": 930549.2727272727,
      "avg_output_tokens_per_task": 44407.63636363636,
      "last_run_timestamp": "2026-09-23T15:38:01.445Z"
    },
    {
      "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": 0.8835014000000001,
      "accuracy_std_dev": 0.025562678362018323,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.14118799797979797,
      "avg_duration_per_task_ms": 41635.47171717172,
      "avg_output_tokens_per_task": 2614.5828282828284,
      "last_run_timestamp": "2026-09-26T08:54:00.492Z"
    },
    {
      "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": 0.8832771111111111,
      "accuracy_std_dev": 0.009589207202132645,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.02110603142536476,
      "avg_duration_per_task_ms": 350549.90852974186,
      "avg_output_tokens_per_task": 24400.115039281707,
      "last_run_timestamp": "2026-09-26T08:58:18.787Z"
    },
    {
      "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": 0.8829965,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13198183333333333,
      "avg_duration_per_task_ms": 254155.68181818182,
      "avg_output_tokens_per_task": 43876.38636363636,
      "last_run_timestamp": "2026-09-01T17:57:36.744Z"
    },
    {
      "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": 0.882716,
      "accuracy_std_dev": 0.006496204173720757,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.027370449494949496,
      "avg_duration_per_task_ms": 130101.09932659933,
      "avg_output_tokens_per_task": 22617.942760942762,
      "last_run_timestamp": "2026-09-26T10:08:37.923Z"
    },
    {
      "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": 0.8824913999999999,
      "accuracy_std_dev": 0.008205959897537906,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.015855806060606062,
      "avg_duration_per_task_ms": 286632.0282828283,
      "avg_output_tokens_per_task": 24219.753535353535,
      "last_run_timestamp": "2026-09-26T09:30:12.511Z"
    },
    {
      "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": 0.8823229999999999,
      "accuracy_std_dev": 0.01560213662932099,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.11433169949494948,
      "avg_duration_per_task_ms": 635262.6303030303,
      "avg_output_tokens_per_task": 29892.36919191919,
      "last_run_timestamp": "2026-09-26T09:11:00.979Z"
    },
    {
      "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": 0.8821944899391673,
      "accuracy_std_dev": 0.012966199129335096,
      "run_count": 22,
      "total_questions": 4274,
      "avg_cost_per_task_usd": 0.19237118413664017,
      "avg_duration_per_task_ms": 381859.63991576975,
      "avg_output_tokens_per_task": 30958.449695835283,
      "last_run_timestamp": "2026-09-26T08:58:33.306Z"
    },
    {
      "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": 0.8821546666666666,
      "accuracy_std_dev": 0.001374572014200141,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.02880540235690236,
      "avg_duration_per_task_ms": 99924.20707070707,
      "avg_output_tokens_per_task": 23871.781144781144,
      "last_run_timestamp": "2026-09-26T09:49:42.288Z"
    },
    {
      "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": 0.8821545714285715,
      "accuracy_std_dev": 0.014256633562332045,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.18132904473304476,
      "avg_duration_per_task_ms": 297976.3398268398,
      "avg_output_tokens_per_task": 45509.72438672439,
      "last_run_timestamp": "2026-09-26T09:19:29.410Z"
    },
    {
      "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": 0.8818613645976111,
      "accuracy_std_dev": 0.010002367123265743,
      "run_count": 27,
      "total_questions": 5107,
      "avg_cost_per_task_usd": 0.1356296686900333,
      "avg_duration_per_task_ms": 62377.717642451535,
      "avg_output_tokens_per_task": 4931.54278441355,
      "last_run_timestamp": "2026-09-26T08:58:07.474Z"
    },
    {
      "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": 0.8809521930501931,
      "accuracy_std_dev": 0.01086493577983783,
      "run_count": 4,
      "total_questions": 777,
      "avg_cost_per_task_usd": 0.021439528957528958,
      "avg_duration_per_task_ms": 189160.3075933076,
      "avg_output_tokens_per_task": 23673.939510939512,
      "last_run_timestamp": "2026-09-29T03:31:47.451Z"
    },
    {
      "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": 0.8806582222222221,
      "accuracy_std_dev": 0.005109133929167741,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.005640715488215489,
      "avg_duration_per_task_ms": 362315.8209876543,
      "avg_output_tokens_per_task": 28732.147025813694,
      "last_run_timestamp": "2026-09-26T09:34:56.325Z"
    },
    {
      "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-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Wafer",
      "accuracy": 0.8799936573033706,
      "accuracy_std_dev": 0.07692276680482653,
      "run_count": 12,
      "total_questions": 2136,
      "avg_cost_per_task_usd": 0.008731384363295881,
      "avg_duration_per_task_ms": 144796.46582397004,
      "avg_output_tokens_per_task": 23515.395599250936,
      "last_run_timestamp": "2026-09-26T08:54:04.117Z"
    },
    {
      "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": "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "GMICloud",
      "accuracy": 0.8793486666666668,
      "accuracy_std_dev": 0.0031744380430068155,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.07590014478114479,
      "avg_duration_per_task_ms": 323507.5353535354,
      "avg_output_tokens_per_task": 20962.156565656565,
      "last_run_timestamp": "2026-08-29T19:12:43.653Z"
    },
    {
      "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": 0.8792515918367347,
      "accuracy_std_dev": 0.00992611788026112,
      "run_count": 9,
      "total_questions": 1764,
      "avg_cost_per_task_usd": 0.004467375283446712,
      "avg_duration_per_task_ms": 383208.2947845805,
      "avg_output_tokens_per_task": 18476.153628117914,
      "last_run_timestamp": "2026-09-26T10:33:25.209Z"
    },
    {
      "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": "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": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.8787875,
      "accuracy_std_dev": 0.008417500000000022,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0075271136363636365,
      "avg_duration_per_task_ms": 80896.84595959596,
      "avg_output_tokens_per_task": 8213.79797979798,
      "last_run_timestamp": "2026-09-01T17:51:13.979Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Novita",
      "accuracy": 0.8787658893709328,
      "accuracy_std_dev": 0.007021239756217274,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.008512674620390456,
      "avg_duration_per_task_ms": 290385.6428054953,
      "avg_output_tokens_per_task": 30108.634851771512,
      "last_run_timestamp": "2026-08-29T19:12:25.584Z"
    },
    {
      "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": 0.8785529767441861,
      "accuracy_std_dev": 0.101852,
      "run_count": 2,
      "total_questions": 258,
      "avg_cost_per_task_usd": 0.1083256472868217,
      "avg_duration_per_task_ms": 620596.2751937985,
      "avg_output_tokens_per_task": 16165.639534883721,
      "last_run_timestamp": "2026-09-01T16:19:34.300Z"
    },
    {
      "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": 0.8784510000000001,
      "accuracy_std_dev": 0.011536704780828876,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.022268736363636365,
      "avg_duration_per_task_ms": 229673.4696969697,
      "avg_output_tokens_per_task": 24972.11919191919,
      "last_run_timestamp": "2026-09-26T09:19:28.485Z"
    },
    {
      "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": 0.8784135555555556,
      "accuracy_std_dev": 0.017145903852336865,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.023216065656565654,
      "avg_duration_per_task_ms": 214871.89786756455,
      "avg_output_tokens_per_task": 24869.92312008979,
      "last_run_timestamp": "2026-09-26T09:39:38.744Z"
    },
    {
      "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": 0.8783669000000001,
      "accuracy_std_dev": 0.03386573186408348,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.14244708787878788,
      "avg_duration_per_task_ms": 338071.06212121213,
      "avg_output_tokens_per_task": 33002.009090909094,
      "last_run_timestamp": "2026-09-26T08:53:37.509Z"
    },
    {
      "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": 0.8783643322981367,
      "accuracy_std_dev": 0.020055520558938385,
      "run_count": 10,
      "total_questions": 1932,
      "avg_cost_per_task_usd": 0.1293156180124224,
      "avg_duration_per_task_ms": 63254.61180124224,
      "avg_output_tokens_per_task": 4948.346273291925,
      "last_run_timestamp": "2026-09-26T08:52:39.749Z"
    },
    {
      "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": 0.8779459999999999,
      "accuracy_std_dev": 0.014310000000000045,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13208357575757576,
      "avg_duration_per_task_ms": 257170.34595959596,
      "avg_output_tokens_per_task": 43907.46464646464,
      "last_run_timestamp": "2026-09-01T17:57:59.987Z"
    },
    {
      "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": 0.8776173653429603,
      "accuracy_std_dev": 0.01555889568113131,
      "run_count": 7,
      "total_questions": 1385,
      "avg_cost_per_task_usd": 0.08483577111913358,
      "avg_duration_per_task_ms": 220187.85559566788,
      "avg_output_tokens_per_task": 21593.077978339352,
      "last_run_timestamp": "2026-09-26T10:06:10.499Z"
    },
    {
      "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": 0.8771045000000001,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14050905303030303,
      "avg_duration_per_task_ms": 185398.66919191918,
      "avg_output_tokens_per_task": 46713.32070707071,
      "last_run_timestamp": "2026-09-01T17:54:30.227Z"
    },
    {
      "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": 0.8771039999999999,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.007407005050505051,
      "avg_duration_per_task_ms": 78924.21717171717,
      "avg_output_tokens_per_task": 8106.714646464647,
      "last_run_timestamp": "2026-09-01T17:50:43.967Z"
    },
    {
      "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": 0.8770211787234045,
      "accuracy_std_dev": 0.03195742141593335,
      "run_count": 6,
      "total_questions": 1175,
      "avg_cost_per_task_usd": 0.0032351353191489365,
      "avg_duration_per_task_ms": 354647.9063829787,
      "avg_output_tokens_per_task": 26251.109787234043,
      "last_run_timestamp": "2026-09-19T08:54:45.345Z"
    },
    {
      "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": 0.8768792503888027,
      "accuracy_std_dev": 0.0068076105343578604,
      "run_count": 4,
      "total_questions": 643,
      "avg_cost_per_task_usd": 0.007664079315707621,
      "avg_duration_per_task_ms": 987610.4805598756,
      "avg_output_tokens_per_task": 30419.066874027994,
      "last_run_timestamp": "2026-09-26T09:19:44.251Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "WandB",
      "accuracy": 0.8762625,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.008190393939393938,
      "avg_duration_per_task_ms": 699109.8106060605,
      "avg_output_tokens_per_task": 28884.636363636364,
      "last_run_timestamp": "2026-07-18T08:31:23.912Z"
    },
    {
      "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": 0.8762625,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.15284478282828282,
      "avg_duration_per_task_ms": 339007.8914141414,
      "avg_output_tokens_per_task": 34511.28535353536,
      "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-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.8757013333333332,
      "accuracy_std_dev": 0.01234889446693734,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.00532792760942761,
      "avg_duration_per_task_ms": 293763.77946127945,
      "avg_output_tokens_per_task": 28613.604377104377,
      "last_run_timestamp": "2026-08-29T18:22:03.157Z"
    },
    {
      "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": 0.8756311250000001,
      "accuracy_std_dev": 0.007549310638023532,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0497966994949495,
      "avg_duration_per_task_ms": 300230.047979798,
      "avg_output_tokens_per_task": 21153.447601010103,
      "last_run_timestamp": "2026-08-29T18:36:08.027Z"
    },
    {
      "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": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.8748596666666667,
      "accuracy_std_dev": 0.006198499675100621,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.003131114478114478,
      "avg_duration_per_task_ms": 32869.14646464647,
      "avg_output_tokens_per_task": 5711.222222222223,
      "last_run_timestamp": "2026-09-23T06:08:44.627Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Novita",
      "accuracy": 0.8744972514285714,
      "accuracy_std_dev": 0.006555065674728234,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.12269054984126984,
      "avg_duration_per_task_ms": 985188.8444444444,
      "avg_output_tokens_per_task": 43152.0253968254,
      "last_run_timestamp": "2026-08-29T19:07:02.104Z"
    },
    {
      "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": 0.8740740000000001,
      "accuracy_std_dev": 0.03472083287595504,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1182012676767677,
      "avg_duration_per_task_ms": 28405.98686868687,
      "avg_output_tokens_per_task": 2155.238383838384,
      "last_run_timestamp": "2026-09-26T08:52:17.920Z"
    },
    {
      "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": 0.874074,
      "accuracy_std_dev": 0.004178433821421624,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07029311111111111,
      "avg_duration_per_task_ms": 171365.84242424244,
      "avg_output_tokens_per_task": 11443.449494949495,
      "last_run_timestamp": "2026-09-26T09:10:25.930Z"
    },
    {
      "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": 0.8737373,
      "accuracy_std_dev": 0.01570253542616605,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.11062107323232323,
      "avg_duration_per_task_ms": 229265.31616161615,
      "avg_output_tokens_per_task": 31458.501515151514,
      "last_run_timestamp": "2026-09-26T08:52:34.744Z"
    },
    {
      "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": 0.8737372500000001,
      "accuracy_std_dev": 0.019341996831700202,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.12990520643939396,
      "avg_duration_per_task_ms": 63171.816287878784,
      "avg_output_tokens_per_task": 4978.773358585859,
      "last_run_timestamp": "2026-09-26T08:54:39.535Z"
    },
    {
      "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": 0.8727692590673575,
      "accuracy_std_dev": 0.006680537619749563,
      "run_count": 3,
      "total_questions": 579,
      "avg_cost_per_task_usd": 0.10616132815198619,
      "avg_duration_per_task_ms": 1070108.6269430052,
      "avg_output_tokens_per_task": 42555.205526770296,
      "last_run_timestamp": "2026-08-01T17:53:10.275Z"
    },
    {
      "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": 0.8726002303664923,
      "accuracy_std_dev": 0.013852891908270337,
      "run_count": 3,
      "total_questions": 573,
      "avg_cost_per_task_usd": 0.10812300000000001,
      "avg_duration_per_task_ms": 1136520.2198952879,
      "avg_output_tokens_per_task": 43220.56020942408,
      "last_run_timestamp": "2026-08-01T17:46:48.312Z"
    },
    {
      "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": 0.8724287134608445,
      "accuracy_std_dev": 0.032129376762123,
      "run_count": 14,
      "total_questions": 2771,
      "avg_cost_per_task_usd": 0.1523031241429087,
      "avg_duration_per_task_ms": 425564.390111873,
      "avg_output_tokens_per_task": 11498.582100324793,
      "last_run_timestamp": "2026-09-26T09:12:17.926Z"
    },
    {
      "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": 0.872292240506329,
      "accuracy_std_dev": 0.013194491052666223,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.00790528016877637,
      "avg_duration_per_task_ms": 421318.1687763713,
      "avg_output_tokens_per_task": 21316.46075949367,
      "last_run_timestamp": "2026-09-26T09:05:47.438Z"
    },
    {
      "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": 0.8720540000000001,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.22419905555555555,
      "avg_duration_per_task_ms": 276263.6666666667,
      "avg_output_tokens_per_task": 22325.227272727272,
      "last_run_timestamp": "2026-09-01T13:30:31.593Z"
    },
    {
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8714926666666667,
      "accuracy_std_dev": 0.0067804647005611455,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03188070202020202,
      "avg_duration_per_task_ms": 240511.38888888888,
      "avg_output_tokens_per_task": 26418.183501683503,
      "last_run_timestamp": "2026-09-26T08:53:40.686Z"
    },
    {
      "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": 0.871212,
      "accuracy_std_dev": 0.010943000000000036,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.25165717424242423,
      "avg_duration_per_task_ms": 248967.9292929293,
      "avg_output_tokens_per_task": 23863.48232323232,
      "last_run_timestamp": "2026-09-01T13:53:50.899Z"
    },
    {
      "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": 0.8711222842639593,
      "accuracy_std_dev": 0.0010748023074035804,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.04708872588832487,
      "avg_duration_per_task_ms": 168365.97292724196,
      "avg_output_tokens_per_task": 39058.98984771573,
      "last_run_timestamp": "2026-09-26T10:21:09.898Z"
    },
    {
      "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": 0.8710750330630069,
      "accuracy_std_dev": 0.028110456226890915,
      "run_count": 9,
      "total_questions": 1603,
      "avg_cost_per_task_usd": 0.09870775982532752,
      "avg_duration_per_task_ms": 221158.11790393014,
      "avg_output_tokens_per_task": 26829.437305053027,
      "last_run_timestamp": "2026-09-26T09:19:13.012Z"
    },
    {
      "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": 0.8710435000000001,
      "accuracy_std_dev": 0.015469940324707153,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.007902119696969698,
      "avg_duration_per_task_ms": 203714.64696969697,
      "avg_output_tokens_per_task": 28019.556565656567,
      "last_run_timestamp": "2026-09-26T09:36:46.529Z"
    },
    {
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Together",
      "accuracy": 0.8709859808450705,
      "accuracy_std_dev": 0.00995822887470397,
      "run_count": 9,
      "total_questions": 1775,
      "avg_cost_per_task_usd": 0.007974392676056338,
      "avg_duration_per_task_ms": 311914.2202816901,
      "avg_output_tokens_per_task": 28209.84957746479,
      "last_run_timestamp": "2026-09-26T09:58:11.632Z"
    },
    {
      "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": "OpenInference",
      "accuracy": 0.87037,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.013891080808080807,
      "avg_duration_per_task_ms": 291446.38383838383,
      "avg_output_tokens_per_task": 27719.68181818182,
      "last_run_timestamp": "2026-09-29T02:39:06.144Z"
    },
    {
      "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": "WandB",
      "accuracy": 0.870085,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.11565565128205128,
      "avg_duration_per_task_ms": 2243159.2153846156,
      "avg_output_tokens_per_task": 32886.94358974359,
      "last_run_timestamp": "2026-07-11T08:42:46.205Z"
    },
    {
      "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": 0.8699951448047163,
      "accuracy_std_dev": 0.03969212027542979,
      "run_count": 14,
      "total_questions": 2714,
      "avg_cost_per_task_usd": 0.00531859358879882,
      "avg_duration_per_task_ms": 114563.89867354458,
      "avg_output_tokens_per_task": 20095.400515843772,
      "last_run_timestamp": "2026-09-26T08:53:37.618Z"
    },
    {
      "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-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Novita",
      "accuracy": 0.8698091666666667,
      "accuracy_std_dev": 0.01883877588394628,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07924426515151516,
      "avg_duration_per_task_ms": 527347.8063973064,
      "avg_output_tokens_per_task": 34630.43771043771,
      "last_run_timestamp": "2026-08-29T19:21:48.221Z"
    },
    {
      "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": 0.8697745034965034,
      "accuracy_std_dev": 0.09942524551683361,
      "run_count": 11,
      "total_questions": 2145,
      "avg_cost_per_task_usd": 0.026464635431235433,
      "avg_duration_per_task_ms": 212770.69836829836,
      "avg_output_tokens_per_task": 23008.030303030304,
      "last_run_timestamp": "2026-09-26T16:07:03.311Z"
    },
    {
      "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": 0.8695795853658537,
      "accuracy_std_dev": 0.014776870412912198,
      "run_count": 5,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.010841884146341461,
      "avg_duration_per_task_ms": 398387.18699186994,
      "avg_output_tokens_per_task": 22967.41768292683,
      "last_run_timestamp": "2026-09-26T09:27:32.975Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.86938268,
      "accuracy_std_dev": 0.008067879753335774,
      "run_count": 7,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.004681328148148148,
      "avg_duration_per_task_ms": 490155.47185185185,
      "avg_output_tokens_per_task": 28757.50888888889,
      "last_run_timestamp": "2026-08-29T19:32:50.553Z"
    },
    {
      "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": 0.869318125,
      "accuracy_std_dev": 0.03295688160474795,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.10640666414141413,
      "avg_duration_per_task_ms": 513815.1142676768,
      "avg_output_tokens_per_task": 34567.331439393936,
      "last_run_timestamp": "2026-09-26T08:56:11.901Z"
    },
    {
      "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": 0.8693140356415479,
      "accuracy_std_dev": 0.004604825312647615,
      "run_count": 5,
      "total_questions": 982,
      "avg_cost_per_task_usd": 0.11175445315682281,
      "avg_duration_per_task_ms": 722023.3411405295,
      "avg_output_tokens_per_task": 31928.803462321794,
      "last_run_timestamp": "2026-08-29T18:36:08.021Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8691076666666667,
      "accuracy_std_dev": 0.07237401544446428,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.12576101936026934,
      "avg_duration_per_task_ms": 1221265.2171717172,
      "avg_output_tokens_per_task": 55812.33249158249,
      "last_run_timestamp": "2026-08-29T20:15:38.359Z"
    },
    {
      "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": 0.8689024817073171,
      "accuracy_std_dev": 0.04756926554684933,
      "run_count": 3,
      "total_questions": 328,
      "avg_cost_per_task_usd": 0.014946170731707318,
      "avg_duration_per_task_ms": 479610.80792682926,
      "avg_output_tokens_per_task": 32324.884146341465,
      "last_run_timestamp": "2026-09-26T08:55:13.371Z"
    },
    {
      "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-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.868687,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0028392929292929293,
      "avg_duration_per_task_ms": 49008.020202020205,
      "avg_output_tokens_per_task": 5520.065656565656,
      "last_run_timestamp": "2026-09-23T06:08:55.567Z"
    },
    {
      "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": "ModelRun",
      "accuracy": 0.8685324351145038,
      "accuracy_std_dev": 0.008327000000000029,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.08006628498727736,
      "avg_duration_per_task_ms": 605980.251908397,
      "avg_output_tokens_per_task": 42020.10687022901,
      "last_run_timestamp": "2026-07-11T08:27:26.027Z"
    },
    {
      "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": 0.8684572837465565,
      "accuracy_std_dev": 0.006350273025626523,
      "run_count": 4,
      "total_questions": 726,
      "avg_cost_per_task_usd": 0.0059327286501377405,
      "avg_duration_per_task_ms": 153900.47520661156,
      "avg_output_tokens_per_task": 21060.526170798898,
      "last_run_timestamp": "2026-08-08T09:25:45.716Z"
    },
    {
      "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": 0.8683500000000001,
      "accuracy_std_dev": 0.014020687172888494,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.08152287070707072,
      "avg_duration_per_task_ms": 190241.53434343435,
      "avg_output_tokens_per_task": 13202.136363636364,
      "last_run_timestamp": "2026-09-26T08:55:09.429Z"
    },
    {
      "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": 0.86835,
      "accuracy_std_dev": 0.01103428752570821,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07775004545454545,
      "avg_duration_per_task_ms": 107497.15858585859,
      "avg_output_tokens_per_task": 12580.208080808081,
      "last_run_timestamp": "2026-09-26T08:57:01.878Z"
    },
    {
      "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": 0.8681256666666667,
      "accuracy_std_dev": 0.02130941144710999,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.02939427441077441,
      "avg_duration_per_task_ms": 110502.21548821549,
      "avg_output_tokens_per_task": 19394.752525252527,
      "last_run_timestamp": "2026-09-26T09:38:56.771Z"
    },
    {
      "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": 0.8681256666666665,
      "accuracy_std_dev": 0.00396809470093087,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.008190090067340068,
      "avg_duration_per_task_ms": 450083.40993265994,
      "avg_output_tokens_per_task": 29007.585016835015,
      "last_run_timestamp": "2026-08-29T18:13:33.416Z"
    },
    {
      "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": 0.86798182245827,
      "accuracy_std_dev": 0.027206693677108224,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.11290408548305515,
      "avg_duration_per_task_ms": 768015.9382903388,
      "avg_output_tokens_per_task": 30078.80880121396,
      "last_run_timestamp": "2026-09-26T09:22:13.122Z"
    },
    {
      "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": 0.8676579999999999,
      "accuracy_std_dev": 0.007874153076151538,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.006459232323232324,
      "avg_duration_per_task_ms": 253155.9539842873,
      "avg_output_tokens_per_task": 18088.584736251403,
      "last_run_timestamp": "2026-09-26T08:54:07.885Z"
    },
    {
      "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": 0.8672838333333331,
      "accuracy_std_dev": 0.008551917044668319,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.2189638569023569,
      "avg_duration_per_task_ms": 1248558.3703703703,
      "avg_output_tokens_per_task": 55040.72895622896,
      "last_run_timestamp": "2026-08-29T18:24:25.224Z"
    },
    {
      "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": 0.8670031666666665,
      "accuracy_std_dev": 0.004661340237051534,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07032303872053872,
      "avg_duration_per_task_ms": 224493.27861952863,
      "avg_output_tokens_per_task": 11452.973905723906,
      "last_run_timestamp": "2026-09-26T09:37:35.352Z"
    },
    {
      "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": 0.8667788887015176,
      "accuracy_std_dev": 0.01707605497635669,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.08546577121978641,
      "avg_duration_per_task_ms": 213324.25801011804,
      "avg_output_tokens_per_task": 23606.13940415964,
      "last_run_timestamp": "2026-08-29T16:55:12.673Z"
    },
    {
      "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": 0.8667228333333332,
      "accuracy_std_dev": 0.006024375774486715,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0256143728956229,
      "avg_duration_per_task_ms": 228795.63636363635,
      "avg_output_tokens_per_task": 21105.421717171717,
      "last_run_timestamp": "2026-09-19T09:03:56.948Z"
    },
    {
      "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": 0.8659932,
      "accuracy_std_dev": 0.008010237759267841,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07810327272727272,
      "avg_duration_per_task_ms": 87737.49797979798,
      "avg_output_tokens_per_task": 12770.258585858586,
      "last_run_timestamp": "2026-09-26T08:55:49.839Z"
    },
    {
      "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": 0.8658808333333333,
      "accuracy_std_dev": 0.0104532872886422,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07000543097643097,
      "avg_duration_per_task_ms": 223270.64393939395,
      "avg_output_tokens_per_task": 11398.105218855218,
      "last_run_timestamp": "2026-09-26T08:58:06.175Z"
    },
    {
      "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": 0.865758552529183,
      "accuracy_std_dev": 0.008634405610694918,
      "run_count": 8,
      "total_questions": 1542,
      "avg_cost_per_task_usd": 0.008104218547341117,
      "avg_duration_per_task_ms": 281362.64915693906,
      "avg_output_tokens_per_task": 30092.439688715953,
      "last_run_timestamp": "2026-08-29T19:20:21.472Z"
    },
    {
      "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": 0.8656778723404256,
      "accuracy_std_dev": 0.011228325910392875,
      "run_count": 8,
      "total_questions": 1551,
      "avg_cost_per_task_usd": 0.008557784010315926,
      "avg_duration_per_task_ms": 288449.6731141199,
      "avg_output_tokens_per_task": 30252.780786589297,
      "last_run_timestamp": "2026-08-29T18:42:44.868Z"
    },
    {
      "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": 0.86532,
      "accuracy_std_dev": 0.008780061799326916,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.02537659898989899,
      "avg_duration_per_task_ms": 218871.01414141414,
      "avg_output_tokens_per_task": 20892.29595959596,
      "last_run_timestamp": "2026-09-19T08:55:39.417Z"
    },
    {
      "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": 0.8653196666666666,
      "accuracy_std_dev": 0.010378007269649058,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.029422861952861954,
      "avg_duration_per_task_ms": 503984.98484848486,
      "avg_output_tokens_per_task": 34386.4898989899,
      "last_run_timestamp": "2026-08-08T09:24:32.269Z"
    },
    {
      "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": 0.8651804917197452,
      "accuracy_std_dev": 0.013354329136557154,
      "run_count": 4,
      "total_questions": 785,
      "avg_cost_per_task_usd": 0.060597563057324846,
      "avg_duration_per_task_ms": 205391.9694267516,
      "avg_output_tokens_per_task": 23879.336305732482,
      "last_run_timestamp": "2026-09-26T09:20:06.107Z"
    },
    {
      "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": 0.8650391666666667,
      "accuracy_std_dev": 0.011558638146665698,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.03327992845117845,
      "avg_duration_per_task_ms": 670922.4377104377,
      "avg_output_tokens_per_task": 34339.50252525252,
      "last_run_timestamp": "2026-08-29T19:08:19.007Z"
    },
    {
      "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": 0.8649764115566037,
      "accuracy_std_dev": 0.05470934774460394,
      "run_count": 5,
      "total_questions": 848,
      "avg_cost_per_task_usd": 0.08730664976415095,
      "avg_duration_per_task_ms": 645412.391509434,
      "avg_output_tokens_per_task": 26303.754716981133,
      "last_run_timestamp": "2026-08-22T09:51:24.721Z"
    },
    {
      "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": 0.8649570000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.020998656410256413,
      "avg_duration_per_task_ms": 619238.4615384615,
      "avg_output_tokens_per_task": 23774.34358974359,
      "last_run_timestamp": "2026-09-26T01:32:20.726Z"
    },
    {
      "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": 0.8648248729281768,
      "accuracy_std_dev": 0.02631682070007696,
      "run_count": 5,
      "total_questions": 905,
      "avg_cost_per_task_usd": 0.010246941436464088,
      "avg_duration_per_task_ms": 355027.6861878453,
      "avg_output_tokens_per_task": 21157.914917127073,
      "last_run_timestamp": "2026-09-26T09:56:32.081Z"
    },
    {
      "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": 0.8648077338403042,
      "accuracy_std_dev": 0.0058285997192721924,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.03674314195183777,
      "avg_duration_per_task_ms": 350368.4626108999,
      "avg_output_tokens_per_task": 23592.097591888465,
      "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-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Fireworks",
      "accuracy": 0.8646862059405941,
      "accuracy_std_dev": 0.024035029940114947,
      "run_count": 3,
      "total_questions": 505,
      "avg_cost_per_task_usd": 0.10486535445544554,
      "avg_duration_per_task_ms": 438191.62970297027,
      "avg_output_tokens_per_task": 30917.40594059406,
      "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": "ModelRun",
      "accuracy": 0.8646021311475409,
      "accuracy_std_dev": 0.007616834002538237,
      "run_count": 3,
      "total_questions": 549,
      "avg_cost_per_task_usd": 0.1932159398907104,
      "avg_duration_per_task_ms": 529434.2877959927,
      "avg_output_tokens_per_task": 56744.85792349727,
      "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-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Moonshot AI",
      "accuracy": 0.864345941772152,
      "accuracy_std_dev": 0.03040778172991867,
      "run_count": 8,
      "total_questions": 1580,
      "avg_cost_per_task_usd": 0.12872155822784812,
      "avg_duration_per_task_ms": 705035.3753164557,
      "avg_output_tokens_per_task": 43078.59620253165,
      "last_run_timestamp": "2026-08-29T18:25:54.650Z"
    },
    {
      "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": 0.8639189175738725,
      "accuracy_std_dev": 0.02340391523335358,
      "run_count": 4,
      "total_questions": 643,
      "avg_cost_per_task_usd": 0.06298315552099534,
      "avg_duration_per_task_ms": 414941.54743390356,
      "avg_output_tokens_per_task": 21263.947122861588,
      "last_run_timestamp": "2026-08-15T08:47:51.369Z"
    },
    {
      "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": 0.863768366310743,
      "accuracy_std_dev": 0.013824474714642947,
      "run_count": 27,
      "total_questions": 4589,
      "avg_cost_per_task_usd": 0.16332582131183263,
      "avg_duration_per_task_ms": 245950.8069296143,
      "avg_output_tokens_per_task": 29748.09130529527,
      "last_run_timestamp": "2026-09-26T08:54:22.466Z"
    },
    {
      "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": 0.8636842273684211,
      "accuracy_std_dev": 0.01763815223202249,
      "run_count": 5,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.08584259578947369,
      "avg_duration_per_task_ms": 395986.4378947368,
      "avg_output_tokens_per_task": 24656.30947368421,
      "last_run_timestamp": "2026-09-26T08:57:34.809Z"
    },
    {
      "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": 0.8636519239465571,
      "accuracy_std_dev": 0.028707041857903168,
      "run_count": 11,
      "total_questions": 1946,
      "avg_cost_per_task_usd": 0.010248814491264132,
      "avg_duration_per_task_ms": 324581.0220966084,
      "avg_output_tokens_per_task": 18963.51336073998,
      "last_run_timestamp": "2026-09-26T08:55:27.375Z"
    },
    {
      "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": 0.863636375,
      "accuracy_std_dev": 0.017230143926687758,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.07376672032828283,
      "avg_duration_per_task_ms": 300093.57133838383,
      "avg_output_tokens_per_task": 20992.814393939392,
      "last_run_timestamp": "2026-08-29T18:39:14.500Z"
    },
    {
      "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": 0.8636360000000001,
      "accuracy_std_dev": 0.006733999999999962,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13014110606060605,
      "avg_duration_per_task_ms": 586191.3131313132,
      "avg_output_tokens_per_task": 37366.416666666664,
      "last_run_timestamp": "2026-08-29T19:12:34.610Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Novita",
      "accuracy": 0.8633267645181878,
      "accuracy_std_dev": 0.013466913424667874,
      "run_count": 8,
      "total_questions": 1567,
      "avg_cost_per_task_usd": 0.07518679770261646,
      "avg_duration_per_task_ms": 298763.8781110402,
      "avg_output_tokens_per_task": 20760.19719208679,
      "last_run_timestamp": "2026-08-29T18:30:43.161Z"
    },
    {
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Inceptron",
      "accuracy": 0.8631252833052278,
      "accuracy_std_dev": 0.03388611321929843,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.09767507644744237,
      "avg_duration_per_task_ms": 708453.107925801,
      "avg_output_tokens_per_task": 37470.564362001125,
      "last_run_timestamp": "2026-09-21T15:33:37.909Z"
    },
    {
      "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": 0.8631102531752752,
      "accuracy_std_dev": 0.01428710704815903,
      "run_count": 6,
      "total_questions": 1181,
      "avg_cost_per_task_usd": 0.006545317527519051,
      "avg_duration_per_task_ms": 246133.70618120237,
      "avg_output_tokens_per_task": 12855.507197290432,
      "last_run_timestamp": "2026-09-26T09:18:13.150Z"
    },
    {
      "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": 0.8628256666666666,
      "accuracy_std_dev": 0.019902461007624172,
      "run_count": 5,
      "total_questions": 972,
      "avg_cost_per_task_usd": 0.08342864197530865,
      "avg_duration_per_task_ms": 454814.01543209876,
      "avg_output_tokens_per_task": 23652.057613168723,
      "last_run_timestamp": "2026-09-26T08:53:09.429Z"
    },
    {
      "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": 0.8627945,
      "accuracy_std_dev": 0.033022030634865585,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0750204577020202,
      "avg_duration_per_task_ms": 314213.95328282827,
      "avg_output_tokens_per_task": 21335.801767676767,
      "last_run_timestamp": "2026-08-29T20:00:02.637Z"
    },
    {
      "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": 0.8627943749999999,
      "accuracy_std_dev": 0.00929733078546604,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.855192389520202,
      "avg_duration_per_task_ms": 483915.25,
      "avg_output_tokens_per_task": 34017.0119949495,
      "last_run_timestamp": "2026-09-26T08:53:16.951Z"
    },
    {
      "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": 0.8626742857142858,
      "accuracy_std_dev": 0.017858490299448887,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.08792987914862915,
      "avg_duration_per_task_ms": 406224.28066378064,
      "avg_output_tokens_per_task": 39434.920634920636,
      "last_run_timestamp": "2026-09-26T08:56:57.800Z"
    },
    {
      "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": 0.862584125,
      "accuracy_std_dev": 0.009408632345318578,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.8442858333333334,
      "avg_duration_per_task_ms": 415897.4154040404,
      "avg_output_tokens_per_task": 33580.344696969696,
      "last_run_timestamp": "2026-09-26T08:53:36.382Z"
    },
    {
      "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": "BaseTen",
      "accuracy": 0.8622473714739068,
      "accuracy_std_dev": 0.021847945469408447,
      "run_count": 29,
      "total_questions": 5672,
      "avg_cost_per_task_usd": 0.2541298106488012,
      "avg_duration_per_task_ms": 337436.6962270804,
      "avg_output_tokens_per_task": 46060.357369534555,
      "last_run_timestamp": "2026-09-26T11:00:26.140Z"
    },
    {
      "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": 0.8620605995542349,
      "accuracy_std_dev": 0.04244629212767299,
      "run_count": 7,
      "total_questions": 1346,
      "avg_cost_per_task_usd": 0.05972067161961367,
      "avg_duration_per_task_ms": 704231.3573551263,
      "avg_output_tokens_per_task": 21477.753343239227,
      "last_run_timestamp": "2026-09-12T09:19:57.645Z"
    },
    {
      "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": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.8619525,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.223144601010101,
      "avg_duration_per_task_ms": 273332.86363636365,
      "avg_output_tokens_per_task": 22220.545454545456,
      "last_run_timestamp": "2026-09-01T13:23:04.034Z"
    },
    {
      "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": 0.86184890094552,
      "accuracy_std_dev": 0.17338949639000253,
      "run_count": 13,
      "total_questions": 2221,
      "avg_cost_per_task_usd": 0.14356039261593875,
      "avg_duration_per_task_ms": 275384.9054479964,
      "avg_output_tokens_per_task": 11018.233678523187,
      "last_run_timestamp": "2026-09-29T21:35:11.276Z"
    },
    {
      "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": 0.8613916666666667,
      "accuracy_std_dev": 0.019552374609977406,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.15697531565656564,
      "avg_duration_per_task_ms": 1157244.2163299664,
      "avg_output_tokens_per_task": 59089.26262626263,
      "last_run_timestamp": "2026-08-29T19:01:27.418Z"
    },
    {
      "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": "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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Minimax",
      "accuracy": 0.8611646092843325,
      "accuracy_std_dev": 0.009461303461336583,
      "run_count": 8,
      "total_questions": 1551,
      "avg_cost_per_task_usd": 0.05262209671179884,
      "avg_duration_per_task_ms": 438586.79239200515,
      "avg_output_tokens_per_task": 27289.920696324953,
      "last_run_timestamp": "2026-08-01T17:44:01.464Z"
    },
    {
      "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": 0.8609557877094972,
      "accuracy_std_dev": 0.0034322869279178035,
      "run_count": 3,
      "total_questions": 537,
      "avg_cost_per_task_usd": 0.005924823091247672,
      "avg_duration_per_task_ms": 435495.54189944133,
      "avg_output_tokens_per_task": 16657.258845437616,
      "last_run_timestamp": "2026-07-18T08:04:02.357Z"
    },
    {
      "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": 0.8605498333333333,
      "accuracy_std_dev": 0.020281775936512293,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0053592163299663285,
      "avg_duration_per_task_ms": 748446.3863636364,
      "avg_output_tokens_per_task": 29643.154882154882,
      "last_run_timestamp": "2026-08-29T19:07:25.263Z"
    },
    {
      "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": 0.8605498333333332,
      "accuracy_std_dev": 0.014065440607112956,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.16507093434343434,
      "avg_duration_per_task_ms": 1149871.6085858585,
      "avg_output_tokens_per_task": 53779.61784511784,
      "last_run_timestamp": "2026-08-29T18:43:04.938Z"
    },
    {
      "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": 0.8605496666666668,
      "accuracy_std_dev": 0.016734196123175105,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.11623640740740741,
      "avg_duration_per_task_ms": 564457.3611111111,
      "avg_output_tokens_per_task": 34704.39057239057,
      "last_run_timestamp": "2026-08-29T18:39:19.202Z"
    },
    {
      "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": 0.8603745,
      "accuracy_std_dev": 0.008427361219563313,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.8821099195075758,
      "avg_duration_per_task_ms": 416796.10890151514,
      "avg_output_tokens_per_task": 33415.74179292929,
      "last_run_timestamp": "2026-09-26T09:19:04.990Z"
    },
    {
      "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": 0.860269,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.20759804545454547,
      "avg_duration_per_task_ms": 271144.0252525252,
      "avg_output_tokens_per_task": 20665.12626262626,
      "last_run_timestamp": "2026-09-01T13:31:11.391Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Together",
      "accuracy": 0.8599929535603715,
      "accuracy_std_dev": 0.009491817962856216,
      "run_count": 5,
      "total_questions": 969,
      "avg_cost_per_task_usd": 0.1207670278637771,
      "avg_duration_per_task_ms": 352384.2745098039,
      "avg_output_tokens_per_task": 30417.953560371516,
      "last_run_timestamp": "2026-08-22T10:19:36.962Z"
    },
    {
      "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": "Novita",
      "accuracy": 0.8599885,
      "accuracy_std_dev": 0.01684674659738986,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.19415970033670032,
      "avg_duration_per_task_ms": 1192894.914983165,
      "avg_output_tokens_per_task": 57724.79882154882,
      "last_run_timestamp": "2026-08-29T20:16:17.828Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Morph",
      "accuracy": 0.859829,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.07579302051282051,
      "avg_duration_per_task_ms": 503262.1435897436,
      "avg_output_tokens_per_task": 21548.302564102563,
      "last_run_timestamp": "2026-07-04T18:14:56.425Z"
    },
    {
      "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": 0.859638,
      "accuracy_std_dev": 0.008071103007024502,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.815575909090909,
      "avg_duration_per_task_ms": 465568.93876262626,
      "avg_output_tokens_per_task": 32451.6672979798,
      "last_run_timestamp": "2026-09-26T09:01:02.717Z"
    },
    {
      "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": "GMICloud",
      "accuracy": 0.8594275,
      "accuracy_std_dev": 0.013424092492604471,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0069255740740740744,
      "avg_duration_per_task_ms": 366969.99410774413,
      "avg_output_tokens_per_task": 23129.368686868685,
      "last_run_timestamp": "2026-09-26T09:34:13.855Z"
    },
    {
      "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": 0.8593380638297874,
      "accuracy_std_dev": 0.008570441036233525,
      "run_count": 3,
      "total_questions": 564,
      "avg_cost_per_task_usd": 0.25169353191489363,
      "avg_duration_per_task_ms": 300098.05673758866,
      "avg_output_tokens_per_task": 62735.48936170213,
      "last_run_timestamp": "2026-07-18T08:26:40.483Z"
    },
    {
      "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": 0.8593072857142857,
      "accuracy_std_dev": 0.014195632454227251,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.05442466955266955,
      "avg_duration_per_task_ms": 720332.65007215,
      "avg_output_tokens_per_task": 32949.00865800866,
      "last_run_timestamp": "2026-08-29T19:37:04.387Z"
    },
    {
      "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": 0.8591573843017328,
      "accuracy_std_dev": 0.02115249701193692,
      "run_count": 5,
      "total_questions": 981,
      "avg_cost_per_task_usd": 0.10401832415902142,
      "avg_duration_per_task_ms": 479500.45667686034,
      "avg_output_tokens_per_task": 30593.96636085627,
      "last_run_timestamp": "2026-08-22T16:49:15.579Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.8590881190053286,
      "accuracy_std_dev": 0.02052541308201858,
      "run_count": 4,
      "total_questions": 563,
      "avg_cost_per_task_usd": 0.08836215808170515,
      "avg_duration_per_task_ms": 517852.0088809947,
      "avg_output_tokens_per_task": 22138.96802841918,
      "last_run_timestamp": "2026-08-15T08:47:41.523Z"
    },
    {
      "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": 0.8589315283806344,
      "accuracy_std_dev": 0.01816958311832932,
      "run_count": 7,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.15454068447412356,
      "avg_duration_per_task_ms": 551935.939899833,
      "avg_output_tokens_per_task": 51245.50918196995,
      "last_run_timestamp": "2026-08-29T19:17:01.480Z"
    },
    {
      "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": 0.8587961250000001,
      "accuracy_std_dev": 0.026570942984948328,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0075656887626262635,
      "avg_duration_per_task_ms": 497586.5126262626,
      "avg_output_tokens_per_task": 24566.76767676768,
      "last_run_timestamp": "2026-09-26T08:53:09.295Z"
    },
    {
      "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": 0.8580246666666665,
      "accuracy_std_dev": 0.007694244595078101,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.06644984511784513,
      "avg_duration_per_task_ms": 213157.5361952862,
      "avg_output_tokens_per_task": 10806.75925925926,
      "last_run_timestamp": "2026-09-26T09:18:24.826Z"
    },
    {
      "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": 0.857954375,
      "accuracy_std_dev": 0.013038726633930733,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.7458450347222222,
      "avg_duration_per_task_ms": 493231.4678030303,
      "avg_output_tokens_per_task": 29674.352904040403,
      "last_run_timestamp": "2026-09-26T09:34:52.748Z"
    },
    {
      "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": 0.8577779119999999,
      "accuracy_std_dev": 0.006748318809155334,
      "run_count": 4,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.23260101733333335,
      "avg_duration_per_task_ms": 433991.3,
      "avg_output_tokens_per_task": 51935.09333333333,
      "last_run_timestamp": "2026-08-15T08:57:45.830Z"
    },
    {
      "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": 0.8577439999999998,
      "accuracy_std_dev": 0.005891999999999953,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.008277098484848486,
      "avg_duration_per_task_ms": 710306.3914141414,
      "avg_output_tokens_per_task": 29437.88888888889,
      "last_run_timestamp": "2026-07-18T08:30:06.336Z"
    },
    {
      "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": 0.8577028343558282,
      "accuracy_std_dev": 0.011904490003355866,
      "run_count": 5,
      "total_questions": 978,
      "avg_cost_per_task_usd": 0.14577651738241312,
      "avg_duration_per_task_ms": 258434.56646216768,
      "avg_output_tokens_per_task": 28881.36707566462,
      "last_run_timestamp": "2026-09-26T09:21:12.329Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Decart",
      "accuracy": 0.857191527607362,
      "accuracy_std_dev": 0.013500948849617942,
      "run_count": 5,
      "total_questions": 978,
      "avg_cost_per_task_usd": 0.1052768057259714,
      "avg_duration_per_task_ms": 186949.86605316974,
      "avg_output_tokens_per_task": 27345.398773006134,
      "last_run_timestamp": "2026-09-26T08:51:46.229Z"
    },
    {
      "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": 0.8571024481012658,
      "accuracy_std_dev": 0.017386228014021777,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.05838445316455696,
      "avg_duration_per_task_ms": 511596.7248945148,
      "avg_output_tokens_per_task": 33795.75696202532,
      "last_run_timestamp": "2026-08-29T19:11:30.664Z"
    },
    {
      "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": 0.8569513442340791,
      "accuracy_std_dev": 0.013334663992859697,
      "run_count": 9,
      "total_questions": 1743,
      "avg_cost_per_task_usd": 0.10127325989672976,
      "avg_duration_per_task_ms": 639996.5054503729,
      "avg_output_tokens_per_task": 35441.2759609868,
      "last_run_timestamp": "2026-09-19T08:55:10.654Z"
    },
    {
      "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": 0.8569022000000001,
      "accuracy_std_dev": 0.011711955283384562,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.005455330303030303,
      "avg_duration_per_task_ms": 645344.1272727273,
      "avg_output_tokens_per_task": 28522.870707070706,
      "last_run_timestamp": "2026-08-22T17:16:03.168Z"
    },
    {
      "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": 0.8569022,
      "accuracy_std_dev": 0.010212534228094395,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.8410627575757575,
      "avg_duration_per_task_ms": 450697.24494949495,
      "avg_output_tokens_per_task": 33467.82424242424,
      "last_run_timestamp": "2026-09-26T09:29:09.285Z"
    },
    {
      "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": 0.8566216666666666,
      "accuracy_std_dev": 0.010073872139792559,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.004407090909090909,
      "avg_duration_per_task_ms": 239041.3745791246,
      "avg_output_tokens_per_task": 27919.25,
      "last_run_timestamp": "2026-08-29T19:40:39.914Z"
    },
    {
      "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": 0.8565655999999999,
      "accuracy_std_dev": 0.01503514097838793,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.09365999898989899,
      "avg_duration_per_task_ms": 671877.2212121212,
      "avg_output_tokens_per_task": 28686.372727272726,
      "last_run_timestamp": "2026-09-26T08:52:43.449Z"
    },
    {
      "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": 0.8565324139194139,
      "accuracy_std_dev": 0.018712538162953735,
      "run_count": 4,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.00399579304029304,
      "avg_duration_per_task_ms": 408743.60622710624,
      "avg_output_tokens_per_task": 7814.793040293041,
      "last_run_timestamp": "2026-09-28T08:34:14.152Z"
    },
    {
      "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": 0.8564577918781726,
      "accuracy_std_dev": 0.028181512947320644,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.10150419204737732,
      "avg_duration_per_task_ms": 330925.20050761424,
      "avg_output_tokens_per_task": 22836.06937394247,
      "last_run_timestamp": "2026-09-26T09:19:52.896Z"
    },
    {
      "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": 0.8563503216174183,
      "accuracy_std_dev": 0.016757180106276432,
      "run_count": 17,
      "total_questions": 3215,
      "avg_cost_per_task_usd": 0.1523061856920684,
      "avg_duration_per_task_ms": 200594.1632970451,
      "avg_output_tokens_per_task": 27265.382270606533,
      "last_run_timestamp": "2026-09-26T09:17:01.108Z"
    },
    {
      "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": 0.8563168223350255,
      "accuracy_std_dev": 0.012593349140849445,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.01789011252115059,
      "avg_duration_per_task_ms": 343652.3730964467,
      "avg_output_tokens_per_task": 35569.74111675127,
      "last_run_timestamp": "2026-09-26T08:50:11.879Z"
    },
    {
      "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": 0.8562288,
      "accuracy_std_dev": 0.011497073373689477,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1913848090909091,
      "avg_duration_per_task_ms": 350329.27575757576,
      "avg_output_tokens_per_task": 55974.5,
      "last_run_timestamp": "2026-08-29T18:20:49.176Z"
    },
    {
      "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": 0.8560649439728354,
      "accuracy_std_dev": 0.042572315121043564,
      "run_count": 9,
      "total_questions": 1767,
      "avg_cost_per_task_usd": 0.08027640350877194,
      "avg_duration_per_task_ms": 491193.9134125637,
      "avg_output_tokens_per_task": 21784.452744765138,
      "last_run_timestamp": "2026-09-26T08:55:14.254Z"
    },
    {
      "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": 0.8557868225629791,
      "accuracy_std_dev": 0.008479443730551645,
      "run_count": 14,
      "total_questions": 2739,
      "avg_cost_per_task_usd": 0.061858265790434463,
      "avg_duration_per_task_ms": 609941.4859437752,
      "avg_output_tokens_per_task": 34015.95436290617,
      "last_run_timestamp": "2026-08-29T19:16:17.710Z"
    },
    {
      "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": 0.8555556,
      "accuracy_std_dev": 0.0176375124050984,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.011364495959595959,
      "avg_duration_per_task_ms": 652022.0676767677,
      "avg_output_tokens_per_task": 33080.43131313131,
      "last_run_timestamp": "2026-09-26T08:57:03.952Z"
    },
    {
      "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": 0.8550504,
      "accuracy_std_dev": 0.04995446213743072,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.07905246363636365,
      "avg_duration_per_task_ms": 503288.9419191919,
      "avg_output_tokens_per_task": 37401.228282828284,
      "last_run_timestamp": "2026-09-26T09:14:27.364Z"
    },
    {
      "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": 0.8549529134157944,
      "accuracy_std_dev": 0.0077363797752788375,
      "run_count": 16,
      "total_questions": 3153,
      "avg_cost_per_task_usd": 0.05448097875039644,
      "avg_duration_per_task_ms": 435614.29495718365,
      "avg_output_tokens_per_task": 27906.339993656835,
      "last_run_timestamp": "2026-08-29T19:17:08.288Z"
    },
    {
      "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": 0.8547375714285715,
      "accuracy_std_dev": 0.04301245273458486,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.1766843759018759,
      "avg_duration_per_task_ms": 45692.21067821068,
      "avg_output_tokens_per_task": 3336.1709956709956,
      "last_run_timestamp": "2026-09-26T09:24:09.411Z"
    },
    {
      "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": 0.8546573333333333,
      "accuracy_std_dev": 0.015141104458467416,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.007946281144781145,
      "avg_duration_per_task_ms": 637289.4747474748,
      "avg_output_tokens_per_task": 29143.16835016835,
      "last_run_timestamp": "2026-08-29T18:38:15.907Z"
    },
    {
      "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": 0.8543893879003559,
      "accuracy_std_dev": 0.010879822884587815,
      "run_count": 3,
      "total_questions": 562,
      "avg_cost_per_task_usd": 0.0749505462633452,
      "avg_duration_per_task_ms": 1305986.9822064056,
      "avg_output_tokens_per_task": 38419.777580071175,
      "last_run_timestamp": "2026-07-18T16:22:39.277Z"
    },
    {
      "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": 0.8541664999999999,
      "accuracy_std_dev": 0.020077776601257404,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.8391522569444444,
      "avg_duration_per_task_ms": 453772.9583333333,
      "avg_output_tokens_per_task": 33398.183712121216,
      "last_run_timestamp": "2026-09-26T09:34:32.494Z"
    },
    {
      "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": 0.8535522308340354,
      "accuracy_std_dev": 0.010038567733441294,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.01690193850042123,
      "avg_duration_per_task_ms": 472592.7733782645,
      "avg_output_tokens_per_task": 33567.315080033695,
      "last_run_timestamp": "2026-09-26T08:54:58.400Z"
    },
    {
      "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": 0.8529230909090909,
      "accuracy_std_dev": 0.020659392015407452,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.09120310835629016,
      "avg_duration_per_task_ms": 630647.5684113866,
      "avg_output_tokens_per_task": 41423.92011019284,
      "last_run_timestamp": "2026-09-26T08:59:56.037Z"
    },
    {
      "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": 0.8526935,
      "accuracy_std_dev": 0.0025254999999999583,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20698726515151514,
      "avg_duration_per_task_ms": 257611.93181818182,
      "avg_output_tokens_per_task": 19625.295454545456,
      "last_run_timestamp": "2026-09-01T13:50:27.830Z"
    },
    {
      "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": "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": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Azure",
      "accuracy": 0.851851625,
      "accuracy_std_dev": 0.018151435989595288,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.7516495170454545,
      "avg_duration_per_task_ms": 500493.72222222225,
      "avg_output_tokens_per_task": 29958.29797979798,
      "last_run_timestamp": "2026-09-26T08:55:10.633Z"
    },
    {
      "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": 0.851679139925373,
      "accuracy_std_dev": 0.03687109443675745,
      "run_count": 28,
      "total_questions": 4824,
      "avg_cost_per_task_usd": 0.011843935737976785,
      "avg_duration_per_task_ms": 411338.9653814262,
      "avg_output_tokens_per_task": 16150.16169154229,
      "last_run_timestamp": "2026-09-26T08:52:37.124Z"
    },
    {
      "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": 0.8516111428571428,
      "accuracy_std_dev": 0.032073947329028685,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.831446111111111,
      "avg_duration_per_task_ms": 427691.9718614719,
      "avg_output_tokens_per_task": 33200.131313131315,
      "last_run_timestamp": "2026-09-26T08:56:21.997Z"
    },
    {
      "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": 0.8512395454545454,
      "accuracy_std_dev": 0.008130117793663436,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.07382550229568412,
      "avg_duration_per_task_ms": 230546.61202938476,
      "avg_output_tokens_per_task": 22967.3406795225,
      "last_run_timestamp": "2026-09-26T08:53:37.648Z"
    },
    {
      "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": 0.8511034444444444,
      "accuracy_std_dev": 0.03586305000387234,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.17666739506172838,
      "avg_duration_per_task_ms": 44619.677328843994,
      "avg_output_tokens_per_task": 3328.8120089786757,
      "last_run_timestamp": "2026-09-26T08:53:27.502Z"
    },
    {
      "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": 0.8510597058823527,
      "accuracy_std_dev": 0.02313439249450838,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.1586523764111705,
      "avg_duration_per_task_ms": 153997.80986333927,
      "avg_output_tokens_per_task": 15572.684789067142,
      "last_run_timestamp": "2026-09-26T09:21:26.479Z"
    },
    {
      "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": 0.8510099000000001,
      "accuracy_std_dev": 0.015100002764569275,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.7538109368686869,
      "avg_duration_per_task_ms": 501864.8787878788,
      "avg_output_tokens_per_task": 30068.36717171717,
      "last_run_timestamp": "2026-09-26T09:17:18.349Z"
    },
    {
      "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": 0.8508604502868068,
      "accuracy_std_dev": 0.02478573563844512,
      "run_count": 6,
      "total_questions": 1046,
      "avg_cost_per_task_usd": 0.010149271510516252,
      "avg_duration_per_task_ms": 344062.4971319312,
      "avg_output_tokens_per_task": 20091.341300191205,
      "last_run_timestamp": "2026-09-26T09:16:37.664Z"
    },
    {
      "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": 0.8508416000000001,
      "accuracy_std_dev": 0.01139804777319343,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.10227179090909091,
      "avg_duration_per_task_ms": 614412.1656565657,
      "avg_output_tokens_per_task": 32654.38888888889,
      "last_run_timestamp": "2026-08-29T19:18:11.839Z"
    },
    {
      "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": 0.8506943125000002,
      "accuracy_std_dev": 0.0170931360994653,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.8004648680555555,
      "avg_duration_per_task_ms": 479005.81407828286,
      "avg_output_tokens_per_task": 30341.69191919192,
      "last_run_timestamp": "2026-09-26T09:19:41.155Z"
    },
    {
      "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": "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": 0.8505651288135594,
      "accuracy_std_dev": 0.014603362238425309,
      "run_count": 6,
      "total_questions": 1180,
      "avg_cost_per_task_usd": 0.008425746610169493,
      "avg_duration_per_task_ms": 368412.6118644068,
      "avg_output_tokens_per_task": 24250.365254237287,
      "last_run_timestamp": "2026-09-26T09:22:08.527Z"
    },
    {
      "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": 0.8504212593103447,
      "accuracy_std_dev": 0.008642258916137237,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.05126995816091954,
      "avg_duration_per_task_ms": 291139.3857471264,
      "avg_output_tokens_per_task": 45945.63908045977,
      "last_run_timestamp": "2026-09-26T16:05:38.016Z"
    },
    {
      "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": 0.8503937480314961,
      "accuracy_std_dev": 0.00023499999999998522,
      "run_count": 2,
      "total_questions": 381,
      "avg_cost_per_task_usd": 0.023863905511811022,
      "avg_duration_per_task_ms": 137111.656167979,
      "avg_output_tokens_per_task": 36528.412073490814,
      "last_run_timestamp": "2026-09-26T09:52:38.111Z"
    },
    {
      "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": 0.8503154279475983,
      "accuracy_std_dev": 0.03140494390885563,
      "run_count": 7,
      "total_questions": 1374,
      "avg_cost_per_task_usd": 0.09674545997088793,
      "avg_duration_per_task_ms": 1349515.0895196507,
      "avg_output_tokens_per_task": 43276.89374090247,
      "last_run_timestamp": "2026-08-29T18:36:10.818Z"
    },
    {
      "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": 0.8501715195876289,
      "accuracy_std_dev": 0.027779548602524146,
      "run_count": 5,
      "total_questions": 970,
      "avg_cost_per_task_usd": 0.13975330515463918,
      "avg_duration_per_task_ms": 383469.5556701031,
      "avg_output_tokens_per_task": 31875.5793814433,
      "last_run_timestamp": "2026-09-26T08:53:10.326Z"
    },
    {
      "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": 0.8501683333333333,
      "accuracy_std_dev": 0.041790840030375614,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.17284996857463525,
      "avg_duration_per_task_ms": 43624.98316498317,
      "avg_output_tokens_per_task": 3253.4483726150393,
      "last_run_timestamp": "2026-09-26T08:54:57.321Z"
    },
    {
      "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": 0.8500861979274612,
      "accuracy_std_dev": 0.01857196094762209,
      "run_count": 5,
      "total_questions": 965,
      "avg_cost_per_task_usd": 0.016829795854922277,
      "avg_duration_per_task_ms": 389325.4870466321,
      "avg_output_tokens_per_task": 22513.535751295338,
      "last_run_timestamp": "2026-08-29T18:07:13.805Z"
    },
    {
      "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": 0.849820223628692,
      "accuracy_std_dev": 0.010339301856316641,
      "run_count": 11,
      "total_questions": 2133,
      "avg_cost_per_task_usd": 0.027238617440225037,
      "avg_duration_per_task_ms": 658650.0131270511,
      "avg_output_tokens_per_task": 31066.365682137835,
      "last_run_timestamp": "2026-09-26T09:22:28.862Z"
    },
    {
      "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": 0.8498050435706695,
      "accuracy_std_dev": 0.03351477910773096,
      "run_count": 5,
      "total_questions": 941,
      "avg_cost_per_task_usd": 0.11714972582359193,
      "avg_duration_per_task_ms": 505045.9829968119,
      "avg_output_tokens_per_task": 33189.51115834219,
      "last_run_timestamp": "2026-09-26T08:54:17.442Z"
    },
    {
      "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": 0.8496956203301478,
      "accuracy_std_dev": 0.012943756573044104,
      "run_count": 6,
      "total_questions": 1151,
      "avg_cost_per_task_usd": 0.076488348392702,
      "avg_duration_per_task_ms": 490389.1841876629,
      "avg_output_tokens_per_task": 33522.794960903564,
      "last_run_timestamp": "2026-08-22T17:03:07.750Z"
    },
    {
      "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": 0.8495983231939164,
      "accuracy_std_dev": 0.010674015891406563,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.09633560709759188,
      "avg_duration_per_task_ms": 314235.76045627374,
      "avg_output_tokens_per_task": 30867.523447401774,
      "last_run_timestamp": "2026-09-26T09:23:32.894Z"
    },
    {
      "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": 0.8493265000000001,
      "accuracy_std_dev": 0.01119888692013629,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.026091900252525255,
      "avg_duration_per_task_ms": 290649.8787878788,
      "avg_output_tokens_per_task": 21576.02525252525,
      "last_run_timestamp": "2026-08-01T17:57:21.740Z"
    },
    {
      "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": 0.8492335655172413,
      "accuracy_std_dev": 0.026170202058877375,
      "run_count": 9,
      "total_questions": 1740,
      "avg_cost_per_task_usd": 0.029375192528735626,
      "avg_duration_per_task_ms": 542113.9201149425,
      "avg_output_tokens_per_task": 28046.42528735632,
      "last_run_timestamp": "2026-09-26T08:57:03.866Z"
    },
    {
      "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": 0.8492264101265824,
      "accuracy_std_dev": 0.04905497536664576,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.1081006405063291,
      "avg_duration_per_task_ms": 612102.1265822785,
      "avg_output_tokens_per_task": 27354.124050632912,
      "last_run_timestamp": "2026-08-29T11:31:24.094Z"
    },
    {
      "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": 0.8490153107221006,
      "accuracy_std_dev": 0.02104665580299624,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.04150975711159738,
      "avg_duration_per_task_ms": 669606.0547045952,
      "avg_output_tokens_per_task": 34749.842450765864,
      "last_run_timestamp": "2026-08-29T19:20:22.044Z"
    },
    {
      "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": 0.8486918037542662,
      "accuracy_std_dev": 0.014893804491204496,
      "run_count": 6,
      "total_questions": 1172,
      "avg_cost_per_task_usd": 0.017789553754266208,
      "avg_duration_per_task_ms": 299607.09129692835,
      "avg_output_tokens_per_task": 35353.55375426621,
      "last_run_timestamp": "2026-09-26T08:54:42.351Z"
    },
    {
      "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": 0.848485,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.20520298484848487,
      "avg_duration_per_task_ms": 278139.5202020202,
      "avg_output_tokens_per_task": 20422.565656565657,
      "last_run_timestamp": "2026-09-01T13:34:32.291Z"
    },
    {
      "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": 0.8482611842239186,
      "accuracy_std_dev": 0.011728653128556577,
      "run_count": 10,
      "total_questions": 1965,
      "avg_cost_per_task_usd": 0.12833305750636134,
      "avg_duration_per_task_ms": 389662.5424936387,
      "avg_output_tokens_per_task": 33100.7262086514,
      "last_run_timestamp": "2026-09-26T08:55:48.508Z"
    },
    {
      "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": 0.8482548437025798,
      "accuracy_std_dev": 0.010506007044067693,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.029977341932220538,
      "avg_duration_per_task_ms": 657961.3520485584,
      "avg_output_tokens_per_task": 31303.588770864946,
      "last_run_timestamp": "2026-09-26T09:15:06.166Z"
    },
    {
      "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": 0.8481760229007633,
      "accuracy_std_dev": 0.021743000000000012,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.025479938931297706,
      "avg_duration_per_task_ms": 435380.475826972,
      "avg_output_tokens_per_task": 21459.994910941477,
      "last_run_timestamp": "2026-07-18T16:18:27.151Z"
    },
    {
      "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": 0.8480639,
      "accuracy_std_dev": 0.00907569944907828,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.04004892272727273,
      "avg_duration_per_task_ms": 395495.8308080808,
      "avg_output_tokens_per_task": 40373.71717171717,
      "last_run_timestamp": "2026-09-26T16:12:42.605Z"
    },
    {
      "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": 0.8480638333333334,
      "accuracy_std_dev": 0.0534225021796891,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.06181613468013468,
      "avg_duration_per_task_ms": 279974.0715488215,
      "avg_output_tokens_per_task": 20518.97474747475,
      "last_run_timestamp": "2026-08-15T15:55:02.752Z"
    },
    {
      "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": 0.847643,
      "accuracy_std_dev": 0.01430999999999999,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02130802777777778,
      "avg_duration_per_task_ms": 487650.29545454547,
      "avg_output_tokens_per_task": 42398.98737373737,
      "last_run_timestamp": "2026-09-26T09:15:51.836Z"
    },
    {
      "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": 0.8475419999999999,
      "accuracy_std_dev": 0.02235490172109912,
      "run_count": 25,
      "total_questions": 4950,
      "avg_cost_per_task_usd": 0.16931346101010097,
      "avg_duration_per_task_ms": 157193.26565656567,
      "avg_output_tokens_per_task": 15870.87696969697,
      "last_run_timestamp": "2026-09-26T09:21:17.407Z"
    },
    {
      "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": 0.8473202765017668,
      "accuracy_std_dev": 0.011288322955111134,
      "run_count": 6,
      "total_questions": 1132,
      "avg_cost_per_task_usd": 0.016132280918727912,
      "avg_duration_per_task_ms": 230171.71908127208,
      "avg_output_tokens_per_task": 32052.750883392226,
      "last_run_timestamp": "2026-09-26T08:49:42.815Z"
    },
    {
      "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": 0.8472522501081783,
      "accuracy_std_dev": 0.022478090167458387,
      "run_count": 12,
      "total_questions": 2311,
      "avg_cost_per_task_usd": 0.153898237126785,
      "avg_duration_per_task_ms": 298616.1008221549,
      "avg_output_tokens_per_task": 40013.343574210296,
      "last_run_timestamp": "2026-09-26T08:55:26.766Z"
    },
    {
      "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": 0.8468012500000001,
      "accuracy_std_dev": 0.031201721542047957,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.11716699116161616,
      "avg_duration_per_task_ms": 561187.7866161616,
      "avg_output_tokens_per_task": 31215.503787878788,
      "last_run_timestamp": "2026-08-29T18:26:17.965Z"
    },
    {
      "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": 0.8466330000000001,
      "accuracy_std_dev": 0.02123421248363121,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.13107612676767677,
      "avg_duration_per_task_ms": 434259.98131313134,
      "avg_output_tokens_per_task": 34303.56616161616,
      "last_run_timestamp": "2026-09-26T10:36:35.700Z"
    },
    {
      "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": 0.8462764299380979,
      "accuracy_std_dev": 0.015770782183096275,
      "run_count": 9,
      "total_questions": 1777,
      "avg_cost_per_task_usd": 0.08913706978052896,
      "avg_duration_per_task_ms": 416673.65053460887,
      "avg_output_tokens_per_task": 22175.449071468767,
      "last_run_timestamp": "2026-09-26T09:15:30.970Z"
    },
    {
      "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": 0.8461668398357288,
      "accuracy_std_dev": 0.06450182954877243,
      "run_count": 6,
      "total_questions": 974,
      "avg_cost_per_task_usd": 0.13265561088295688,
      "avg_duration_per_task_ms": 1458439.1293634498,
      "avg_output_tokens_per_task": 46081.89835728953,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Makora",
      "accuracy": 0.8460472371705964,
      "accuracy_std_dev": 0.020750268207603947,
      "run_count": 4,
      "total_questions": 721,
      "avg_cost_per_task_usd": 0.1417408987517337,
      "avg_duration_per_task_ms": 223845.73370319002,
      "avg_output_tokens_per_task": 32811.28294036061,
      "last_run_timestamp": "2026-09-26T09:01:35.352Z"
    },
    {
      "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": 0.8458658888888889,
      "accuracy_std_dev": 0.03972899765618852,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.17558623344556676,
      "avg_duration_per_task_ms": 45322.60044893378,
      "avg_output_tokens_per_task": 3305.652076318743,
      "last_run_timestamp": "2026-09-26T08:54:44.118Z"
    },
    {
      "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": 0.8458657777777778,
      "accuracy_std_dev": 0.025409284277723633,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.15791502805836138,
      "avg_duration_per_task_ms": 153529.50112233445,
      "avg_output_tokens_per_task": 15577.850168350169,
      "last_run_timestamp": "2026-09-26T09:19:06.330Z"
    },
    {
      "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": 0.8452106827586208,
      "accuracy_std_dev": 0.035124267859903004,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.054803951724137934,
      "avg_duration_per_task_ms": 227710.32091954022,
      "avg_output_tokens_per_task": 22733.708505747127,
      "last_run_timestamp": "2026-09-26T09:01:34.657Z"
    },
    {
      "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": 0.8451892474226804,
      "accuracy_std_dev": 0.006483228158872687,
      "run_count": 5,
      "total_questions": 970,
      "avg_cost_per_task_usd": 0.007323164948453608,
      "avg_duration_per_task_ms": 502972.8824742268,
      "avg_output_tokens_per_task": 18045.067010309278,
      "last_run_timestamp": "2026-08-29T18:17:30.072Z"
    },
    {
      "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": 0.8451776040609137,
      "accuracy_std_dev": 0.007926079064287278,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.020253781725888324,
      "avg_duration_per_task_ms": 578832.0761421319,
      "avg_output_tokens_per_task": 22875.73434856176,
      "last_run_timestamp": "2026-08-15T14:57:08.773Z"
    },
    {
      "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": 0.8451178333333332,
      "accuracy_std_dev": 0.062380510090403164,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.10489853282828282,
      "avg_duration_per_task_ms": 490349.9595959596,
      "avg_output_tokens_per_task": 26821.65404040404,
      "last_run_timestamp": "2026-09-26T08:57:17.745Z"
    },
    {
      "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": 0.8451176666666667,
      "accuracy_std_dev": 0.014740528220597193,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1084811388888889,
      "avg_duration_per_task_ms": 330556.0925925926,
      "avg_output_tokens_per_task": 28808.55387205387,
      "last_run_timestamp": "2026-09-26T09:15:10.787Z"
    },
    {
      "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": 0.8450335999999999,
      "accuracy_std_dev": 0.011207895897089693,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.10169760353535354,
      "avg_duration_per_task_ms": 588854.2631313131,
      "avg_output_tokens_per_task": 32291.069191919192,
      "last_run_timestamp": "2026-09-26T10:00:58.971Z"
    },
    {
      "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": 0.8447463261390887,
      "accuracy_std_dev": 0.02515683625839815,
      "run_count": 19,
      "total_questions": 3753,
      "avg_cost_per_task_usd": 0.23025201278976826,
      "avg_duration_per_task_ms": 509807.78284039436,
      "avg_output_tokens_per_task": 43633.612310151875,
      "last_run_timestamp": "2026-09-26T08:55:16.971Z"
    },
    {
      "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": 0.8446432684563759,
      "accuracy_std_dev": 0.0059905795195115885,
      "run_count": 7,
      "total_questions": 1341,
      "avg_cost_per_task_usd": 0.007240005219985085,
      "avg_duration_per_task_ms": 381452.95600298286,
      "avg_output_tokens_per_task": 17843.34004474273,
      "last_run_timestamp": "2026-08-29T18:39:38.845Z"
    },
    {
      "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": "DekaLLM",
      "accuracy": 0.844444,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 180,
      "avg_cost_per_task_usd": 0.08250182777777779,
      "avg_duration_per_task_ms": 453463.44444444444,
      "avg_output_tokens_per_task": 29087.933333333334,
      "last_run_timestamp": "2026-07-04T09:58:40.318Z"
    },
    {
      "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": 0.8443695555555555,
      "accuracy_std_dev": 0.00866138720107314,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.14565384680134677,
      "avg_duration_per_task_ms": 171228.07014590348,
      "avg_output_tokens_per_task": 32219.744668911335,
      "last_run_timestamp": "2026-09-26T09:17:52.892Z"
    },
    {
      "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": 0.8443149751647239,
      "accuracy_std_dev": 0.019658458576399102,
      "run_count": 10,
      "total_questions": 1973,
      "avg_cost_per_task_usd": 0.2023946046629498,
      "avg_duration_per_task_ms": 318245.6705524582,
      "avg_output_tokens_per_task": 45936.32032437912,
      "last_run_timestamp": "2026-09-26T09:23:36.793Z"
    },
    {
      "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": 0.844276,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.059444419191919194,
      "avg_duration_per_task_ms": 215120.31818181818,
      "avg_output_tokens_per_task": 34746.81060606061,
      "last_run_timestamp": "2026-09-26T10:57:32.238Z"
    },
    {
      "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": 0.844202597826087,
      "accuracy_std_dev": 0.014526749979936743,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.03675822753623188,
      "avg_duration_per_task_ms": 642776.863768116,
      "avg_output_tokens_per_task": 34394.93550724638,
      "last_run_timestamp": "2026-08-29T19:21:17.134Z"
    },
    {
      "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": 0.8440308723559445,
      "accuracy_std_dev": 0.010942287133606307,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.006902234865061998,
      "avg_duration_per_task_ms": 486147.99270605395,
      "avg_output_tokens_per_task": 17097.730123997084,
      "last_run_timestamp": "2026-08-29T19:18:32.746Z"
    },
    {
      "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": 0.8439305549132948,
      "accuracy_std_dev": 0.013499116239012633,
      "run_count": 8,
      "total_questions": 1557,
      "avg_cost_per_task_usd": 0.08022249261400129,
      "avg_duration_per_task_ms": 891006.2575465639,
      "avg_output_tokens_per_task": 40664.11753371869,
      "last_run_timestamp": "2026-08-30T00:19:02.095Z"
    },
    {
      "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": 0.8436028,
      "accuracy_std_dev": 0.0247016256015672,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.06619451666666666,
      "avg_duration_per_task_ms": 657610.2742424243,
      "avg_output_tokens_per_task": 49788.666666666664,
      "last_run_timestamp": "2026-09-26T09:04:08.258Z"
    },
    {
      "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": 0.8431370045248868,
      "accuracy_std_dev": 0.004260933614280586,
      "run_count": 7,
      "total_questions": 1326,
      "avg_cost_per_task_usd": 0.039398003770739064,
      "avg_duration_per_task_ms": 50754.730015082954,
      "avg_output_tokens_per_task": 14781.932126696833,
      "last_run_timestamp": "2026-09-01T17:50:42.331Z"
    },
    {
      "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": 0.8431245977011494,
      "accuracy_std_dev": 0.011756871278107985,
      "run_count": 8,
      "total_questions": 1566,
      "avg_cost_per_task_usd": 0.027553961047254147,
      "avg_duration_per_task_ms": 530629.3461047254,
      "avg_output_tokens_per_task": 22775.175606641125,
      "last_run_timestamp": "2026-08-29T18:16:39.840Z"
    },
    {
      "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": 0.8428064354527938,
      "accuracy_std_dev": 0.014271656998128247,
      "run_count": 6,
      "total_questions": 1038,
      "avg_cost_per_task_usd": 0.006940633911368016,
      "avg_duration_per_task_ms": 644934.6493256262,
      "avg_output_tokens_per_task": 17139.102119460502,
      "last_run_timestamp": "2026-08-29T18:15:38.703Z"
    },
    {
      "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": 0.842761,
      "accuracy_std_dev": 0.013366603308245514,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.11932980505050504,
      "avg_duration_per_task_ms": 317023.23434343433,
      "avg_output_tokens_per_task": 26929.522222222222,
      "last_run_timestamp": "2026-09-26T08:52:53.172Z"
    },
    {
      "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": 0.8427608000000001,
      "accuracy_std_dev": 0.011595230077924271,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.08432281212121212,
      "avg_duration_per_task_ms": 322252.27373737376,
      "avg_output_tokens_per_task": 20908.12626262626,
      "last_run_timestamp": "2026-09-26T08:55:48.606Z"
    },
    {
      "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": 0.8427092264988897,
      "accuracy_std_dev": 0.008653421154287425,
      "run_count": 7,
      "total_questions": 1351,
      "avg_cost_per_task_usd": 0.026629270910436716,
      "avg_duration_per_task_ms": 68698.27461139897,
      "avg_output_tokens_per_task": 18885.883049592892,
      "last_run_timestamp": "2026-09-29T21:03:10.646Z"
    },
    {
      "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": 0.8425925,
      "accuracy_std_dev": 0.005892500000000023,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.06301305808080808,
      "avg_duration_per_task_ms": 90351.56313131313,
      "avg_output_tokens_per_task": 13876.954545454546,
      "last_run_timestamp": "2026-09-01T17:51:50.923Z"
    },
    {
      "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": 0.8425925,
      "accuracy_std_dev": 0.02455548471828647,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.06700256376262627,
      "avg_duration_per_task_ms": 555614.4097222222,
      "avg_output_tokens_per_task": 21893.804924242424,
      "last_run_timestamp": "2026-08-29T19:19:13.521Z"
    },
    {
      "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": 0.8425925,
      "accuracy_std_dev": 0.008788056170166437,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.14231570202020202,
      "avg_duration_per_task_ms": 163901.45454545456,
      "avg_output_tokens_per_task": 32169.752525252527,
      "last_run_timestamp": "2026-09-26T09:14:46.960Z"
    },
    {
      "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": 0.8420405496688741,
      "accuracy_std_dev": 0.03337997922071716,
      "run_count": 7,
      "total_questions": 1359,
      "avg_cost_per_task_usd": 0.03366557321559971,
      "avg_duration_per_task_ms": 468305.63208241353,
      "avg_output_tokens_per_task": 31573.827814569537,
      "last_run_timestamp": "2026-08-29T19:03:08.917Z"
    },
    {
      "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": 0.8418721404174574,
      "accuracy_std_dev": 0.020746850670267405,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.025684514231499055,
      "avg_duration_per_task_ms": 264287.531941809,
      "avg_output_tokens_per_task": 21340.85009487666,
      "last_run_timestamp": "2026-08-29T18:12:43.943Z"
    },
    {
      "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": 0.8417577430497052,
      "accuracy_std_dev": 0.014641061103015125,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.13158069755686608,
      "avg_duration_per_task_ms": 164554.15669755687,
      "avg_output_tokens_per_task": 29682.097725358046,
      "last_run_timestamp": "2026-09-26T08:53:13.623Z"
    },
    {
      "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": 0.8417004740890687,
      "accuracy_std_dev": 0.014854076686525622,
      "run_count": 14,
      "total_questions": 2470,
      "avg_cost_per_task_usd": 0.01521865870445344,
      "avg_duration_per_task_ms": 573091.9230769231,
      "avg_output_tokens_per_task": 16302.382995951417,
      "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-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Relace",
      "accuracy": 0.8415826000000001,
      "accuracy_std_dev": 0.01677086684223569,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.013069406060606059,
      "avg_duration_per_task_ms": 474851.301010101,
      "avg_output_tokens_per_task": 38641.41818181818,
      "last_run_timestamp": "2026-09-26T08:53:10.669Z"
    },
    {
      "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": 0.8410445402985075,
      "accuracy_std_dev": 0.013784355106942281,
      "run_count": 7,
      "total_questions": 1340,
      "avg_cost_per_task_usd": 0.21741457462686567,
      "avg_duration_per_task_ms": 421390.1029850746,
      "avg_output_tokens_per_task": 49267.26119402985,
      "last_run_timestamp": "2026-08-29T19:00:02.941Z"
    },
    {
      "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": 0.8410292857142857,
      "accuracy_std_dev": 0.03771800444544936,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.6859342640692642,
      "avg_duration_per_task_ms": 470200.2972582973,
      "avg_output_tokens_per_task": 27535.334054834057,
      "last_run_timestamp": "2026-09-26T08:58:47.240Z"
    },
    {
      "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": 0.8409129024016351,
      "accuracy_std_dev": 0.025338336264245934,
      "run_count": 10,
      "total_questions": 1957,
      "avg_cost_per_task_usd": 0.20147738068472154,
      "avg_duration_per_task_ms": 293735.5861011753,
      "avg_output_tokens_per_task": 45634.63157894737,
      "last_run_timestamp": "2026-09-26T08:54:40.874Z"
    },
    {
      "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": 0.8408546584615385,
      "accuracy_std_dev": 0.004112945970955614,
      "run_count": 5,
      "total_questions": 975,
      "avg_cost_per_task_usd": 0.09125773538461539,
      "avg_duration_per_task_ms": 289113.2892307692,
      "avg_output_tokens_per_task": 22386.701538461537,
      "last_run_timestamp": "2026-09-19T08:59:39.935Z"
    },
    {
      "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": 0.840661254601227,
      "accuracy_std_dev": 0.04601154819825125,
      "run_count": 5,
      "total_questions": 978,
      "avg_cost_per_task_usd": 0.16520423415132923,
      "avg_duration_per_task_ms": 1050780.9069529653,
      "avg_output_tokens_per_task": 49997.56237218814,
      "last_run_timestamp": "2026-08-29T19:03:21.016Z"
    },
    {
      "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": 0.840404,
      "accuracy_std_dev": 0.008869764145680555,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07076540101010101,
      "avg_duration_per_task_ms": 297317.25757575757,
      "avg_output_tokens_per_task": 20594.954545454544,
      "last_run_timestamp": "2026-09-26T08:53:05.194Z"
    },
    {
      "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": 0.8394122720364743,
      "accuracy_std_dev": 0.0596591189780741,
      "run_count": 10,
      "total_questions": 1974,
      "avg_cost_per_task_usd": 0.06978041185410336,
      "avg_duration_per_task_ms": 685532.4701114488,
      "avg_output_tokens_per_task": 44128.422998986825,
      "last_run_timestamp": "2026-09-26T09:41:13.780Z"
    },
    {
      "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": 0.8393938763636365,
      "accuracy_std_dev": 0.012496125008266452,
      "run_count": 3,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.03367001454545455,
      "avg_duration_per_task_ms": 269728.67454545456,
      "avg_output_tokens_per_task": 28518.758181818182,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.8392968267419962,
      "accuracy_std_dev": 0.0037396706405778625,
      "run_count": 4,
      "total_questions": 531,
      "avg_cost_per_task_usd": 0.12392348210922788,
      "avg_duration_per_task_ms": 100276.54613935969,
      "avg_output_tokens_per_task": 27948.913370998118,
      "last_run_timestamp": "2026-09-26T09:05:37.095Z"
    },
    {
      "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": 0.8388888000000001,
      "accuracy_std_dev": 0.07596307646192327,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.1830352515151515,
      "avg_duration_per_task_ms": 532300.757070707,
      "avg_output_tokens_per_task": 41655.45656565657,
      "last_run_timestamp": "2026-09-26T08:52:39.171Z"
    },
    {
      "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": 0.8388092148272959,
      "accuracy_std_dev": 0.039447312816872845,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.16212616849199665,
      "avg_duration_per_task_ms": 1421951.0867733783,
      "avg_output_tokens_per_task": 51701.65206402696,
      "last_run_timestamp": "2026-08-29T18:13:11.491Z"
    },
    {
      "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": 0.8387204,
      "accuracy_std_dev": 0.02117180060930105,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.020740932323232323,
      "avg_duration_per_task_ms": 437294.63434343436,
      "avg_output_tokens_per_task": 37324.31616161616,
      "last_run_timestamp": "2026-09-26T09:10:17.414Z"
    },
    {
      "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": 0.838579522693997,
      "accuracy_std_dev": 0.017426082982771814,
      "run_count": 7,
      "total_questions": 1366,
      "avg_cost_per_task_usd": 0.0794436691068814,
      "avg_duration_per_task_ms": 602046.7005856516,
      "avg_output_tokens_per_task": 29605.27232796486,
      "last_run_timestamp": "2026-09-26T09:19:10.825Z"
    },
    {
      "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": 0.8385383078393881,
      "accuracy_std_dev": 0.038297652145521575,
      "run_count": 8,
      "total_questions": 1569,
      "avg_cost_per_task_usd": 0.10716400509878904,
      "avg_duration_per_task_ms": 1170088.323773104,
      "avg_output_tokens_per_task": 44102.23709369025,
      "last_run_timestamp": "2026-08-29T18:42:20.541Z"
    },
    {
      "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": 0.838284705882353,
      "accuracy_std_dev": 0.05170697825901454,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.18276167468805704,
      "avg_duration_per_task_ms": 40754.59982174688,
      "avg_output_tokens_per_task": 3290.764111705288,
      "last_run_timestamp": "2026-09-26T08:53:17.875Z"
    },
    {
      "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": 0.8381625397574123,
      "accuracy_std_dev": 0.007901272276048144,
      "run_count": 9,
      "total_questions": 1484,
      "avg_cost_per_task_usd": 0.017624041105121297,
      "avg_duration_per_task_ms": 189126.76886792452,
      "avg_output_tokens_per_task": 35057.77156334232,
      "last_run_timestamp": "2026-09-26T09:29:08.510Z"
    },
    {
      "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": 0.8380671171597633,
      "accuracy_std_dev": 0.02773812208207325,
      "run_count": 5,
      "total_questions": 845,
      "avg_cost_per_task_usd": 0.07336154792899409,
      "avg_duration_per_task_ms": 417593.0852071006,
      "avg_output_tokens_per_task": 24105.62721893491,
      "last_run_timestamp": "2026-09-26T08:54:14.754Z"
    },
    {
      "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": 0.8380537102396514,
      "accuracy_std_dev": 0.018206946326350625,
      "run_count": 7,
      "total_questions": 1377,
      "avg_cost_per_task_usd": 0.04750454974582426,
      "avg_duration_per_task_ms": 652669.2650689905,
      "avg_output_tokens_per_task": 37622.816993464054,
      "last_run_timestamp": "2026-08-29T19:06:57.766Z"
    },
    {
      "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": 0.8377145373291274,
      "accuracy_std_dev": 0.08364258942907019,
      "run_count": 5,
      "total_questions": 951,
      "avg_cost_per_task_usd": 0.09799871188222924,
      "avg_duration_per_task_ms": 610509.8033648791,
      "avg_output_tokens_per_task": 25161.19242902208,
      "last_run_timestamp": "2026-09-26T08:53:09.416Z"
    },
    {
      "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": 0.8376334722439405,
      "accuracy_std_dev": 0.027219006084059767,
      "run_count": 7,
      "total_questions": 1279,
      "avg_cost_per_task_usd": 0.0889748506645817,
      "avg_duration_per_task_ms": 1571235.6520719312,
      "avg_output_tokens_per_task": 41303.13604378421,
      "last_run_timestamp": "2026-08-29T19:12:33.941Z"
    },
    {
      "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": 0.8376065934065935,
      "accuracy_std_dev": 0.00571412028808169,
      "run_count": 3,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.036677783882783886,
      "avg_duration_per_task_ms": 47951.65201465201,
      "avg_output_tokens_per_task": 14583.362637362638,
      "last_run_timestamp": "2026-09-01T17:50:52.743Z"
    },
    {
      "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": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "DeepInfra",
      "accuracy": 0.8371357823193916,
      "accuracy_std_dev": 0.009460368146759528,
      "run_count": 6,
      "total_questions": 1052,
      "avg_cost_per_task_usd": 0.08486342680608365,
      "avg_duration_per_task_ms": 599922.150190114,
      "avg_output_tokens_per_task": 21002.96958174905,
      "last_run_timestamp": "2026-09-19T08:48:08.824Z"
    },
    {
      "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": 0.8371241903959562,
      "accuracy_std_dev": 0.03104266294105938,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.17545385846672285,
      "avg_duration_per_task_ms": 1006738.7118786857,
      "avg_output_tokens_per_task": 53398.904802021905,
      "last_run_timestamp": "2026-08-29T18:43:32.988Z"
    },
    {
      "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": 0.8369565347826087,
      "accuracy_std_dev": 0.019774448331529436,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.07724686594202897,
      "avg_duration_per_task_ms": 754357.3224637681,
      "avg_output_tokens_per_task": 41683.36666666667,
      "last_run_timestamp": "2026-08-22T17:01:35.600Z"
    },
    {
      "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": 0.8367000000000001,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.06203523232323233,
      "avg_duration_per_task_ms": 92697.9696969697,
      "avg_output_tokens_per_task": 13658.787878787878,
      "last_run_timestamp": "2026-09-01T17:49:43.168Z"
    },
    {
      "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": 0.8366999999999999,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04958680555555556,
      "avg_duration_per_task_ms": 536787.6388888889,
      "avg_output_tokens_per_task": 49485.26515151515,
      "last_run_timestamp": "2026-09-26T16:26:15.411Z"
    },
    {
      "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": 0.8366909151079136,
      "accuracy_std_dev": 0.01869795941406443,
      "run_count": 4,
      "total_questions": 695,
      "avg_cost_per_task_usd": 0.020156113669064748,
      "avg_duration_per_task_ms": 242800.49064748202,
      "avg_output_tokens_per_task": 40096.50647482014,
      "last_run_timestamp": "2026-09-26T09:48:09.064Z"
    },
    {
      "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": 0.8366301736263737,
      "accuracy_std_dev": 0.012674357340709624,
      "run_count": 3,
      "total_questions": 455,
      "avg_cost_per_task_usd": 0.04995332967032967,
      "avg_duration_per_task_ms": 107459.2945054945,
      "avg_output_tokens_per_task": 15090.074725274726,
      "last_run_timestamp": "2026-09-26T08:53:14.372Z"
    },
    {
      "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": 0.8365793063670413,
      "accuracy_std_dev": 0.023337994672784947,
      "run_count": 7,
      "total_questions": 1335,
      "avg_cost_per_task_usd": 0.01897062172284644,
      "avg_duration_per_task_ms": 793025.5857677902,
      "avg_output_tokens_per_task": 41115.47715355805,
      "last_run_timestamp": "2026-09-26T12:41:27.111Z"
    },
    {
      "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": 0.83627925,
      "accuracy_std_dev": 0.037120416716242575,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.015161977272727275,
      "avg_duration_per_task_ms": 425327.9482323232,
      "avg_output_tokens_per_task": 27500.13005050505,
      "last_run_timestamp": "2026-08-29T19:58:27.585Z"
    },
    {
      "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": 0.836234166760247,
      "accuracy_std_dev": 0.02982784551020275,
      "run_count": 27,
      "total_questions": 5343,
      "avg_cost_per_task_usd": 0.1657105358412877,
      "avg_duration_per_task_ms": 150009.68556990454,
      "avg_output_tokens_per_task": 15316.003930376193,
      "last_run_timestamp": "2026-09-26T08:59:40.295Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Parasail",
      "accuracy": 0.8359790095238095,
      "accuracy_std_dev": 0.03660393871642171,
      "run_count": 6,
      "total_questions": 630,
      "avg_cost_per_task_usd": 0.10253282222222224,
      "avg_duration_per_task_ms": 301607.7301587302,
      "avg_output_tokens_per_task": 23860.25873015873,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Novita",
      "accuracy": 0.8358585000000001,
      "accuracy_std_dev": 0.02753853053904659,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1297035244107744,
      "avg_duration_per_task_ms": 386511.07744107745,
      "avg_output_tokens_per_task": 35313.07323232323,
      "last_run_timestamp": "2026-09-26T08:54:10.966Z"
    },
    {
      "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": 0.8355473623188407,
      "accuracy_std_dev": 0.012606082718439233,
      "run_count": 11,
      "total_questions": 1656,
      "avg_cost_per_task_usd": 0.03472509842995169,
      "avg_duration_per_task_ms": 236602.51086956522,
      "avg_output_tokens_per_task": 27679.197463768116,
      "last_run_timestamp": "2026-09-26T16:00:25.681Z"
    },
    {
      "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": 0.835238659717552,
      "accuracy_std_dev": 0.014890026205318157,
      "run_count": 9,
      "total_questions": 1487,
      "avg_cost_per_task_usd": 0.09666733691997309,
      "avg_duration_per_task_ms": 626516.3597848016,
      "avg_output_tokens_per_task": 30133.80094149294,
      "last_run_timestamp": "2026-09-26T09:56:28.417Z"
    },
    {
      "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": 0.8350426615384616,
      "accuracy_std_dev": 0.022313590127081184,
      "run_count": 8,
      "total_questions": 1560,
      "avg_cost_per_task_usd": 0.08725940705128205,
      "avg_duration_per_task_ms": 939709.0711538461,
      "avg_output_tokens_per_task": 40061.0217948718,
      "last_run_timestamp": "2026-08-30T00:28:09.702Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openrouter/auto-beta",
      "model_slug": "openrouter/auto-beta",
      "model_name": "Auto Router (Beta)",
      "provider_name": "auto-routing",
      "accuracy": 0.835017,
      "accuracy_std_dev": 0.062145249638783075,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.025753547138047138,
      "avg_duration_per_task_ms": 248297.88383838383,
      "avg_output_tokens_per_task": 30752.395622895623,
      "last_run_timestamp": "2026-09-15T14:04:40.150Z"
    },
    {
      "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": 0.835017,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.036737896464646466,
      "avg_duration_per_task_ms": 49536.31060606061,
      "avg_output_tokens_per_task": 14606.348484848484,
      "last_run_timestamp": "2026-09-01T17:49:40.120Z"
    },
    {
      "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": 0.8348017092511013,
      "accuracy_std_dev": 0.013128455115293883,
      "run_count": 7,
      "total_questions": 1362,
      "avg_cost_per_task_usd": 0.03501912334801763,
      "avg_duration_per_task_ms": 566392.0389133627,
      "avg_output_tokens_per_task": 32664.345814977973,
      "last_run_timestamp": "2026-08-29T18:47:45.892Z"
    },
    {
      "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": 0.8345117,
      "accuracy_std_dev": 0.03065694363777969,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.21535924595959594,
      "avg_duration_per_task_ms": 332148.0095959596,
      "avg_output_tokens_per_task": 48826.94343434343,
      "last_run_timestamp": "2026-09-26T08:56:02.276Z"
    },
    {
      "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": 0.8344516577181208,
      "accuracy_std_dev": 0.048501499999999975,
      "run_count": 2,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.0903846644295302,
      "avg_duration_per_task_ms": 304392.4194630872,
      "avg_output_tokens_per_task": 22866.939597315435,
      "last_run_timestamp": "2026-07-18T08:27:05.304Z"
    },
    {
      "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": 0.8343404728033473,
      "accuracy_std_dev": 0.00859056480406806,
      "run_count": 11,
      "total_questions": 2151,
      "avg_cost_per_task_usd": 0.07748330729893073,
      "avg_duration_per_task_ms": 794488.6968851697,
      "avg_output_tokens_per_task": 46585.74012087401,
      "last_run_timestamp": "2026-09-26T16:00:20.168Z"
    },
    {
      "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/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.834175,
      "accuracy_std_dev": 0.009259000000000017,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.028012333333333334,
      "avg_duration_per_task_ms": 352518.654040404,
      "avg_output_tokens_per_task": 14875.09595959596,
      "last_run_timestamp": "2026-09-26T09:19:39.685Z"
    },
    {
      "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": 0.8340065999999998,
      "accuracy_std_dev": 0.032567889465545674,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.12823596313131314,
      "avg_duration_per_task_ms": 558007.0247474747,
      "avg_output_tokens_per_task": 38277.854545454546,
      "last_run_timestamp": "2026-09-26T08:50:56.054Z"
    },
    {
      "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": 0.8339505777777776,
      "accuracy_std_dev": 0.01071049161258874,
      "run_count": 6,
      "total_questions": 1080,
      "avg_cost_per_task_usd": 0.025328656481481478,
      "avg_duration_per_task_ms": 650579.8324074075,
      "avg_output_tokens_per_task": 40296.40185185185,
      "last_run_timestamp": "2026-09-26T09:00:45.770Z"
    },
    {
      "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-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Venice",
      "accuracy": 0.8338271333333335,
      "accuracy_std_dev": 0.012377564032868325,
      "run_count": 7,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.007238588888888889,
      "avg_duration_per_task_ms": 324197.44074074074,
      "avg_output_tokens_per_task": 26128.143703703703,
      "last_run_timestamp": "2026-08-29T18:49:32.713Z"
    },
    {
      "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": 0.8337061610738257,
      "accuracy_std_dev": 0.0693376467642218,
      "run_count": 5,
      "total_questions": 894,
      "avg_cost_per_task_usd": 0.009961087248322148,
      "avg_duration_per_task_ms": 513095.80536912754,
      "avg_output_tokens_per_task": 19701.401565995526,
      "last_run_timestamp": "2026-09-26T08:54:24.468Z"
    },
    {
      "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": 0.8334047490347491,
      "accuracy_std_dev": 0.03381421535995953,
      "run_count": 12,
      "total_questions": 2331,
      "avg_cost_per_task_usd": 0.056030425568425565,
      "avg_duration_per_task_ms": 279504.7528957529,
      "avg_output_tokens_per_task": 24216.713856713857,
      "last_run_timestamp": "2026-09-26T10:00:05.608Z"
    },
    {
      "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": 0.8333332499999999,
      "accuracy_std_dev": 0.01666596387664092,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.024300950126262627,
      "avg_duration_per_task_ms": 440107.09532828286,
      "avg_output_tokens_per_task": 22549.840909090908,
      "last_run_timestamp": "2026-08-29T18:09:03.379Z"
    },
    {
      "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": 0.8333332105263158,
      "accuracy_std_dev": 0.01599350000000005,
      "run_count": 2,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.16854262573099413,
      "avg_duration_per_task_ms": 685525.8274853801,
      "avg_output_tokens_per_task": 47357.02046783626,
      "last_run_timestamp": "2026-07-18T08:29:37.313Z"
    },
    {
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Parasail",
      "accuracy": 0.833159912591051,
      "accuracy_std_dev": 0.01602299589215448,
      "run_count": 5,
      "total_questions": 961,
      "avg_cost_per_task_usd": 0.14537621123829345,
      "avg_duration_per_task_ms": 277832.8116545265,
      "avg_output_tokens_per_task": 32850.81373569199,
      "last_run_timestamp": "2026-09-26T09:19:06.323Z"
    },
    {
      "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": 0.8331351764705882,
      "accuracy_std_dev": 0.011719105633588751,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.25533701039809864,
      "avg_duration_per_task_ms": 277560.3027332145,
      "avg_output_tokens_per_task": 55800.19964349376,
      "last_run_timestamp": "2026-09-26T09:16:57.240Z"
    },
    {
      "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": 0.8327431107738998,
      "accuracy_std_dev": 0.03352054924908005,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.1104804410723318,
      "avg_duration_per_task_ms": 986804.3333333334,
      "avg_output_tokens_per_task": 34519.57764289327,
      "last_run_timestamp": "2026-09-26T10:24:21.190Z"
    },
    {
      "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": 0.8326118571428571,
      "accuracy_std_dev": 0.010792989132217443,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.039859138528138535,
      "avg_duration_per_task_ms": 221347.44805194804,
      "avg_output_tokens_per_task": 14514.689033189034,
      "last_run_timestamp": "2026-09-26T09:19:17.245Z"
    },
    {
      "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-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "StreamLake",
      "accuracy": 0.8323770956022944,
      "accuracy_std_dev": 0.018917767858550327,
      "run_count": 8,
      "total_questions": 1569,
      "avg_cost_per_task_usd": 0.07184262970044615,
      "avg_duration_per_task_ms": 878259.2613129382,
      "avg_output_tokens_per_task": 40890.33397068196,
      "last_run_timestamp": "2026-08-30T00:37:59.094Z"
    },
    {
      "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": 0.8323511666666668,
      "accuracy_std_dev": 0.013383536583139077,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1359044436026936,
      "avg_duration_per_task_ms": 151136.74915824915,
      "avg_output_tokens_per_task": 30646.316498316497,
      "last_run_timestamp": "2026-09-26T08:54:57.237Z"
    },
    {
      "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": 0.8319833976435935,
      "accuracy_std_dev": 0.0051619355292529005,
      "run_count": 7,
      "total_questions": 1358,
      "avg_cost_per_task_usd": 0.03598950073637703,
      "avg_duration_per_task_ms": 200380.98085419735,
      "avg_output_tokens_per_task": 13861.594256259204,
      "last_run_timestamp": "2026-09-26T09:19:46.321Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.8315411499051233,
      "accuracy_std_dev": 0.028738766047579353,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.02477825616698292,
      "avg_duration_per_task_ms": 479102.12713472487,
      "avg_output_tokens_per_task": 23444.680581910183,
      "last_run_timestamp": "2026-08-29T18:05:36.579Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Baidu",
      "accuracy": 0.8311784671163575,
      "accuracy_std_dev": 0.009516052855602353,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.007403454749859472,
      "avg_duration_per_task_ms": 398416.0635188308,
      "avg_output_tokens_per_task": 19072.874648679033,
      "last_run_timestamp": "2026-08-29T16:40:58.367Z"
    },
    {
      "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": 0.8309764,
      "accuracy_std_dev": 0.005061551248382267,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.24307877272727274,
      "avg_duration_per_task_ms": 341071.7717171717,
      "avg_output_tokens_per_task": 40681.778787878786,
      "last_run_timestamp": "2026-09-26T08:54:33.880Z"
    },
    {
      "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": 0.8309340194750211,
      "accuracy_std_dev": 0.0227049955198312,
      "run_count": 6,
      "total_questions": 1181,
      "avg_cost_per_task_usd": 0.0975369652836579,
      "avg_duration_per_task_ms": 214278.25232853513,
      "avg_output_tokens_per_task": 28093.950889077052,
      "last_run_timestamp": "2026-08-29T18:58:14.477Z"
    },
    {
      "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": 0.8308604629080119,
      "accuracy_std_dev": 0.014193547884550253,
      "run_count": 7,
      "total_questions": 1348,
      "avg_cost_per_task_usd": 0.037891371661721074,
      "avg_duration_per_task_ms": 428497.83827893174,
      "avg_output_tokens_per_task": 44654.94213649852,
      "last_run_timestamp": "2026-09-26T08:56:01.204Z"
    },
    {
      "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": 0.830808,
      "accuracy_std_dev": 0.008551060402078817,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.09822325252525253,
      "avg_duration_per_task_ms": 297130.1747474747,
      "avg_output_tokens_per_task": 24088.215151515153,
      "last_run_timestamp": "2026-09-19T08:48:09.416Z"
    },
    {
      "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": 0.8306482182585346,
      "accuracy_std_dev": 0.018880885612038913,
      "run_count": 14,
      "total_questions": 2607,
      "avg_cost_per_task_usd": 0.18192265515918682,
      "avg_duration_per_task_ms": 363453.91829689295,
      "avg_output_tokens_per_task": 37006.87917146145,
      "last_run_timestamp": "2026-09-26T08:52:36.362Z"
    },
    {
      "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": 0.8298668701517706,
      "accuracy_std_dev": 0.04231782629659911,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.16520087015177065,
      "avg_duration_per_task_ms": 161607.96964586846,
      "avg_output_tokens_per_task": 16300.88813940416,
      "last_run_timestamp": "2026-09-26T08:54:23.573Z"
    },
    {
      "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": 0.8296198191945158,
      "accuracy_std_dev": 0.010647141525623993,
      "run_count": 6,
      "total_questions": 1167,
      "avg_cost_per_task_usd": 0.1271454147386461,
      "avg_duration_per_task_ms": 247054.93658954583,
      "avg_output_tokens_per_task": 29738.498714652957,
      "last_run_timestamp": "2026-09-26T08:53:15.658Z"
    },
    {
      "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": 0.8295974490238612,
      "accuracy_std_dev": 0.024864472254399347,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.15712047794649311,
      "avg_duration_per_task_ms": 856729.4591467824,
      "avg_output_tokens_per_task": 51626.77368040492,
      "last_run_timestamp": "2026-08-29T20:19:40.446Z"
    },
    {
      "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": 0.8294428739495799,
      "accuracy_std_dev": 0.015239815528915427,
      "run_count": 11,
      "total_questions": 2142,
      "avg_cost_per_task_usd": 0.009810573762838469,
      "avg_duration_per_task_ms": 575007.9150326797,
      "avg_output_tokens_per_task": 34796.17366946779,
      "last_run_timestamp": "2026-09-26T08:54:38.079Z"
    },
    {
      "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": 0.8291486389610391,
      "accuracy_std_dev": 0.04390329909775197,
      "run_count": 6,
      "total_questions": 1155,
      "avg_cost_per_task_usd": 0.028787966233766234,
      "avg_duration_per_task_ms": 286215.91515151516,
      "avg_output_tokens_per_task": 8890.160173160173,
      "last_run_timestamp": "2026-09-26T09:45:13.813Z"
    },
    {
      "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": 0.8287637142857143,
      "accuracy_std_dev": 0.08258088982275724,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.007342964646464646,
      "avg_duration_per_task_ms": 721777.8398268399,
      "avg_output_tokens_per_task": 30149.034632034633,
      "last_run_timestamp": "2026-09-26T11:30:22.248Z"
    },
    {
      "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": 0.8286193999999999,
      "accuracy_std_dev": 0.012838925852266618,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.01266686868686869,
      "avg_duration_per_task_ms": 861046.6656565657,
      "avg_output_tokens_per_task": 37375.524242424246,
      "last_run_timestamp": "2026-09-26T08:09:25.366Z"
    },
    {
      "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-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AkashML",
      "accuracy": 0.828283,
      "accuracy_std_dev": 0.016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.00548029292929293,
      "avg_duration_per_task_ms": 1403265.9545454546,
      "avg_output_tokens_per_task": 25437.79292929293,
      "last_run_timestamp": "2026-07-18T08:42:46.001Z"
    },
    {
      "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": 0.8280124253521126,
      "accuracy_std_dev": 0.05892056354319715,
      "run_count": 11,
      "total_questions": 2130,
      "avg_cost_per_task_usd": 0.052598692957746475,
      "avg_duration_per_task_ms": 609008.6507042254,
      "avg_output_tokens_per_task": 25184.795774647886,
      "last_run_timestamp": "2026-09-26T08:01:18.114Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.827862,
      "accuracy_std_dev": 0.003834504661621904,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.009011289141414141,
      "avg_duration_per_task_ms": 370127.09217171714,
      "avg_output_tokens_per_task": 21495.685606060608,
      "last_run_timestamp": "2026-08-01T20:03:34.768Z"
    },
    {
      "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": 0.8278202748091604,
      "accuracy_std_dev": 0.008498300197686612,
      "run_count": 4,
      "total_questions": 786,
      "avg_cost_per_task_usd": 0.007172601781170484,
      "avg_duration_per_task_ms": 632005.1844783715,
      "avg_output_tokens_per_task": 17107.625954198473,
      "last_run_timestamp": "2026-08-01T18:25:37.046Z"
    },
    {
      "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": 0.8275528689788053,
      "accuracy_std_dev": 0.01888290543204505,
      "run_count": 3,
      "total_questions": 519,
      "avg_cost_per_task_usd": 0.03400616570327553,
      "avg_duration_per_task_ms": 162226.8689788054,
      "avg_output_tokens_per_task": 28516.98458574181,
      "last_run_timestamp": "2026-07-18T08:31:24.271Z"
    },
    {
      "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": 0.8269733333333333,
      "accuracy_std_dev": 0.015136600014534322,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.06164716273849608,
      "avg_duration_per_task_ms": 496054.37934904604,
      "avg_output_tokens_per_task": 48584.97418630752,
      "last_run_timestamp": "2026-09-26T16:19:15.650Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "StreamLake",
      "accuracy": 0.8268798333333333,
      "accuracy_std_dev": 0.036978311140526084,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1489224217171717,
      "avg_duration_per_task_ms": 1008310.4612794613,
      "avg_output_tokens_per_task": 52023.04461279461,
      "last_run_timestamp": "2026-08-29T18:40:24.764Z"
    },
    {
      "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": 0.826599,
      "accuracy_std_dev": 0.002749143987983661,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04722597138047138,
      "avg_duration_per_task_ms": 258616.0824915825,
      "avg_output_tokens_per_task": 15105.597643097643,
      "last_run_timestamp": "2026-09-26T09:35:33.510Z"
    },
    {
      "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": 0.826575279270147,
      "accuracy_std_dev": 0.033978737517600645,
      "run_count": 10,
      "total_questions": 1973,
      "avg_cost_per_task_usd": 0.06587307298530158,
      "avg_duration_per_task_ms": 1029346.9406994425,
      "avg_output_tokens_per_task": 33210.919412062845,
      "last_run_timestamp": "2026-09-26T10:24:45.234Z"
    },
    {
      "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": 0.8265687442542787,
      "accuracy_std_dev": 0.1992607600768644,
      "run_count": 12,
      "total_questions": 2045,
      "avg_cost_per_task_usd": 0.20695833007334963,
      "avg_duration_per_task_ms": 332203.7985330073,
      "avg_output_tokens_per_task": 37422.67432762836,
      "last_run_timestamp": "2026-09-26T08:56:04.781Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Novita",
      "accuracy": 0.8263186666666665,
      "accuracy_std_dev": 0.011640048205322105,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.04355452946127946,
      "avg_duration_per_task_ms": 227785.6026936027,
      "avg_output_tokens_per_task": 14401.111111111111,
      "last_run_timestamp": "2026-09-26T09:34:17.391Z"
    },
    {
      "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": 0.8262184258271078,
      "accuracy_std_dev": 0.01942525748194862,
      "run_count": 5,
      "total_questions": 937,
      "avg_cost_per_task_usd": 0.02446647598719317,
      "avg_duration_per_task_ms": 331857.0437566702,
      "avg_output_tokens_per_task": 48723.26467449306,
      "last_run_timestamp": "2026-09-26T08:54:40.393Z"
    },
    {
      "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/qwen3.5-35b-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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8255894000000001,
      "accuracy_std_dev": 0.08582614429554669,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.007835238383838384,
      "avg_duration_per_task_ms": 420094.35454545455,
      "avg_output_tokens_per_task": 28284.40707070707,
      "last_run_timestamp": "2026-08-29T18:30:11.054Z"
    },
    {
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "GMICloud",
      "accuracy": 0.8249159000000001,
      "accuracy_std_dev": 0.02643692337035457,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.12457927070707071,
      "avg_duration_per_task_ms": 523158.02121212124,
      "avg_output_tokens_per_task": 36100.31464646465,
      "last_run_timestamp": "2026-09-26T08:54:25.911Z"
    },
    {
      "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": 0.82491575,
      "accuracy_std_dev": 0.04835478209481976,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.09927686868686869,
      "avg_duration_per_task_ms": 820653.2960858586,
      "avg_output_tokens_per_task": 39684.74116161616,
      "last_run_timestamp": "2026-08-29T18:40:35.044Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.8246351666666666,
      "accuracy_std_dev": 0.013862659393933843,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.034730324915824916,
      "avg_duration_per_task_ms": 240485.28282828283,
      "avg_output_tokens_per_task": 13780.222222222223,
      "last_run_timestamp": "2026-09-26T09:15:01.388Z"
    },
    {
      "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": 0.8246351666666666,
      "accuracy_std_dev": 0.0064044884369392945,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.13257150673400672,
      "avg_duration_per_task_ms": 144313.25757575757,
      "avg_output_tokens_per_task": 32437.94191919192,
      "last_run_timestamp": "2026-09-26T08:54:42.422Z"
    },
    {
      "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": 0.824495,
      "accuracy_std_dev": 0.00300556117888158,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.04055735353535353,
      "avg_duration_per_task_ms": 210547.6452020202,
      "avg_output_tokens_per_task": 15056.00883838384,
      "last_run_timestamp": "2026-09-26T09:19:00.848Z"
    },
    {
      "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": 0.82449475,
      "accuracy_std_dev": 0.010869783171135461,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.06350338131313131,
      "avg_duration_per_task_ms": 123563.64393939394,
      "avg_output_tokens_per_task": 13357.755050505051,
      "last_run_timestamp": "2026-09-01T17:56:39.303Z"
    },
    {
      "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": 0.8244612319391635,
      "accuracy_std_dev": 0.0189305399949394,
      "run_count": 8,
      "total_questions": 1578,
      "avg_cost_per_task_usd": 0.021243255386565272,
      "avg_duration_per_task_ms": 482977.82002534857,
      "avg_output_tokens_per_task": 22160.196451204054,
      "last_run_timestamp": "2026-08-29T18:16:26.161Z"
    },
    {
      "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": 0.8239101113924051,
      "accuracy_std_dev": 0.05687472058441742,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.21822399915611815,
      "avg_duration_per_task_ms": 1618621.7156118143,
      "avg_output_tokens_per_task": 51785.03459915612,
      "last_run_timestamp": "2026-08-29T18:42:16.346Z"
    },
    {
      "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": 0.823772114732725,
      "accuracy_std_dev": 0.15112751242638614,
      "run_count": 8,
      "total_questions": 1534,
      "avg_cost_per_task_usd": 0.06556984028683181,
      "avg_duration_per_task_ms": 284366.7822685789,
      "avg_output_tokens_per_task": 18120.139504563234,
      "last_run_timestamp": "2026-08-29T19:36:49.687Z"
    },
    {
      "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": 0.8237374,
      "accuracy_std_dev": 0.08853685691078036,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.1022956090909091,
      "avg_duration_per_task_ms": 896671.251010101,
      "avg_output_tokens_per_task": 43699.027272727275,
      "last_run_timestamp": "2026-09-26T09:26:01.477Z"
    },
    {
      "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": 0.8237132857142857,
      "accuracy_std_dev": 0.024208164205574968,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.037303974025974025,
      "avg_duration_per_task_ms": 279740.5851370851,
      "avg_output_tokens_per_task": 14323.054112554113,
      "last_run_timestamp": "2026-09-26T09:20:53.278Z"
    },
    {
      "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": 0.8234849,
      "accuracy_std_dev": 0.025448233802957704,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.04444430303030303,
      "avg_duration_per_task_ms": 358421.2823232323,
      "avg_output_tokens_per_task": 46840.29242424243,
      "last_run_timestamp": "2026-09-26T16:24:56.918Z"
    },
    {
      "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": 0.8232579938524589,
      "accuracy_std_dev": 0.05951885425846167,
      "run_count": 5,
      "total_questions": 976,
      "avg_cost_per_task_usd": 0.08256164446721312,
      "avg_duration_per_task_ms": 628781.7581967213,
      "avg_output_tokens_per_task": 23867.87807377049,
      "last_run_timestamp": "2026-09-26T09:35:44.547Z"
    },
    {
      "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": 0.823232375,
      "accuracy_std_dev": 0.046813672196105,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.08898095391414142,
      "avg_duration_per_task_ms": 476497.32765151514,
      "avg_output_tokens_per_task": 20279.160984848484,
      "last_run_timestamp": "2026-08-29T18:50:46.982Z"
    },
    {
      "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": 0.8228419167330677,
      "accuracy_std_dev": 0.11242608196694337,
      "run_count": 13,
      "total_questions": 2510,
      "avg_cost_per_task_usd": 0.007463425099601593,
      "avg_duration_per_task_ms": 612656.6079681275,
      "avg_output_tokens_per_task": 26820.300398406376,
      "last_run_timestamp": "2026-09-26T09:19:42.439Z"
    },
    {
      "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": 0.8227513720634919,
      "accuracy_std_dev": 0.011297792117682792,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.025156736507936508,
      "avg_duration_per_task_ms": 449467.82285714283,
      "avg_output_tokens_per_task": 25121.824126984127,
      "last_run_timestamp": "2026-08-29T16:15:12.678Z"
    },
    {
      "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": 0.822297,
      "accuracy_std_dev": 0.014221076658725011,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.31472648260381597,
      "avg_duration_per_task_ms": 289101.1150392817,
      "avg_output_tokens_per_task": 20792.76374859708,
      "last_run_timestamp": "2026-09-26T08:55:59.561Z"
    },
    {
      "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": 0.8222220723809524,
      "accuracy_std_dev": 0.00641547133888074,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.007690131428571429,
      "avg_duration_per_task_ms": 591255.7904761905,
      "avg_output_tokens_per_task": 17900.798095238097,
      "last_run_timestamp": "2026-08-29T18:40:24.101Z"
    },
    {
      "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": 0.8221650670103093,
      "accuracy_std_dev": 0.01835186937268852,
      "run_count": 6,
      "total_questions": 1164,
      "avg_cost_per_task_usd": 0.13827411597938147,
      "avg_duration_per_task_ms": 261027.06529209623,
      "avg_output_tokens_per_task": 31195.84621993127,
      "last_run_timestamp": "2026-09-26T09:19:47.728Z"
    },
    {
      "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": 0.8221100000000001,
      "accuracy_std_dev": 0.019325780306224248,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.011398650392817058,
      "avg_duration_per_task_ms": 623049.0549943884,
      "avg_output_tokens_per_task": 33683.20538720539,
      "last_run_timestamp": "2026-09-26T09:34:52.654Z"
    },
    {
      "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": 0.8221098333333334,
      "accuracy_std_dev": 0.033805456691766336,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0906014957912458,
      "avg_duration_per_task_ms": 1460808.6498316498,
      "avg_output_tokens_per_task": 42233.86784511784,
      "last_run_timestamp": "2026-08-29T19:12:56.521Z"
    },
    {
      "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": 0.821969625,
      "accuracy_std_dev": 0.0188925810831759,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.10310438383838386,
      "avg_duration_per_task_ms": 642163.4210858586,
      "avg_output_tokens_per_task": 38222.52777777778,
      "last_run_timestamp": "2026-08-29T18:26:14.942Z"
    },
    {
      "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": "SiliconFlow",
      "accuracy": 0.8216439045553144,
      "accuracy_std_dev": 0.014190341770056845,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.18553550686912507,
      "avg_duration_per_task_ms": 863914.5343456254,
      "avg_output_tokens_per_task": 49671.55531453362,
      "last_run_timestamp": "2026-08-29T18:39:39.402Z"
    },
    {
      "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": 0.821549,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.001158979797979798,
      "avg_duration_per_task_ms": 153563.0101010101,
      "avg_output_tokens_per_task": 18533.79292929293,
      "last_run_timestamp": "2026-09-23T18:11:58.526Z"
    },
    {
      "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": "Ionstream",
      "accuracy": 0.8215488,
      "accuracy_std_dev": 0.011711955283384605,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.039391856565656566,
      "avg_duration_per_task_ms": 244469.52525252526,
      "avg_output_tokens_per_task": 15370.382828282829,
      "last_run_timestamp": "2026-09-26T09:19:26.857Z"
    },
    {
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Google",
      "accuracy": 0.821233,
      "accuracy_std_dev": 0.011708622773409337,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.09879417929292929,
      "avg_duration_per_task_ms": 368109.4835858586,
      "avg_output_tokens_per_task": 45695.4898989899,
      "last_run_timestamp": "2026-08-30T00:09:46.578Z"
    },
    {
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "ModelRun",
      "accuracy": 0.8211350956521739,
      "accuracy_std_dev": 0.03084416399732497,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.013802629710144926,
      "avg_duration_per_task_ms": 81921.22101449275,
      "avg_output_tokens_per_task": 17742.007971014493,
      "last_run_timestamp": "2026-08-29T18:36:28.592Z"
    },
    {
      "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": 0.820920596529284,
      "accuracy_std_dev": 0.007605756273362683,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.019052912509038324,
      "avg_duration_per_task_ms": 539380.6218365871,
      "avg_output_tokens_per_task": 18914.664497469268,
      "last_run_timestamp": "2026-08-29T18:35:57.732Z"
    },
    {
      "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": 0.8207635888324873,
      "accuracy_std_dev": 0.020270393054191847,
      "run_count": 7,
      "total_questions": 1379,
      "avg_cost_per_task_usd": 0.03637786294416244,
      "avg_duration_per_task_ms": 250330.74111675128,
      "avg_output_tokens_per_task": 14481.398839738942,
      "last_run_timestamp": "2026-09-26T08:56:59.861Z"
    },
    {
      "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": 0.8206135555555556,
      "accuracy_std_dev": 0.02409886708315059,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.03784499887766555,
      "avg_duration_per_task_ms": 323218.42199775536,
      "avg_output_tokens_per_task": 39347.21212121212,
      "last_run_timestamp": "2026-09-26T16:29:49.007Z"
    },
    {
      "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": 0.8201054571428571,
      "accuracy_std_dev": 0.026122159766927572,
      "run_count": 3,
      "total_questions": 315,
      "avg_cost_per_task_usd": 0.055772555555555554,
      "avg_duration_per_task_ms": 137698.42222222223,
      "avg_output_tokens_per_task": 23066.479365079365,
      "last_run_timestamp": "2026-09-12T16:00:25.695Z"
    },
    {
      "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": 0.8198652222222221,
      "accuracy_std_dev": 0.06079266623857736,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.027504470819304155,
      "avg_duration_per_task_ms": 620185.7014590348,
      "avg_output_tokens_per_task": 28854.86756453423,
      "last_run_timestamp": "2026-09-26T09:17:20.976Z"
    },
    {
      "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-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Venice",
      "accuracy": 0.8193043333333333,
      "accuracy_std_dev": 0.0031744380430068155,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.14333356734006736,
      "avg_duration_per_task_ms": 321739.1531986532,
      "avg_output_tokens_per_task": 28661.96127946128,
      "last_run_timestamp": "2026-09-26T09:35:22.612Z"
    },
    {
      "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": 0.8193041666666666,
      "accuracy_std_dev": 0.02287040255742972,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.006429800505050506,
      "avg_duration_per_task_ms": 501214.0008417508,
      "avg_output_tokens_per_task": 17257.11026936027,
      "last_run_timestamp": "2026-08-29T19:56:51.477Z"
    },
    {
      "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": 0.8191512936660268,
      "accuracy_std_dev": 0.031450804747007274,
      "run_count": 8,
      "total_questions": 1563,
      "avg_cost_per_task_usd": 0.025433834932821493,
      "avg_duration_per_task_ms": 455424.4030710173,
      "avg_output_tokens_per_task": 21193.655790147153,
      "last_run_timestamp": "2026-08-29T19:19:06.625Z"
    },
    {
      "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": 0.8186799411764705,
      "accuracy_std_dev": 0.009390308467211003,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.019136678051865905,
      "avg_duration_per_task_ms": 558518.3769765971,
      "avg_output_tokens_per_task": 18938.588867805185,
      "last_run_timestamp": "2026-08-30T00:16:38.394Z"
    },
    {
      "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": 0.8186026249999999,
      "accuracy_std_dev": 0.021390337134659064,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.04284533396464647,
      "avg_duration_per_task_ms": 182582.7582070707,
      "avg_output_tokens_per_task": 23700.921085858587,
      "last_run_timestamp": "2026-08-29T18:13:32.148Z"
    },
    {
      "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": 0.8184969532163743,
      "accuracy_std_dev": 0.012420924411426656,
      "run_count": 8,
      "total_questions": 1539,
      "avg_cost_per_task_usd": 0.04671114684860298,
      "avg_duration_per_task_ms": 443069.649122807,
      "avg_output_tokens_per_task": 25928.164392462637,
      "last_run_timestamp": "2026-08-29T16:44:34.185Z"
    },
    {
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "SambaNova",
      "accuracy": 0.8181631193111932,
      "accuracy_std_dev": 0.03197110297229838,
      "run_count": 7,
      "total_questions": 813,
      "avg_cost_per_task_usd": 0.06679120172201722,
      "avg_duration_per_task_ms": 131925.82779827798,
      "avg_output_tokens_per_task": 28854.319803198032,
      "last_run_timestamp": "2026-08-15T14:54:40.470Z"
    },
    {
      "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": 0.8180902577475433,
      "accuracy_std_dev": 0.11577773679656365,
      "run_count": 8,
      "total_questions": 1323,
      "avg_cost_per_task_usd": 0.046483944066515494,
      "avg_duration_per_task_ms": 395321.3703703704,
      "avg_output_tokens_per_task": 15410.478458049887,
      "last_run_timestamp": "2026-08-29T18:43:58.355Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.817971375,
      "accuracy_std_dev": 0.021218711111525495,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0422043106060606,
      "avg_duration_per_task_ms": 258582.00315656565,
      "avg_output_tokens_per_task": 14114.647727272728,
      "last_run_timestamp": "2026-09-26T09:36:02.338Z"
    },
    {
      "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": 0.8177008571428571,
      "accuracy_std_dev": 0.044711167868022074,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.16188713708513708,
      "avg_duration_per_task_ms": 160453.3152958153,
      "avg_output_tokens_per_task": 16045.246753246753,
      "last_run_timestamp": "2026-09-26T08:55:42.287Z"
    },
    {
      "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": 0.8176514288840263,
      "accuracy_std_dev": 0.03360829910201664,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.22697232020423047,
      "avg_duration_per_task_ms": 936805.9460247994,
      "avg_output_tokens_per_task": 51848.20277169949,
      "last_run_timestamp": "2026-08-29T18:07:06.751Z"
    },
    {
      "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": 0.8165784761904761,
      "accuracy_std_dev": 0.01245241634466425,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.3341618246753246,
      "avg_duration_per_task_ms": 300449.1738816739,
      "avg_output_tokens_per_task": 21181.114718614717,
      "last_run_timestamp": "2026-09-26T08:51:53.334Z"
    },
    {
      "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": 0.8158729085714285,
      "accuracy_std_dev": 0.03348131954164889,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.07726252825396825,
      "avg_duration_per_task_ms": 943986.486984127,
      "avg_output_tokens_per_task": 39783.29777777778,
      "last_run_timestamp": "2026-08-30T00:30:19.495Z"
    },
    {
      "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": 0.8157559809402797,
      "accuracy_std_dev": 0.02360825327613207,
      "run_count": 8,
      "total_questions": 1574,
      "avg_cost_per_task_usd": 0.021597616899618807,
      "avg_duration_per_task_ms": 295953.15184243966,
      "avg_output_tokens_per_task": 17957.35959339263,
      "last_run_timestamp": "2026-09-26T08:54:33.107Z"
    },
    {
      "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": 0.8155159532293987,
      "accuracy_std_dev": 0.03005979094119869,
      "run_count": 7,
      "total_questions": 1347,
      "avg_cost_per_task_usd": 0.07733159317000743,
      "avg_duration_per_task_ms": 1465901.4966592428,
      "avg_output_tokens_per_task": 44227.633259094284,
      "last_run_timestamp": "2026-08-22T17:50:04.003Z"
    },
    {
      "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": 0.8154496285714286,
      "accuracy_std_dev": 0.0068932906227268985,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.006205518095238096,
      "avg_duration_per_task_ms": 548094.434920635,
      "avg_output_tokens_per_task": 18728.389206349206,
      "last_run_timestamp": "2026-08-29T19:09:30.030Z"
    },
    {
      "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": 0.8154462499999999,
      "accuracy_std_dev": 0.00789342203594232,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.024953141414141414,
      "avg_duration_per_task_ms": 185879.93371212122,
      "avg_output_tokens_per_task": 16543.260101010103,
      "last_run_timestamp": "2026-09-26T08:53:11.662Z"
    },
    {
      "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": 0.8153759444444445,
      "accuracy_std_dev": 0.010779567108769665,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.36484757070707075,
      "avg_duration_per_task_ms": 359825.24242424243,
      "avg_output_tokens_per_task": 22980.845679012345,
      "last_run_timestamp": "2026-09-26T08:54:41.498Z"
    },
    {
      "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": 0.8151659858437146,
      "accuracy_std_dev": 0.06632272963083027,
      "run_count": 19,
      "total_questions": 3532,
      "avg_cost_per_task_usd": 0.00728941053227633,
      "avg_duration_per_task_ms": 452529.82446206117,
      "avg_output_tokens_per_task": 16964.4419592299,
      "last_run_timestamp": "2026-08-29T20:12:46.798Z"
    },
    {
      "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": 0.8144508860510805,
      "accuracy_std_dev": 0.03052507251322263,
      "run_count": 8,
      "total_questions": 1527,
      "avg_cost_per_task_usd": 0.1842238932547479,
      "avg_duration_per_task_ms": 842214.8297314997,
      "avg_output_tokens_per_task": 52169.59397511461,
      "last_run_timestamp": "2026-08-29T18:14:11.455Z"
    },
    {
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8138834622823985,
      "accuracy_std_dev": 0.0221782773339139,
      "run_count": 8,
      "total_questions": 1551,
      "avg_cost_per_task_usd": 0.16634050999355257,
      "avg_duration_per_task_ms": 905404.1012250162,
      "avg_output_tokens_per_task": 52917.24887169568,
      "last_run_timestamp": "2026-08-29T19:16:33.097Z"
    },
    {
      "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": 0.8136879467680608,
      "accuracy_std_dev": 0.01151239505652493,
      "run_count": 8,
      "total_questions": 1578,
      "avg_cost_per_task_usd": 0.007660202154626108,
      "avg_duration_per_task_ms": 445275.62420785805,
      "avg_output_tokens_per_task": 19701.929657794677,
      "last_run_timestamp": "2026-08-29T19:19:45.463Z"
    },
    {
      "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": "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": "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": 0.8122604827586207,
      "accuracy_std_dev": 0.00611459963257337,
      "run_count": 3,
      "total_questions": 522,
      "avg_cost_per_task_usd": 0.07136742145593869,
      "avg_duration_per_task_ms": 345248.30842911877,
      "avg_output_tokens_per_task": 17346.159003831417,
      "last_run_timestamp": "2026-07-18T08:30:38.228Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Alibaba",
      "accuracy": 0.8112167885714286,
      "accuracy_std_dev": 0.01823961098089469,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.022302389206349204,
      "avg_duration_per_task_ms": 338892.4336507936,
      "avg_output_tokens_per_task": 19867.52,
      "last_run_timestamp": "2026-08-29T18:59:02.809Z"
    },
    {
      "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": 0.8108436061269148,
      "accuracy_std_dev": 0.03441946501380775,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.21617651641137853,
      "avg_duration_per_task_ms": 867715.7636761488,
      "avg_output_tokens_per_task": 54511.27789934354,
      "last_run_timestamp": "2026-08-29T18:09:15.344Z"
    },
    {
      "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": 0.8108163749999999,
      "accuracy_std_dev": 0.010578591517039247,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.03051571590909091,
      "avg_duration_per_task_ms": 224324.35542929292,
      "avg_output_tokens_per_task": 24211.31691919192,
      "last_run_timestamp": "2026-08-29T19:24:09.844Z"
    },
    {
      "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": 0.8107742,
      "accuracy_std_dev": 0.00793923498581567,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.032759715151515156,
      "avg_duration_per_task_ms": 674236.5636363636,
      "avg_output_tokens_per_task": 15717.559595959596,
      "last_run_timestamp": "2026-09-26T09:29:37.459Z"
    },
    {
      "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": 0.8099527992277992,
      "accuracy_std_dev": 0.02090205677678631,
      "run_count": 4,
      "total_questions": 777,
      "avg_cost_per_task_usd": 0.00813298326898327,
      "avg_duration_per_task_ms": 668505.0875160876,
      "avg_output_tokens_per_task": 19405.019305019305,
      "last_run_timestamp": "2026-08-01T18:07:34.003Z"
    },
    {
      "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": 0.8097504682741116,
      "accuracy_std_dev": 0.10898800173843678,
      "run_count": 8,
      "total_questions": 1576,
      "avg_cost_per_task_usd": 0.1267229282994924,
      "avg_duration_per_task_ms": 700808.2106598985,
      "avg_output_tokens_per_task": 40110.88578680203,
      "last_run_timestamp": "2026-08-29T18:23:09.250Z"
    },
    {
      "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": 0.8094891786492375,
      "accuracy_std_dev": 0.052310168338791006,
      "run_count": 7,
      "total_questions": 1377,
      "avg_cost_per_task_usd": 0.23138632244008714,
      "avg_duration_per_task_ms": 967370.0573710966,
      "avg_output_tokens_per_task": 53108.82788671024,
      "last_run_timestamp": "2026-08-29T18:21:26.899Z"
    },
    {
      "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": 0.8092030833333335,
      "accuracy_std_dev": 0.01675517045301904,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.33934875,
      "avg_duration_per_task_ms": 354134.20664983167,
      "avg_output_tokens_per_task": 22457.28787878788,
      "last_run_timestamp": "2026-09-26T08:56:43.900Z"
    },
    {
      "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": 0.8091603740458015,
      "accuracy_std_dev": 0.1072178269085883,
      "run_count": 12,
      "total_questions": 2358,
      "avg_cost_per_task_usd": 0.04859381764206954,
      "avg_duration_per_task_ms": 198361.56318914334,
      "avg_output_tokens_per_task": 23259.173452078034,
      "last_run_timestamp": "2026-09-26T09:18:49.857Z"
    },
    {
      "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": 0.8088945473145778,
      "accuracy_std_dev": 0.022541716685711595,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.00648230179028133,
      "avg_duration_per_task_ms": 718133.9377664109,
      "avg_output_tokens_per_task": 17528.813299232737,
      "last_run_timestamp": "2026-08-29T19:15:43.723Z"
    },
    {
      "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": 0.8081382125237192,
      "accuracy_std_dev": 0.024044703190099895,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.1062299411764706,
      "avg_duration_per_task_ms": 740176.1967109424,
      "avg_output_tokens_per_task": 54188.096774193546,
      "last_run_timestamp": "2026-08-29T19:17:49.455Z"
    },
    {
      "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": 0.808081,
      "accuracy_std_dev": 0.12407749055328285,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.10305112289562289,
      "avg_duration_per_task_ms": 1499684.5976430976,
      "avg_output_tokens_per_task": 25661.05218855219,
      "last_run_timestamp": "2026-09-26T09:01:35.348Z"
    },
    {
      "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": 0.8074670728744938,
      "accuracy_std_dev": 0.01649167806046429,
      "run_count": 4,
      "total_questions": 741,
      "avg_cost_per_task_usd": 0.07785456275303644,
      "avg_duration_per_task_ms": 468008.1632928475,
      "avg_output_tokens_per_task": 33473.376518218625,
      "last_run_timestamp": "2026-08-01T17:52:55.742Z"
    },
    {
      "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": 0.8073004516129032,
      "accuracy_std_dev": 0.05300431140963913,
      "run_count": 10,
      "total_questions": 1767,
      "avg_cost_per_task_usd": 0.046523333333333326,
      "avg_duration_per_task_ms": 546722.9988681381,
      "avg_output_tokens_per_task": 25674.418222976798,
      "last_run_timestamp": "2026-09-26T09:56:00.145Z"
    },
    {
      "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": 0.8072388749999999,
      "accuracy_std_dev": 0.022365952279287706,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.007244868686868686,
      "avg_duration_per_task_ms": 515270.8522727273,
      "avg_output_tokens_per_task": 18306.11994949495,
      "last_run_timestamp": "2026-08-29T19:12:36.104Z"
    },
    {
      "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": 0.8068935127478752,
      "accuracy_std_dev": 0.009178499999999978,
      "run_count": 2,
      "total_questions": 353,
      "avg_cost_per_task_usd": 0.033640501416430595,
      "avg_duration_per_task_ms": 174824.1133144476,
      "avg_output_tokens_per_task": 26792.110481586402,
      "last_run_timestamp": "2026-07-18T08:04:57.082Z"
    },
    {
      "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": "CoreWeave",
      "accuracy": 0.8063971999999998,
      "accuracy_std_dev": 0.01659741351415936,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.00614619090909091,
      "avg_duration_per_task_ms": 535780.8212121212,
      "avg_output_tokens_per_task": 18359.183838383837,
      "last_run_timestamp": "2026-08-29T19:12:20.528Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.8054487816091953,
      "accuracy_std_dev": 0.036888297226220654,
      "run_count": 4,
      "total_questions": 783,
      "avg_cost_per_task_usd": 0.008708275862068965,
      "avg_duration_per_task_ms": 1038271.9885057472,
      "avg_output_tokens_per_task": 20826.485312899105,
      "last_run_timestamp": "2026-08-01T22:57:01.346Z"
    },
    {
      "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": 0.8051947142857142,
      "accuracy_std_dev": 0.014445950710090603,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.022680748196248195,
      "avg_duration_per_task_ms": 254100.05483405484,
      "avg_output_tokens_per_task": 17520.435786435788,
      "last_run_timestamp": "2026-09-26T08:51:26.240Z"
    },
    {
      "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": 0.805091904400607,
      "accuracy_std_dev": 0.04892398259146531,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.08381860748609003,
      "avg_duration_per_task_ms": 790750.7895801719,
      "avg_output_tokens_per_task": 31756.94537177542,
      "last_run_timestamp": "2026-09-26T10:35:48.204Z"
    },
    {
      "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": 0.8049008888888889,
      "accuracy_std_dev": 0.023075324904150284,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.3397021683501683,
      "avg_duration_per_task_ms": 355471.79349046014,
      "avg_output_tokens_per_task": 22495.737373737375,
      "last_run_timestamp": "2026-09-26T09:00:45.748Z"
    },
    {
      "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": "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": 0.804676697257384,
      "accuracy_std_dev": 0.059236782672255274,
      "run_count": 5,
      "total_questions": 948,
      "avg_cost_per_task_usd": 0.023476022151898732,
      "avg_duration_per_task_ms": 1321432.723628692,
      "avg_output_tokens_per_task": 50466.419831223626,
      "last_run_timestamp": "2026-09-30T00:39:58.995Z"
    },
    {
      "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": 0.8038987567251462,
      "accuracy_std_dev": 0.03143660042823293,
      "run_count": 8,
      "total_questions": 855,
      "avg_cost_per_task_usd": 0.02817445146198831,
      "avg_duration_per_task_ms": 97917.72046783626,
      "avg_output_tokens_per_task": 30151.957894736843,
      "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": 0.802867256166983,
      "accuracy_std_dev": 0.022011618153474646,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.02734331878557875,
      "avg_duration_per_task_ms": 434011.0303605313,
      "avg_output_tokens_per_task": 22775.223276407338,
      "last_run_timestamp": "2026-08-29T18:09:57.600Z"
    },
    {
      "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": 0.8025491428571428,
      "accuracy_std_dev": 0.025111070964409142,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.33407534559884555,
      "avg_duration_per_task_ms": 352449.7582972583,
      "avg_output_tokens_per_task": 22169.862914862915,
      "last_run_timestamp": "2026-09-26T08:59:18.032Z"
    },
    {
      "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": 0.8023003671874999,
      "accuracy_std_dev": 0.01483082484009234,
      "run_count": 8,
      "total_questions": 1536,
      "avg_cost_per_task_usd": 0.016652225911458332,
      "avg_duration_per_task_ms": 384547.912109375,
      "avg_output_tokens_per_task": 15953.287760416666,
      "last_run_timestamp": "2026-08-29T16:10:33.070Z"
    },
    {
      "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": 0.8021885000000001,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04024864141414142,
      "avg_duration_per_task_ms": 73206.09595959596,
      "avg_output_tokens_per_task": 26709.396464646463,
      "last_run_timestamp": "2026-09-01T17:53:21.017Z"
    },
    {
      "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": "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": 0.8013465,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.041531338383838386,
      "avg_duration_per_task_ms": 80828.55555555556,
      "avg_output_tokens_per_task": 27572.47474747475,
      "last_run_timestamp": "2026-09-01T17:50:00.305Z"
    },
    {
      "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": 0.8009767048665621,
      "accuracy_std_dev": 0.1773041368489748,
      "run_count": 10,
      "total_questions": 1911,
      "avg_cost_per_task_usd": 0.0910193940345369,
      "avg_duration_per_task_ms": 503465.8440607012,
      "avg_output_tokens_per_task": 24546.500784929358,
      "last_run_timestamp": "2026-09-26T08:54:41.414Z"
    },
    {
      "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": 0.8007155,
      "accuracy_std_dev": 0.04390787837905629,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.039541025883838385,
      "avg_duration_per_task_ms": 942411.1104797979,
      "avg_output_tokens_per_task": 24048.43876262626,
      "last_run_timestamp": "2026-09-26T09:34:13.954Z"
    },
    {
      "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": 0.8001721395348838,
      "accuracy_std_dev": 0.0356745,
      "run_count": 2,
      "total_questions": 387,
      "avg_cost_per_task_usd": 0.184302338501292,
      "avg_duration_per_task_ms": 1402883.9819121447,
      "avg_output_tokens_per_task": 51056.865633074936,
      "last_run_timestamp": "2026-09-28T18:08:56.704Z"
    },
    {
      "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": 0.8001197163375224,
      "accuracy_std_dev": 0.08791149969002918,
      "run_count": 8,
      "total_questions": 1114,
      "avg_cost_per_task_usd": 0.010599297127468581,
      "avg_duration_per_task_ms": 62868.36624775583,
      "avg_output_tokens_per_task": 10694.411131059245,
      "last_run_timestamp": "2026-08-29T18:12:40.868Z"
    },
    {
      "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": 0.7988215,
      "accuracy_std_dev": 0.010942500000000022,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.25381196717171717,
      "avg_duration_per_task_ms": 225693.02525252526,
      "avg_output_tokens_per_task": 25403.512626262625,
      "last_run_timestamp": "2026-09-01T13:22:15.928Z"
    },
    {
      "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": 0.7987011428571428,
      "accuracy_std_dev": 0.00925850994535102,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.04827445887445887,
      "avg_duration_per_task_ms": 463705.8015873016,
      "avg_output_tokens_per_task": 25011.621933621933,
      "last_run_timestamp": "2026-08-22T17:18:01.115Z"
    },
    {
      "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": 0.79840075,
      "accuracy_std_dev": 0.009249753655503483,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.014099073863636364,
      "avg_duration_per_task_ms": 333383.3731060606,
      "avg_output_tokens_per_task": 14661.142676767677,
      "last_run_timestamp": "2026-08-30T00:01:09.438Z"
    },
    {
      "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": 0.79798,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.2411195303030303,
      "avg_duration_per_task_ms": 184685.6616161616,
      "avg_output_tokens_per_task": 24018.91919191919,
      "last_run_timestamp": "2026-09-01T13:34:01.175Z"
    },
    {
      "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-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Mancer 2",
      "accuracy": 0.7979798,
      "accuracy_std_dev": 0.09258320044241286,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.010301554545454545,
      "avg_duration_per_task_ms": 568986.8757575757,
      "avg_output_tokens_per_task": 24756.19292929293,
      "last_run_timestamp": "2026-08-29T19:51:27.312Z"
    },
    {
      "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": 0.7979040119760479,
      "accuracy_std_dev": 0.02389207889003172,
      "run_count": 4,
      "total_questions": 668,
      "avg_cost_per_task_usd": 0.15626441916167666,
      "avg_duration_per_task_ms": 1788980.868263473,
      "avg_output_tokens_per_task": 46933.58532934132,
      "last_run_timestamp": "2026-08-22T16:44:47.658Z"
    },
    {
      "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": 0.7974987142857143,
      "accuracy_std_dev": 0.009164318316387839,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.01350666378066378,
      "avg_duration_per_task_ms": 299956.4531024531,
      "avg_output_tokens_per_task": 14063.395382395382,
      "last_run_timestamp": "2026-08-29T19:12:57.906Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.7969696999999999,
      "accuracy_std_dev": 0.10557091126825607,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.09371603636363636,
      "avg_duration_per_task_ms": 530749.2,
      "avg_output_tokens_per_task": 34572.13787878788,
      "last_run_timestamp": "2026-09-26T09:22:12.564Z"
    },
    {
      "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": 0.7969276249999999,
      "accuracy_std_dev": 0.020340632081731754,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.027590136994949492,
      "avg_duration_per_task_ms": 278109.5505050505,
      "avg_output_tokens_per_task": 18237.960227272728,
      "last_run_timestamp": "2026-08-29T18:07:47.743Z"
    },
    {
      "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": 0.7968618091603052,
      "accuracy_std_dev": 0.006526907862839814,
      "run_count": 8,
      "total_questions": 1572,
      "avg_cost_per_task_usd": 0.055094477735368956,
      "avg_duration_per_task_ms": 1236078.9319338421,
      "avg_output_tokens_per_task": 25116.7868956743,
      "last_run_timestamp": "2026-08-29T20:02:23.660Z"
    },
    {
      "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": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.7962965,
      "accuracy_std_dev": 0.008417500000000022,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.044405648989898995,
      "avg_duration_per_task_ms": 249218.0202020202,
      "avg_output_tokens_per_task": 22108.909090909092,
      "last_run_timestamp": "2026-09-01T18:00:27.682Z"
    },
    {
      "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": 0.796296,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04404542424242425,
      "avg_duration_per_task_ms": 221111.49494949495,
      "avg_output_tokens_per_task": 21928.032828282827,
      "last_run_timestamp": "2026-09-01T18:07:51.939Z"
    },
    {
      "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": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Novita",
      "accuracy": 0.7955469441233141,
      "accuracy_std_dev": 0.07292560689454357,
      "run_count": 8,
      "total_questions": 1557,
      "avg_cost_per_task_usd": 0.18441899165061015,
      "avg_duration_per_task_ms": 1050598.7052023122,
      "avg_output_tokens_per_task": 59274.58445728966,
      "last_run_timestamp": "2026-08-29T18:11:51.472Z"
    },
    {
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.7945011215854313,
      "accuracy_std_dev": 0.05670915857959413,
      "run_count": 12,
      "total_questions": 1867,
      "avg_cost_per_task_usd": 0.05605855168719871,
      "avg_duration_per_task_ms": 749155.3829673273,
      "avg_output_tokens_per_task": 45440.918050348155,
      "last_run_timestamp": "2026-08-29T19:07:57.500Z"
    },
    {
      "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": 0.7942543479087453,
      "accuracy_std_dev": 0.04886147848238298,
      "run_count": 8,
      "total_questions": 1578,
      "avg_cost_per_task_usd": 0.1139862959442332,
      "avg_duration_per_task_ms": 955060.933460076,
      "avg_output_tokens_per_task": 52224.47401774398,
      "last_run_timestamp": "2026-08-29T18:40:30.684Z"
    },
    {
      "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": 0.79419175,
      "accuracy_std_dev": 0.04229542576553049,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.027481944444444444,
      "avg_duration_per_task_ms": 377974.2102272727,
      "avg_output_tokens_per_task": 25769.56186868687,
      "last_run_timestamp": "2026-08-29T19:18:38.408Z"
    },
    {
      "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": 0.7936028,
      "accuracy_std_dev": 0.016273367695716804,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.017573431313131313,
      "avg_duration_per_task_ms": 59390.138383838384,
      "avg_output_tokens_per_task": 5706.428282828283,
      "last_run_timestamp": "2026-08-29T18:05:48.084Z"
    },
    {
      "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": 0.7930948249452954,
      "accuracy_std_dev": 0.06762071684847804,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.17062498103574034,
      "avg_duration_per_task_ms": 873257.2749817652,
      "avg_output_tokens_per_task": 58045.98614150255,
      "last_run_timestamp": "2026-08-29T19:22:50.890Z"
    },
    {
      "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": 0.7929289999999999,
      "accuracy_std_dev": 0.016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03961820454545454,
      "avg_duration_per_task_ms": 540887.0883838384,
      "avg_output_tokens_per_task": 13782.598484848484,
      "last_run_timestamp": "2026-08-08T09:30:55.143Z"
    },
    {
      "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": 0.7922558,
      "accuracy_std_dev": 0.03991858831121162,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.04258375404040404,
      "avg_duration_per_task_ms": 463527.07676767674,
      "avg_output_tokens_per_task": 42373.13080808081,
      "last_run_timestamp": "2026-09-26T16:31:04.519Z"
    },
    {
      "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": 0.7921716999999999,
      "accuracy_std_dev": 0.051077414533333594,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.17702268080808084,
      "avg_duration_per_task_ms": 826914.541919192,
      "avg_output_tokens_per_task": 57978.44292929293,
      "last_run_timestamp": "2026-09-26T09:47:43.340Z"
    },
    {
      "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": 0.7918424506329113,
      "accuracy_std_dev": 0.007923520859160772,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.18873121181434602,
      "avg_duration_per_task_ms": 1008777.9578059071,
      "avg_output_tokens_per_task": 47257.703797468355,
      "last_run_timestamp": "2026-08-29T20:08:40.362Z"
    },
    {
      "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": 0.7915958188529246,
      "accuracy_std_dev": 0.016282913854712845,
      "run_count": 10,
      "total_questions": 1761,
      "avg_cost_per_task_usd": 0.004947599091425326,
      "avg_duration_per_task_ms": 1178125.1942078364,
      "avg_output_tokens_per_task": 32554.07495741056,
      "last_run_timestamp": "2026-09-26T21:21:20.008Z"
    },
    {
      "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": 0.7908496524064171,
      "accuracy_std_dev": 0.011274463999331494,
      "run_count": 6,
      "total_questions": 1122,
      "avg_cost_per_task_usd": 0.013792627450980394,
      "avg_duration_per_task_ms": 448206.95098039217,
      "avg_output_tokens_per_task": 14371.752228163992,
      "last_run_timestamp": "2026-08-15T14:56:55.269Z"
    },
    {
      "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": 0.7908466547406083,
      "accuracy_std_dev": 0.0326514777957277,
      "run_count": 7,
      "total_questions": 1118,
      "avg_cost_per_task_usd": 0.0524859221824687,
      "avg_duration_per_task_ms": 809286.446332737,
      "avg_output_tokens_per_task": 22749.67352415027,
      "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-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Chutes",
      "accuracy": 0.7907120376940132,
      "accuracy_std_dev": 0.04068089691283433,
      "run_count": 7,
      "total_questions": 1353,
      "avg_cost_per_task_usd": 0.1386707028824834,
      "avg_duration_per_task_ms": 1542626.872135994,
      "avg_output_tokens_per_task": 46143.74427198817,
      "last_run_timestamp": "2026-08-29T18:08:11.373Z"
    },
    {
      "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": 0.790404,
      "accuracy_std_dev": 0.009259000000000017,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0189197702020202,
      "avg_duration_per_task_ms": 459328.30303030304,
      "avg_output_tokens_per_task": 25138.35606060606,
      "last_run_timestamp": "2026-08-29T19:17:42.373Z"
    },
    {
      "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": 0.7902669767441859,
      "accuracy_std_dev": 0.0482520453161055,
      "run_count": 8,
      "total_questions": 1548,
      "avg_cost_per_task_usd": 0.09614282041343669,
      "avg_duration_per_task_ms": 503336.3320413437,
      "avg_output_tokens_per_task": 28837.058785529716,
      "last_run_timestamp": "2026-08-29T18:54:50.382Z"
    },
    {
      "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": 0.7885914761215629,
      "accuracy_std_dev": 0.02234077768730239,
      "run_count": 7,
      "total_questions": 1382,
      "avg_cost_per_task_usd": 0.06106606946454414,
      "avg_duration_per_task_ms": 179070.87192474675,
      "avg_output_tokens_per_task": 19226.02532561505,
      "last_run_timestamp": "2026-09-26T09:36:55.311Z"
    },
    {
      "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": 0.7884858931297708,
      "accuracy_std_dev": 0.009706627796897349,
      "run_count": 8,
      "total_questions": 1572,
      "avg_cost_per_task_usd": 0.01825112468193384,
      "avg_duration_per_task_ms": 147479.76017811705,
      "avg_output_tokens_per_task": 15055.673664122138,
      "last_run_timestamp": "2026-09-26T09:15:51.063Z"
    },
    {
      "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": 0.7884401666666667,
      "accuracy_std_dev": 0.023761946338748348,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.04315188552188553,
      "avg_duration_per_task_ms": 84212.74747474748,
      "avg_output_tokens_per_task": 26885.046296296296,
      "last_run_timestamp": "2026-09-01T17:55:54.208Z"
    },
    {
      "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": 0.7878980417457305,
      "accuracy_std_dev": 0.013974852474623847,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.05333916508538899,
      "avg_duration_per_task_ms": 933037.6869070209,
      "avg_output_tokens_per_task": 24223.848829854524,
      "last_run_timestamp": "2026-08-29T19:47:52.368Z"
    },
    {
      "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": 0.7877272418181819,
      "accuracy_std_dev": 0.05234185111680073,
      "run_count": 6,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.007575785454545454,
      "avg_duration_per_task_ms": 581307.9736363636,
      "avg_output_tokens_per_task": 17149.295454545456,
      "last_run_timestamp": "2026-08-29T19:08:46.030Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Phala",
      "accuracy": 0.7875873916083916,
      "accuracy_std_dev": 0.10155079262062182,
      "run_count": 9,
      "total_questions": 1716,
      "avg_cost_per_task_usd": 0.08341448135198136,
      "avg_duration_per_task_ms": 882544.3578088578,
      "avg_output_tokens_per_task": 30968.915501165502,
      "last_run_timestamp": "2026-09-26T10:34:37.144Z"
    },
    {
      "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": "WandB",
      "accuracy": 0.7867566666666665,
      "accuracy_std_dev": 0.004199386807089298,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.021028161616161615,
      "avg_duration_per_task_ms": 321362.4663299663,
      "avg_output_tokens_per_task": 12498.824915824916,
      "last_run_timestamp": "2026-07-18T16:16:42.667Z"
    },
    {
      "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": 0.7848129301310043,
      "accuracy_std_dev": 0.07293711306212178,
      "run_count": 7,
      "total_questions": 1374,
      "avg_cost_per_task_usd": 0.19101991703056767,
      "avg_duration_per_task_ms": 1250087.5604075692,
      "avg_output_tokens_per_task": 61492.3442503639,
      "last_run_timestamp": "2026-08-29T18:49:03.323Z"
    },
    {
      "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": 0.7847221249999999,
      "accuracy_std_dev": 0.15720335838129978,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.017239123106060607,
      "avg_duration_per_task_ms": 568625.6464646464,
      "avg_output_tokens_per_task": 27456.18876262626,
      "last_run_timestamp": "2026-09-26T08:08:15.090Z"
    },
    {
      "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": 0.784512,
      "accuracy_std_dev": 0.043771000000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.030820969696969697,
      "avg_duration_per_task_ms": 734052.446969697,
      "avg_output_tokens_per_task": 25681.72474747475,
      "last_run_timestamp": "2026-07-18T09:15:42.531Z"
    },
    {
      "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": 0.7841148783269962,
      "accuracy_std_dev": 0.01731134085469697,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.016873125475285175,
      "avg_duration_per_task_ms": 173156.85044359948,
      "avg_output_tokens_per_task": 6557.945500633714,
      "last_run_timestamp": "2026-08-01T17:44:11.144Z"
    },
    {
      "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": 0.7839965421750663,
      "accuracy_std_dev": 0.016550632751952416,
      "run_count": 10,
      "total_questions": 1885,
      "avg_cost_per_task_usd": 0.008975832360742704,
      "avg_duration_per_task_ms": 426995.76127320953,
      "avg_output_tokens_per_task": 35628.419628647214,
      "last_run_timestamp": "2026-09-26T17:33:23.516Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "WandB",
      "accuracy": 0.7830773446153846,
      "accuracy_std_dev": 0.015297999999999978,
      "run_count": 2,
      "total_questions": 325,
      "avg_cost_per_task_usd": 0.07515540923076923,
      "avg_duration_per_task_ms": 234345.21538461538,
      "avg_output_tokens_per_task": 20795.20923076923,
      "last_run_timestamp": "2026-07-18T08:37:00.147Z"
    },
    {
      "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": 0.7826841480519481,
      "accuracy_std_dev": 0.051647,
      "run_count": 2,
      "total_questions": 385,
      "avg_cost_per_task_usd": 0.02775078961038961,
      "avg_duration_per_task_ms": 131229.6181818182,
      "avg_output_tokens_per_task": 22060.571428571428,
      "last_run_timestamp": "2026-07-18T16:00:08.986Z"
    },
    {
      "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": 0.7824072499999999,
      "accuracy_std_dev": 0.0034452088742919643,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.04613476767676768,
      "avg_duration_per_task_ms": 199616.12626262626,
      "avg_output_tokens_per_task": 21878.295454545456,
      "last_run_timestamp": "2026-09-01T19:10:24.853Z"
    },
    {
      "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": 0.782216,
      "accuracy_std_dev": 0.021976765048568925,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.0050231294765840226,
      "avg_duration_per_task_ms": 272954.46831955924,
      "avg_output_tokens_per_task": 32942.98806244261,
      "last_run_timestamp": "2026-09-26T16:14:01.905Z"
    },
    {
      "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": 0.7819865,
      "accuracy_std_dev": 0.016169511173295674,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.252381888047138,
      "avg_duration_per_task_ms": 227878.92592592593,
      "avg_output_tokens_per_task": 25272.329124579126,
      "last_run_timestamp": "2026-09-01T13:35:35.064Z"
    },
    {
      "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": 0.7808641666666666,
      "accuracy_std_dev": 0.010351374971739527,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.021081101010101008,
      "avg_duration_per_task_ms": 286187.4873737374,
      "avg_output_tokens_per_task": 12540.748316498317,
      "last_run_timestamp": "2026-08-30T00:44:42.602Z"
    },
    {
      "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": 0.7805236273764259,
      "accuracy_std_dev": 0.04118353988226817,
      "run_count": 8,
      "total_questions": 1578,
      "avg_cost_per_task_usd": 0.013252240811153358,
      "avg_duration_per_task_ms": 364659.76552598225,
      "avg_output_tokens_per_task": 13873.679974651457,
      "last_run_timestamp": "2026-08-29T16:52:23.150Z"
    },
    {
      "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": 0.7804641392649903,
      "accuracy_std_dev": 0.04111153382981473,
      "run_count": 5,
      "total_questions": 517,
      "avg_cost_per_task_usd": 0.012311303675048357,
      "avg_duration_per_task_ms": 53239.41392649903,
      "avg_output_tokens_per_task": 16349.947775628627,
      "last_run_timestamp": "2026-08-29T19:07:43.372Z"
    },
    {
      "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": 0.7796308899266178,
      "accuracy_std_dev": 0.043835688799196484,
      "run_count": 8,
      "total_questions": 1499,
      "avg_cost_per_task_usd": 0.04210462508338893,
      "avg_duration_per_task_ms": 691803.9633088725,
      "avg_output_tokens_per_task": 19458.18612408272,
      "last_run_timestamp": "2026-08-29T16:22:26.395Z"
    },
    {
      "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": 0.7794614999999999,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.013140674242424242,
      "avg_duration_per_task_ms": 69975.75757575757,
      "avg_output_tokens_per_task": 10390.04797979798,
      "last_run_timestamp": "2026-09-01T17:51:03.252Z"
    },
    {
      "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": 0.7792592914285713,
      "accuracy_std_dev": 0.02420591387160119,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.05840109523809524,
      "avg_duration_per_task_ms": 642342.9288888889,
      "avg_output_tokens_per_task": 25397.8926984127,
      "last_run_timestamp": "2026-08-29T16:48:15.193Z"
    },
    {
      "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": 0.7787957975567189,
      "accuracy_std_dev": 0.01661363892243051,
      "run_count": 7,
      "total_questions": 1146,
      "avg_cost_per_task_usd": 0.054749945898778365,
      "avg_duration_per_task_ms": 994167.5584642234,
      "avg_output_tokens_per_task": 28666.913612565444,
      "last_run_timestamp": "2026-08-29T18:31:11.241Z"
    },
    {
      "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": 0.7781144000000001,
      "accuracy_std_dev": 0.02296971973359712,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.20166651515151518,
      "avg_duration_per_task_ms": 657386.6868686868,
      "avg_output_tokens_per_task": 48003.93838383839,
      "last_run_timestamp": "2026-08-29T18:10:32.359Z"
    },
    {
      "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": 0.7777777692307692,
      "accuracy_std_dev": 0.019987758265998724,
      "run_count": 8,
      "total_questions": 1560,
      "avg_cost_per_task_usd": 0.015498306410256409,
      "avg_duration_per_task_ms": 422600.68076923076,
      "avg_output_tokens_per_task": 15705.085256410257,
      "last_run_timestamp": "2026-08-29T18:21:38.666Z"
    },
    {
      "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": 0.7752525,
      "accuracy_std_dev": 0.021487984431227306,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.02151301430976431,
      "avg_duration_per_task_ms": 770288.1026936027,
      "avg_output_tokens_per_task": 37496.628787878784,
      "last_run_timestamp": "2026-08-29T18:36:10.280Z"
    },
    {
      "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": 0.7744108571428571,
      "accuracy_std_dev": 0.04740344394188064,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.17528118686868685,
      "avg_duration_per_task_ms": 1415587.722943723,
      "avg_output_tokens_per_task": 45015.56204906205,
      "last_run_timestamp": "2026-08-29T19:12:26.877Z"
    },
    {
      "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": 0.774186536121673,
      "accuracy_std_dev": 0.045556886694000515,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.0076262953105196455,
      "avg_duration_per_task_ms": 513514.7541191382,
      "avg_output_tokens_per_task": 15020.147021546261,
      "last_run_timestamp": "2026-09-26T08:55:07.944Z"
    },
    {
      "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": 0.7741623318872016,
      "accuracy_std_dev": 0.020647498917569188,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.05963888141720896,
      "avg_duration_per_task_ms": 613056.7744034707,
      "avg_output_tokens_per_task": 27371.603036876357,
      "last_run_timestamp": "2026-08-29T18:26:43.144Z"
    },
    {
      "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": 0.7722750914285714,
      "accuracy_std_dev": 0.032206594225552256,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.11735591428571429,
      "avg_duration_per_task_ms": 772416.0933333334,
      "avg_output_tokens_per_task": 45815.12253968254,
      "last_run_timestamp": "2026-08-29T19:11:37.135Z"
    },
    {
      "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": 0.7722621887655128,
      "accuracy_std_dev": 0.13097706644765858,
      "run_count": 8,
      "total_questions": 1531,
      "avg_cost_per_task_usd": 0.02291441149575441,
      "avg_duration_per_task_ms": 346959.66819072503,
      "avg_output_tokens_per_task": 23385.622468974525,
      "last_run_timestamp": "2026-08-29T20:00:43.878Z"
    },
    {
      "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": 0.7722220999999999,
      "accuracy_std_dev": 0.025764007034815056,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.005924544949494949,
      "avg_duration_per_task_ms": 705308.8757575757,
      "avg_output_tokens_per_task": 37071.080808080806,
      "last_run_timestamp": "2026-09-26T19:06:56.023Z"
    },
    {
      "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": 0.7718254330357142,
      "accuracy_std_dev": 0.0208914466356596,
      "run_count": 7,
      "total_questions": 1344,
      "avg_cost_per_task_usd": 0.016976154761904765,
      "avg_duration_per_task_ms": 826873.3988095238,
      "avg_output_tokens_per_task": 42408.13392857143,
      "last_run_timestamp": "2026-08-29T16:45:34.440Z"
    },
    {
      "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": "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": 0.7703205704697986,
      "accuracy_std_dev": 0.02824210987088577,
      "run_count": 7,
      "total_questions": 1341,
      "avg_cost_per_task_usd": 0.016340362416107385,
      "avg_duration_per_task_ms": 580033.3325876212,
      "avg_output_tokens_per_task": 40993.44817300522,
      "last_run_timestamp": "2026-08-29T16:43:11.936Z"
    },
    {
      "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": 0.7688491071428571,
      "accuracy_std_dev": 0.036143499999999995,
      "run_count": 2,
      "total_questions": 168,
      "avg_cost_per_task_usd": 0.013457690476190477,
      "avg_duration_per_task_ms": 239500.45238095237,
      "avg_output_tokens_per_task": 24312.886904761905,
      "last_run_timestamp": "2026-07-18T09:01:08.846Z"
    },
    {
      "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": 0.7687991666666666,
      "accuracy_std_dev": 0.023722281329421564,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.017075120370370372,
      "avg_duration_per_task_ms": 891145.8181818182,
      "avg_output_tokens_per_task": 43562.35101010101,
      "last_run_timestamp": "2026-08-29T18:49:24.022Z"
    },
    {
      "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": 0.7686387142857142,
      "accuracy_std_dev": 0.039191927622605004,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.05355179797979799,
      "avg_duration_per_task_ms": 647628.132034632,
      "avg_output_tokens_per_task": 24095.015151515152,
      "last_run_timestamp": "2026-08-22T17:09:41.940Z"
    },
    {
      "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": 0.7685185,
      "accuracy_std_dev": 0.005892499999999967,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.013512386363636364,
      "avg_duration_per_task_ms": 71792.87121212122,
      "avg_output_tokens_per_task": 10688.070707070707,
      "last_run_timestamp": "2026-09-01T17:55:24.769Z"
    },
    {
      "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": 0.7676545740944017,
      "accuracy_std_dev": 0.04018594976550331,
      "run_count": 6,
      "total_questions": 911,
      "avg_cost_per_task_usd": 0.14126989352360045,
      "avg_duration_per_task_ms": 1023387.3852908892,
      "avg_output_tokens_per_task": 40525.553238199784,
      "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-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "Azure",
      "accuracy": 0.7672560000000002,
      "accuracy_std_dev": 0.0089379450938121,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.013739436868686868,
      "avg_duration_per_task_ms": 117371.45833333333,
      "avg_output_tokens_per_task": 10352.156565656565,
      "last_run_timestamp": "2026-09-01T17:58:15.198Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Reka",
      "accuracy": 0.7671957142857142,
      "accuracy_std_dev": 0.009679973770836427,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.017645118326118325,
      "avg_duration_per_task_ms": 105010.88528138529,
      "avg_output_tokens_per_task": 6033.81746031746,
      "last_run_timestamp": "2026-09-26T09:17:54.791Z"
    },
    {
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Mara",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05139547999999999,
      "avg_duration_per_task_ms": 172458.62,
      "avg_output_tokens_per_task": 21216.74,
      "last_run_timestamp": "2026-09-14T19:42:46.225Z"
    },
    {
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "BaseTen",
      "accuracy": 0.766309,
      "accuracy_std_dev": 0.007116882358167777,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.014960216540404043,
      "avg_duration_per_task_ms": 157155.3813131313,
      "avg_output_tokens_per_task": 29913.797348484848,
      "last_run_timestamp": "2026-08-29T18:39:41.257Z"
    },
    {
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "SambaNova",
      "accuracy": 0.7659360351925629,
      "accuracy_std_dev": 0.01122486854878489,
      "run_count": 8,
      "total_questions": 1506,
      "avg_cost_per_task_usd": 0.024484757636122175,
      "avg_duration_per_task_ms": 73000.31474103585,
      "avg_output_tokens_per_task": 25656.97609561753,
      "last_run_timestamp": "2026-08-29T19:51:33.484Z"
    },
    {
      "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": 0.7653383396584439,
      "accuracy_std_dev": 0.03752103948599206,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.01665043263757116,
      "avg_duration_per_task_ms": 488848.29917773564,
      "avg_output_tokens_per_task": 16454.457305502845,
      "last_run_timestamp": "2026-08-30T00:18:59.365Z"
    },
    {
      "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": 0.7647061294117647,
      "accuracy_std_dev": 0.11983655888350603,
      "run_count": 5,
      "total_questions": 680,
      "avg_cost_per_task_usd": 0.015231964705882355,
      "avg_duration_per_task_ms": 558302.2632352941,
      "avg_output_tokens_per_task": 21406.642647058823,
      "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": "Parasail",
      "accuracy": 0.7644630358126723,
      "accuracy_std_dev": 0.017799283895356573,
      "run_count": 4,
      "total_questions": 726,
      "avg_cost_per_task_usd": 0.019927535812672176,
      "avg_duration_per_task_ms": 237713.65702479339,
      "avg_output_tokens_per_task": 26486.89256198347,
      "last_run_timestamp": "2026-08-29T18:11:51.470Z"
    },
    {
      "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": 0.7643024133099825,
      "accuracy_std_dev": 0.031491855215101075,
      "run_count": 6,
      "total_questions": 1142,
      "avg_cost_per_task_usd": 0.04750308056042031,
      "avg_duration_per_task_ms": 265287.42907180387,
      "avg_output_tokens_per_task": 17893.702276707532,
      "last_run_timestamp": "2026-09-26T09:16:55.239Z"
    },
    {
      "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": 0.7611532500000001,
      "accuracy_std_dev": 0.007114550894996836,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.017736026515151517,
      "avg_duration_per_task_ms": 86462.99116161616,
      "avg_output_tokens_per_task": 29343.33207070707,
      "last_run_timestamp": "2026-08-29T18:36:51.876Z"
    },
    {
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "AkashML",
      "accuracy": 0.760674964876033,
      "accuracy_std_dev": 0.01645419404771925,
      "run_count": 5,
      "total_questions": 968,
      "avg_cost_per_task_usd": 0.010708528925619836,
      "avg_duration_per_task_ms": 425619.78925619833,
      "avg_output_tokens_per_task": 25673.311983471074,
      "last_run_timestamp": "2026-08-29T18:56:40.728Z"
    },
    {
      "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": 0.7606652169197398,
      "accuracy_std_dev": 0.04591514075032818,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.004993729573391178,
      "avg_duration_per_task_ms": 317882.5820679682,
      "avg_output_tokens_per_task": 12242.755603759942,
      "last_run_timestamp": "2026-08-29T18:46:19.350Z"
    },
    {
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.75974,
      "accuracy_std_dev": 0.052236688581877,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0108920873015873,
      "avg_duration_per_task_ms": 734819.8246753247,
      "avg_output_tokens_per_task": 32324.756132756134,
      "last_run_timestamp": "2026-09-12T09:07:23.668Z"
    },
    {
      "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": "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Nebius",
      "accuracy": 0.758207125,
      "accuracy_std_dev": 0.00985024896179661,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.01902252525252525,
      "avg_duration_per_task_ms": 176301.23106060605,
      "avg_output_tokens_per_task": 31511.469696969696,
      "last_run_timestamp": "2026-08-29T20:01:12.707Z"
    },
    {
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Parasail",
      "accuracy": 0.7572815631067961,
      "accuracy_std_dev": 0.10286820253600068,
      "run_count": 9,
      "total_questions": 721,
      "avg_cost_per_task_usd": 0.002934966712898752,
      "avg_duration_per_task_ms": 194322.5145631068,
      "avg_output_tokens_per_task": 13374.00693481276,
      "last_run_timestamp": "2026-09-26T16:00:24.612Z"
    },
    {
      "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": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7559323627118644,
      "accuracy_std_dev": 0.007798453236949536,
      "run_count": 3,
      "total_questions": 590,
      "avg_cost_per_task_usd": 0.002668813559322034,
      "avg_duration_per_task_ms": 91653.82203389831,
      "avg_output_tokens_per_task": 31189.77966101695,
      "last_run_timestamp": "2026-09-01T13:23:13.492Z"
    },
    {
      "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": 0.7558922,
      "accuracy_std_dev": 0.10699962710663996,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.13022594646464647,
      "avg_duration_per_task_ms": 1023351.8464646465,
      "avg_output_tokens_per_task": 39743.50404040404,
      "last_run_timestamp": "2026-08-22T17:17:31.993Z"
    },
    {
      "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": 0.7558506982922201,
      "accuracy_std_dev": 0.013474818336988598,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.018912632194813407,
      "avg_duration_per_task_ms": 195928.35926628715,
      "avg_output_tokens_per_task": 31299.83870967742,
      "last_run_timestamp": "2026-08-29T16:18:21.978Z"
    },
    {
      "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": 0.7551707142857145,
      "accuracy_std_dev": 0.008190904148841594,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.02236280663780664,
      "avg_duration_per_task_ms": 98622.52741702742,
      "avg_output_tokens_per_task": 14763.670274170274,
      "last_run_timestamp": "2026-08-29T19:17:49.331Z"
    },
    {
      "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": 0.7548771259541985,
      "accuracy_std_dev": 0.012936525817791087,
      "run_count": 8,
      "total_questions": 1572,
      "avg_cost_per_task_usd": 0.016541959287531806,
      "avg_duration_per_task_ms": 307024.9662849873,
      "avg_output_tokens_per_task": 28293.188295165393,
      "last_run_timestamp": "2026-08-29T18:30:03.855Z"
    },
    {
      "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": 0.754696375,
      "accuracy_std_dev": 0.022698605518617125,
      "run_count": 7,
      "total_questions": 1384,
      "avg_cost_per_task_usd": 0.020891089595375723,
      "avg_duration_per_task_ms": 144594.61271676302,
      "avg_output_tokens_per_task": 12072.543352601157,
      "last_run_timestamp": "2026-08-22T17:12:08.919Z"
    },
    {
      "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": 0.7545347341659232,
      "accuracy_std_dev": 0.02761225844139473,
      "run_count": 8,
      "total_questions": 1121,
      "avg_cost_per_task_usd": 0.05582876717216771,
      "avg_duration_per_task_ms": 404528.79750223015,
      "avg_output_tokens_per_task": 16141.22033898305,
      "last_run_timestamp": "2026-09-19T08:53:38.031Z"
    },
    {
      "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": 0.754260582822086,
      "accuracy_std_dev": 0.0330868929772635,
      "run_count": 3,
      "total_questions": 489,
      "avg_cost_per_task_usd": 0.010235226993865029,
      "avg_duration_per_task_ms": 357931.47034764825,
      "avg_output_tokens_per_task": 31124.41717791411,
      "last_run_timestamp": "2026-07-18T08:04:08.574Z"
    },
    {
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.752765857142857,
      "accuracy_std_dev": 0.045870754858246085,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.05275206060606061,
      "avg_duration_per_task_ms": 979408.9415584416,
      "avg_output_tokens_per_task": 27425.04473304473,
      "last_run_timestamp": "2026-08-29T18:19:18.341Z"
    },
    {
      "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": 0.7525250000000001,
      "accuracy_std_dev": 0.010101000000000027,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.056976058080808085,
      "avg_duration_per_task_ms": 98919.65151515152,
      "avg_output_tokens_per_task": 22701.11111111111,
      "last_run_timestamp": "2026-09-01T17:54:55.806Z"
    },
    {
      "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": 0.751473,
      "accuracy_std_dev": 0.14287927657379149,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.03140934785353535,
      "avg_duration_per_task_ms": 145498.60984848486,
      "avg_output_tokens_per_task": 24064.433712121212,
      "last_run_timestamp": "2026-08-29T19:13:10.747Z"
    },
    {
      "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": 0.750842,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01581958585858586,
      "avg_duration_per_task_ms": 326624.7373737374,
      "avg_output_tokens_per_task": 13062.823232323231,
      "last_run_timestamp": "2026-09-12T09:42:09.531Z"
    },
    {
      "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": 0.75,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0067602800000000005,
      "avg_duration_per_task_ms": 1824454.58,
      "avg_output_tokens_per_task": 23925.293333333335,
      "last_run_timestamp": "2026-09-22T21:48:57.116Z"
    },
    {
      "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": 0.7493372537960954,
      "accuracy_std_dev": 0.017786308528225433,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.16125216847433116,
      "avg_duration_per_task_ms": 602672.0441070137,
      "avg_output_tokens_per_task": 50469.859725234994,
      "last_run_timestamp": "2026-08-29T18:15:24.940Z"
    },
    {
      "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": 0.7485970000000001,
      "accuracy_std_dev": 0.07972918647312704,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.006700525252525253,
      "avg_duration_per_task_ms": 93845.37542087543,
      "avg_output_tokens_per_task": 16482.319865319867,
      "last_run_timestamp": "2026-08-29T18:45:38.590Z"
    },
    {
      "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": 0.748106,
      "accuracy_std_dev": 0.006730599731821838,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.013436075126262624,
      "avg_duration_per_task_ms": 99232.33333333333,
      "avg_output_tokens_per_task": 14767.26452020202,
      "last_run_timestamp": "2026-08-29T18:17:28.541Z"
    },
    {
      "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": 0.7476619999999999,
      "accuracy_std_dev": 0.008415500381240979,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.014419837822671156,
      "avg_duration_per_task_ms": 105834.91358024691,
      "avg_output_tokens_per_task": 14297.092031425365,
      "last_run_timestamp": "2026-08-29T18:11:58.511Z"
    },
    {
      "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": 0.7475968772093022,
      "accuracy_std_dev": 0.030626747173151808,
      "run_count": 7,
      "total_questions": 1075,
      "avg_cost_per_task_usd": 0.006931252093023256,
      "avg_duration_per_task_ms": 969560.9004651163,
      "avg_output_tokens_per_task": 17890.426046511628,
      "last_run_timestamp": "2026-08-29T19:19:08.877Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7468043761061948,
      "accuracy_std_dev": 0.051672782249735434,
      "run_count": 7,
      "total_questions": 1356,
      "avg_cost_per_task_usd": 0.001709050884955752,
      "avg_duration_per_task_ms": 588006.3584070796,
      "avg_output_tokens_per_task": 11440.941002949852,
      "last_run_timestamp": "2026-08-22T17:47:50.063Z"
    },
    {
      "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7463394360465117,
      "accuracy_std_dev": 0.07095671544505143,
      "run_count": 8,
      "total_questions": 1548,
      "avg_cost_per_task_usd": 0.12446177519379846,
      "avg_duration_per_task_ms": 998515.9916020671,
      "avg_output_tokens_per_task": 55402.337855297155,
      "last_run_timestamp": "2026-08-29T18:23:09.250Z"
    },
    {
      "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": 0.7462122499999999,
      "accuracy_std_dev": 0.013006497229750218,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.017694700757575756,
      "avg_duration_per_task_ms": 99445.50378787878,
      "avg_output_tokens_per_task": 6284.411616161616,
      "last_run_timestamp": "2026-08-01T17:45:27.258Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "GMICloud",
      "accuracy": 0.7462122499999999,
      "accuracy_std_dev": 0.013222542538691302,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0095508547979798,
      "avg_duration_per_task_ms": 202963.16287878787,
      "avg_output_tokens_per_task": 8137.35101010101,
      "last_run_timestamp": "2026-08-29T18:16:24.161Z"
    },
    {
      "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": 0.7455108333333333,
      "accuracy_std_dev": 0.020304992673532824,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.08322730471380473,
      "avg_duration_per_task_ms": 907303.9099326599,
      "avg_output_tokens_per_task": 23781.781144781144,
      "last_run_timestamp": "2026-08-22T17:38:59.682Z"
    },
    {
      "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": 0.7449942438162545,
      "accuracy_std_dev": 0.01819967836712137,
      "run_count": 3,
      "total_questions": 283,
      "avg_cost_per_task_usd": 0.01679327561837456,
      "avg_duration_per_task_ms": 97120.16254416961,
      "avg_output_tokens_per_task": 9655.678445229682,
      "last_run_timestamp": "2026-08-30T00:01:09.485Z"
    },
    {
      "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": 0.7431723333333333,
      "accuracy_std_dev": 0.011056875789400136,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0071705791245791255,
      "avg_duration_per_task_ms": 175275.17003367003,
      "avg_output_tokens_per_task": 8334.754208754208,
      "last_run_timestamp": "2026-08-29T18:27:42.956Z"
    },
    {
      "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": 0.742845125,
      "accuracy_std_dev": 0.010902042072904301,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.022606026515151516,
      "avg_duration_per_task_ms": 99614.75883838384,
      "avg_output_tokens_per_task": 14962.358585858587,
      "last_run_timestamp": "2026-08-29T18:28:10.968Z"
    },
    {
      "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": "Nebius",
      "accuracy": 0.7426647142857141,
      "accuracy_std_dev": 0.007278926328092368,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0053309747474747475,
      "avg_duration_per_task_ms": 143246.7821067821,
      "avg_output_tokens_per_task": 8614.1443001443,
      "last_run_timestamp": "2026-08-29T18:49:21.941Z"
    },
    {
      "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": 0.7423279333333332,
      "accuracy_std_dev": 0.047881077561854994,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.007047760000000001,
      "avg_duration_per_task_ms": 1019395.5453968254,
      "avg_output_tokens_per_task": 18421.787936507935,
      "last_run_timestamp": "2026-08-29T20:11:32.700Z"
    },
    {
      "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": 0.7418629999999999,
      "accuracy_std_dev": 0.004827531943619518,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.01876800505050505,
      "avg_duration_per_task_ms": 32883.444444444445,
      "avg_output_tokens_per_task": 6023.489898989899,
      "last_run_timestamp": "2026-08-01T17:45:49.025Z"
    },
    {
      "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": 0.7413018888888889,
      "accuracy_std_dev": 0.010764979449270195,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.004904684624017957,
      "avg_duration_per_task_ms": 109945.37485970819,
      "avg_output_tokens_per_task": 7999.2755331088665,
      "last_run_timestamp": "2026-08-29T18:15:48.874Z"
    },
    {
      "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": 0.7402804187380497,
      "accuracy_std_dev": 0.076460867937462,
      "run_count": 8,
      "total_questions": 1569,
      "avg_cost_per_task_usd": 0.010684235181644361,
      "avg_duration_per_task_ms": 919191.3709369025,
      "avg_output_tokens_per_task": 11540.983428935628,
      "last_run_timestamp": "2026-08-30T01:20:44.266Z"
    },
    {
      "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": 0.7398376524390244,
      "accuracy_std_dev": 0.030475562908008763,
      "run_count": 5,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.004495964430894309,
      "avg_duration_per_task_ms": 781618.3089430894,
      "avg_output_tokens_per_task": 27377.12906504065,
      "last_run_timestamp": "2026-08-29T18:16:54.151Z"
    },
    {
      "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": 0.7392771671732522,
      "accuracy_std_dev": 0.07975374971554629,
      "run_count": 10,
      "total_questions": 1974,
      "avg_cost_per_task_usd": 0.005161332320162108,
      "avg_duration_per_task_ms": 757299.1337386018,
      "avg_output_tokens_per_task": 34332.25075987842,
      "last_run_timestamp": "2026-09-26T18:04:35.621Z"
    },
    {
      "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": 0.7388467499999999,
      "accuracy_std_dev": 0.07747406520692911,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.01745328409090909,
      "avg_duration_per_task_ms": 2275674.0732323234,
      "avg_output_tokens_per_task": 17131.02398989899,
      "last_run_timestamp": "2026-08-22T18:24:57.217Z"
    },
    {
      "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": 0.7384959999999999,
      "accuracy_std_dev": 0.0076941959727230055,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.058646202861952865,
      "avg_duration_per_task_ms": 123198.74579124579,
      "avg_output_tokens_per_task": 23489.968855218856,
      "last_run_timestamp": "2026-09-01T17:53:51.881Z"
    },
    {
      "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": 0.7380156260869567,
      "accuracy_std_dev": 0.012608807218939267,
      "run_count": 8,
      "total_questions": 1495,
      "avg_cost_per_task_usd": 0.008862770568561874,
      "avg_duration_per_task_ms": 132109.78528428092,
      "avg_output_tokens_per_task": 11237.499665551839,
      "last_run_timestamp": "2026-08-29T18:11:53.163Z"
    },
    {
      "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": 0.7374254884910486,
      "accuracy_std_dev": 0.06440394854096035,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.14750449019607842,
      "avg_duration_per_task_ms": 1116895.3742540495,
      "avg_output_tokens_per_task": 40733.769820971866,
      "last_run_timestamp": "2026-08-22T17:24:19.901Z"
    },
    {
      "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": 0.7373736666666667,
      "accuracy_std_dev": 0.017869435923448234,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.004179791245791245,
      "avg_duration_per_task_ms": 604833.3282828283,
      "avg_output_tokens_per_task": 30919.042087542086,
      "last_run_timestamp": "2026-07-18T16:28:34.215Z"
    },
    {
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Together",
      "accuracy": 0.73569,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.005349656565656565,
      "avg_duration_per_task_ms": 222485.2626262626,
      "avg_output_tokens_per_task": 8728.030303030304,
      "last_run_timestamp": "2026-07-04T17:56:44.437Z"
    },
    {
      "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": 0.7355811297709923,
      "accuracy_std_dev": 0.10861849313307564,
      "run_count": 8,
      "total_questions": 1572,
      "avg_cost_per_task_usd": 0.00556993893129771,
      "avg_duration_per_task_ms": 485638.58396946563,
      "avg_output_tokens_per_task": 24035.00954198473,
      "last_run_timestamp": "2026-09-26T09:37:20.718Z"
    },
    {
      "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": 0.7351291666666667,
      "accuracy_std_dev": 0.05591704578336457,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.015923739057239056,
      "avg_duration_per_task_ms": 844758.271885522,
      "avg_output_tokens_per_task": 41206.31481481482,
      "last_run_timestamp": "2026-08-29T19:32:52.576Z"
    },
    {
      "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": 0.7336406894705532,
      "accuracy_std_dev": 0.016215018082375493,
      "run_count": 17,
      "total_questions": 3362,
      "avg_cost_per_task_usd": 0.0069099631171921465,
      "avg_duration_per_task_ms": 93516.04104699583,
      "avg_output_tokens_per_task": 7210.586555621654,
      "last_run_timestamp": "2026-08-29T19:15:44.037Z"
    },
    {
      "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": 0.7314815,
      "accuracy_std_dev": 0.11424146120352568,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.13496367760942762,
      "avg_duration_per_task_ms": 1650116.6127946128,
      "avg_output_tokens_per_task": 45746.51094276094,
      "last_run_timestamp": "2026-08-29T18:41:58.319Z"
    },
    {
      "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": 0.7312871538461537,
      "accuracy_std_dev": 0.012238898554301565,
      "run_count": 26,
      "total_questions": 5148,
      "avg_cost_per_task_usd": 0.20442489957264962,
      "avg_duration_per_task_ms": 242818.47999223,
      "avg_output_tokens_per_task": 38189.194832944835,
      "last_run_timestamp": "2026-09-26T08:58:38.716Z"
    },
    {
      "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": 0.7309068406130269,
      "accuracy_std_dev": 0.06991914710924596,
      "run_count": 8,
      "total_questions": 1305,
      "avg_cost_per_task_usd": 0.03057965057471263,
      "avg_duration_per_task_ms": 649469.8950191571,
      "avg_output_tokens_per_task": 33236.169348659,
      "last_run_timestamp": "2026-08-29T18:51:07.939Z"
    },
    {
      "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": 0.7307611537001898,
      "accuracy_std_dev": 0.03501680507605419,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.009550828589500316,
      "avg_duration_per_task_ms": 215086.75015812778,
      "avg_output_tokens_per_task": 18410.464263124606,
      "last_run_timestamp": "2026-08-29T19:17:18.768Z"
    },
    {
      "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": 0.7307607964071856,
      "accuracy_std_dev": 0.09245964826175526,
      "run_count": 8,
      "total_questions": 1503,
      "avg_cost_per_task_usd": 0.14931497338656022,
      "avg_duration_per_task_ms": 776563.6879574185,
      "avg_output_tokens_per_task": 54559.805721889556,
      "last_run_timestamp": "2026-08-29T16:33:36.680Z"
    },
    {
      "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": 0.7300085000000001,
      "accuracy_std_dev": 0.01549688586458583,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.005400070075757575,
      "avg_duration_per_task_ms": 260007.1976010101,
      "avg_output_tokens_per_task": 8023.359217171717,
      "last_run_timestamp": "2026-08-29T18:12:05.151Z"
    },
    {
      "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": 0.7298248631578947,
      "accuracy_std_dev": 0.025121594911328576,
      "run_count": 3,
      "total_questions": 570,
      "avg_cost_per_task_usd": 0.004760936842105263,
      "avg_duration_per_task_ms": 667212.4614035088,
      "avg_output_tokens_per_task": 27657.705263157895,
      "last_run_timestamp": "2026-07-18T16:27:55.986Z"
    },
    {
      "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": 0.7297043333333333,
      "accuracy_std_dev": 0.017207018593843885,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.1851553911335578,
      "avg_duration_per_task_ms": 289835.51515151514,
      "avg_output_tokens_per_task": 36869.942199775534,
      "last_run_timestamp": "2026-09-26T09:21:00.998Z"
    },
    {
      "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": 0.729592,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.0030441326530612246,
      "avg_duration_per_task_ms": 200954.75,
      "avg_output_tokens_per_task": 31361.45408163265,
      "last_run_timestamp": "2026-09-01T13:20:55.717Z"
    },
    {
      "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": "Phala",
      "accuracy": 0.7290074809160306,
      "accuracy_std_dev": 0.07204749999999999,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.12394920356234096,
      "avg_duration_per_task_ms": 1193060.445292621,
      "avg_output_tokens_per_task": 40638.114503816796,
      "last_run_timestamp": "2026-07-11T16:02:27.787Z"
    },
    {
      "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": 0.7283226758169934,
      "accuracy_std_dev": 0.19151515448196257,
      "run_count": 4,
      "total_questions": 765,
      "avg_cost_per_task_usd": 0.019633686274509805,
      "avg_duration_per_task_ms": 252454.9568627451,
      "avg_output_tokens_per_task": 16304.679738562092,
      "last_run_timestamp": "2026-08-01T18:21:26.176Z"
    },
    {
      "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": 0.7274615892439593,
      "accuracy_std_dev": 0.08374107689388129,
      "run_count": 7,
      "total_questions": 1283,
      "avg_cost_per_task_usd": 0.030927738893219012,
      "avg_duration_per_task_ms": 152143.38581449728,
      "avg_output_tokens_per_task": 28647.951675759938,
      "last_run_timestamp": "2026-08-29T19:16:48.157Z"
    },
    {
      "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": 0.727459888888889,
      "accuracy_std_dev": 0.010021023926885385,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.1903540600448934,
      "avg_duration_per_task_ms": 241489.60381593715,
      "avg_output_tokens_per_task": 37880.08529741863,
      "last_run_timestamp": "2026-09-26T08:54:32.344Z"
    },
    {
      "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": 0.7272725,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.058774820707070706,
      "avg_duration_per_task_ms": 121424.5,
      "avg_output_tokens_per_task": 23508.14898989899,
      "last_run_timestamp": "2026-09-01T17:49:53.883Z"
    },
    {
      "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": 0.7269571249999999,
      "accuracy_std_dev": 0.006573317074306892,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0035786773989898986,
      "avg_duration_per_task_ms": 533271.5606060605,
      "avg_output_tokens_per_task": 8145.644570707071,
      "last_run_timestamp": "2026-08-29T18:39:40.767Z"
    },
    {
      "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": 0.7259116944971536,
      "accuracy_std_dev": 0.014910946473459525,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.00463467109424415,
      "avg_duration_per_task_ms": 214923.00189753322,
      "avg_output_tokens_per_task": 7865.01201771031,
      "last_run_timestamp": "2026-08-29T18:16:39.843Z"
    },
    {
      "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": 0.7245957898383372,
      "accuracy_std_dev": 0.009250125041530539,
      "run_count": 9,
      "total_questions": 1732,
      "avg_cost_per_task_usd": 0.1865489249422633,
      "avg_duration_per_task_ms": 350199.4688221709,
      "avg_output_tokens_per_task": 37119.776558891455,
      "last_run_timestamp": "2026-09-26T08:53:09.141Z"
    },
    {
      "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": 0.7241462857142857,
      "accuracy_std_dev": 0.008790615315929416,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.006113787157287157,
      "avg_duration_per_task_ms": 404593.2886002886,
      "avg_output_tokens_per_task": 8122.329725829726,
      "last_run_timestamp": "2026-08-22T16:50:17.574Z"
    },
    {
      "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": "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": "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": 0.722853625,
      "accuracy_std_dev": 0.04324929642473245,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.006985363636363637,
      "avg_duration_per_task_ms": 129043.11742424243,
      "avg_output_tokens_per_task": 7756.648358585859,
      "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": "Novita",
      "accuracy": 0.722222125,
      "accuracy_std_dev": 0.01884093724073658,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.006819729797979799,
      "avg_duration_per_task_ms": 350972.87373737374,
      "avg_output_tokens_per_task": 27084.810606060608,
      "last_run_timestamp": "2026-08-29T18:36:32.803Z"
    },
    {
      "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": 0.7219416666666667,
      "accuracy_std_dev": 0.05789083119880804,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.011145885521885522,
      "avg_duration_per_task_ms": 86502.09595959596,
      "avg_output_tokens_per_task": 4478.992424242424,
      "last_run_timestamp": "2026-09-26T09:45:38.864Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7211966144756279,
      "accuracy_std_dev": 0.09962827751441757,
      "run_count": 7,
      "total_questions": 1354,
      "avg_cost_per_task_usd": 0.04244912850812408,
      "avg_duration_per_task_ms": 1668691.0731166913,
      "avg_output_tokens_per_task": 17565.30502215657,
      "last_run_timestamp": "2026-08-29T20:16:12.181Z"
    },
    {
      "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": 0.7204764444444446,
      "accuracy_std_dev": 0.010281167113216102,
      "run_count": 27,
      "total_questions": 5346,
      "avg_cost_per_task_usd": 0.1972407962962963,
      "avg_duration_per_task_ms": 261625.21979049756,
      "avg_output_tokens_per_task": 36783.956041900485,
      "last_run_timestamp": "2026-09-26T09:21:19.221Z"
    },
    {
      "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": 0.71875,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 192,
      "avg_cost_per_task_usd": 0.014737661458333334,
      "avg_duration_per_task_ms": 950475.9635416666,
      "avg_output_tokens_per_task": 52272.296875,
      "last_run_timestamp": "2026-09-29T03:31:44.342Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Nebius",
      "accuracy": 0.7183463229974161,
      "accuracy_std_dev": 0.029264511720208542,
      "run_count": 6,
      "total_questions": 1161,
      "avg_cost_per_task_usd": 0.17074779758828595,
      "avg_duration_per_task_ms": 1524939.6339362618,
      "avg_output_tokens_per_task": 41593.74849267872,
      "last_run_timestamp": "2026-08-22T16:49:30.177Z"
    },
    {
      "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": 0.7170990662824207,
      "accuracy_std_dev": 0.10692249999999998,
      "run_count": 2,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.005610605187319885,
      "avg_duration_per_task_ms": 269295.4322766571,
      "avg_output_tokens_per_task": 27309.671469740635,
      "last_run_timestamp": "2026-09-01T13:46:27.106Z"
    },
    {
      "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": 0.7147435038461539,
      "accuracy_std_dev": 0.013460967196281816,
      "run_count": 8,
      "total_questions": 1560,
      "avg_cost_per_task_usd": 0.00823847564102564,
      "avg_duration_per_task_ms": 74470.71923076923,
      "avg_output_tokens_per_task": 10465.133974358974,
      "last_run_timestamp": "2026-08-29T18:12:18.859Z"
    },
    {
      "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": 0.7142555025380709,
      "accuracy_std_dev": 0.033662331581457615,
      "run_count": 4,
      "total_questions": 788,
      "avg_cost_per_task_usd": 0.0112786192893401,
      "avg_duration_per_task_ms": 562577.1395939087,
      "avg_output_tokens_per_task": 56130.946700507615,
      "last_run_timestamp": "2026-09-26T09:18:34.055Z"
    },
    {
      "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": 0.7132454451219512,
      "accuracy_std_dev": 0.06047841361477665,
      "run_count": 5,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.01395117174796748,
      "avg_duration_per_task_ms": 1210466.6656504066,
      "avg_output_tokens_per_task": 35695.419715447155,
      "last_run_timestamp": "2026-08-22T10:02:15.507Z"
    },
    {
      "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": 0.7118334872169467,
      "accuracy_std_dev": 0.05025937914195097,
      "run_count": 8,
      "total_questions": 1369,
      "avg_cost_per_task_usd": 0.00946499707815924,
      "avg_duration_per_task_ms": 383206.33820306795,
      "avg_output_tokens_per_task": 15939.81811541271,
      "last_run_timestamp": "2026-08-29T19:16:51.883Z"
    },
    {
      "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": 0.7100695976562501,
      "accuracy_std_dev": 0.008928047441035206,
      "run_count": 8,
      "total_questions": 1536,
      "avg_cost_per_task_usd": 0.01613465299479167,
      "avg_duration_per_task_ms": 90625.380859375,
      "avg_output_tokens_per_task": 10675.574869791666,
      "last_run_timestamp": "2026-08-29T19:12:26.066Z"
    },
    {
      "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": 0.708820908625731,
      "accuracy_std_dev": 0.016533293775325746,
      "run_count": 7,
      "total_questions": 1368,
      "avg_cost_per_task_usd": 0.010723485380116957,
      "avg_duration_per_task_ms": 104832.341374269,
      "avg_output_tokens_per_task": 9909.211988304094,
      "last_run_timestamp": "2026-08-29T19:10:03.342Z"
    },
    {
      "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": 0.708754,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01329097474747475,
      "avg_duration_per_task_ms": 223430.93434343435,
      "avg_output_tokens_per_task": 33132.76262626263,
      "last_run_timestamp": "2026-09-01T18:00:29.373Z"
    },
    {
      "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": 0.7077671902912621,
      "accuracy_std_dev": 0.13908047205911944,
      "run_count": 4,
      "total_questions": 515,
      "avg_cost_per_task_usd": 0.03580529126213592,
      "avg_duration_per_task_ms": 1364969.3572815533,
      "avg_output_tokens_per_task": 24706.64077669903,
      "last_run_timestamp": "2026-08-01T23:27:37.975Z"
    },
    {
      "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": 0.7075006595744683,
      "accuracy_std_dev": 0.010031238556847082,
      "run_count": 8,
      "total_questions": 1551,
      "avg_cost_per_task_usd": 0.011043542875564152,
      "avg_duration_per_task_ms": 107250.29013539651,
      "avg_output_tokens_per_task": 10405.526756931013,
      "last_run_timestamp": "2026-08-29T18:42:28.455Z"
    },
    {
      "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": 0.7074916250000001,
      "accuracy_std_dev": 0.009879437230145022,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.008883945707070707,
      "avg_duration_per_task_ms": 142295.2904040404,
      "avg_output_tokens_per_task": 5760.14898989899,
      "last_run_timestamp": "2026-08-29T18:10:22.043Z"
    },
    {
      "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": 0.7070005146036162,
      "accuracy_std_dev": 0.06882289073774799,
      "run_count": 4,
      "total_questions": 719,
      "avg_cost_per_task_usd": 0.07411282058414463,
      "avg_duration_per_task_ms": 612487.7343532684,
      "avg_output_tokens_per_task": 29410.726008344922,
      "last_run_timestamp": "2026-08-15T09:29:41.209Z"
    },
    {
      "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": 0.7067488941418294,
      "accuracy_std_dev": 0.11976383227569164,
      "run_count": 5,
      "total_questions": 973,
      "avg_cost_per_task_usd": 0.22542014594039053,
      "avg_duration_per_task_ms": 1724549.0164439876,
      "avg_output_tokens_per_task": 61812.13874614594,
      "last_run_timestamp": "2026-08-22T09:29:12.890Z"
    },
    {
      "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": 0.7059191574642126,
      "accuracy_std_dev": 0.05245735925848092,
      "run_count": 15,
      "total_questions": 2934,
      "avg_cost_per_task_usd": 0.008334460804362644,
      "avg_duration_per_task_ms": 468794.78834355826,
      "avg_output_tokens_per_task": 25898.645875937287,
      "last_run_timestamp": "2026-08-29T19:12:34.710Z"
    },
    {
      "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": 0.7055976250000001,
      "accuracy_std_dev": 0.010151381518511415,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.00599510542929293,
      "avg_duration_per_task_ms": 107464.72411616161,
      "avg_output_tokens_per_task": 5562.948863636364,
      "last_run_timestamp": "2026-08-29T19:09:55.303Z"
    },
    {
      "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": 0.7053707181996087,
      "accuracy_std_dev": 0.016091084149595845,
      "run_count": 8,
      "total_questions": 1533,
      "avg_cost_per_task_usd": 0.01048878082191781,
      "avg_duration_per_task_ms": 154483.25962165688,
      "avg_output_tokens_per_task": 9469.20156555773,
      "last_run_timestamp": "2026-08-29T18:09:55.969Z"
    },
    {
      "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": 0.7050869311010947,
      "accuracy_std_dev": 0.036824859327421404,
      "run_count": 8,
      "total_questions": 1553,
      "avg_cost_per_task_usd": 0.01090031036703155,
      "avg_duration_per_task_ms": 296639.77012234385,
      "avg_output_tokens_per_task": 47139.87057308435,
      "last_run_timestamp": "2026-09-26T08:54:38.846Z"
    },
    {
      "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": "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": "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": 0.7024411249999999,
      "accuracy_std_dev": 0.013087863819561053,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.01130629292929293,
      "avg_duration_per_task_ms": 202130.5946969697,
      "avg_output_tokens_per_task": 5888.216540404041,
      "last_run_timestamp": "2026-08-29T18:30:44.353Z"
    },
    {
      "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": 0.7002430000000002,
      "accuracy_std_dev": 0.010305162546564278,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0022877149270482604,
      "avg_duration_per_task_ms": 203069.73456790124,
      "avg_output_tokens_per_task": 5122.809203142537,
      "last_run_timestamp": "2026-08-29T16:33:36.891Z"
    },
    {
      "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": 0.6984128571428573,
      "accuracy_std_dev": 0.02323031510412064,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.011283265512265512,
      "avg_duration_per_task_ms": 64876.18759018759,
      "avg_output_tokens_per_task": 9342.030303030304,
      "last_run_timestamp": "2026-08-29T18:10:09.843Z"
    },
    {
      "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": 0.6979317142857143,
      "accuracy_std_dev": 0.03403895099405253,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.012199084415584417,
      "avg_duration_per_task_ms": 1274707.3217893217,
      "avg_output_tokens_per_task": 11028.597402597403,
      "last_run_timestamp": "2026-08-22T17:28:27.797Z"
    },
    {
      "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": 0.695926669535284,
      "accuracy_std_dev": 0.09652330217621027,
      "run_count": 6,
      "total_questions": 1162,
      "avg_cost_per_task_usd": 0.009071061101549054,
      "avg_duration_per_task_ms": 470826.79862306366,
      "avg_output_tokens_per_task": 31779.3339070568,
      "last_run_timestamp": "2026-08-29T20:02:35.387Z"
    },
    {
      "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": 0.69549675,
      "accuracy_std_dev": 0.017785276598566026,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.008984150252525254,
      "avg_duration_per_task_ms": 269134.8390151515,
      "avg_output_tokens_per_task": 6325.36047979798,
      "last_run_timestamp": "2026-08-29T18:14:24.357Z"
    },
    {
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Venice",
      "accuracy": 0.6940944351371386,
      "accuracy_std_dev": 0.0463087281397186,
      "run_count": 7,
      "total_questions": 1349,
      "avg_cost_per_task_usd": 0.005615137879911044,
      "avg_duration_per_task_ms": 396156.75315048185,
      "avg_output_tokens_per_task": 15382.206078576724,
      "last_run_timestamp": "2026-08-29T19:25:16.673Z"
    },
    {
      "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": 0.6931105699373695,
      "accuracy_std_dev": 0.08497842021164353,
      "run_count": 6,
      "total_questions": 958,
      "avg_cost_per_task_usd": 0.0069657160751565765,
      "avg_duration_per_task_ms": 481855.8267223382,
      "avg_output_tokens_per_task": 32574.692066805845,
      "last_run_timestamp": "2026-08-29T12:30:09.617Z"
    },
    {
      "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": 0.6917087499999999,
      "accuracy_std_dev": 0.2503720824541696,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.025392156565656566,
      "avg_duration_per_task_ms": 357863.4059343434,
      "avg_output_tokens_per_task": 15932.088383838383,
      "last_run_timestamp": "2026-08-29T19:03:14.914Z"
    },
    {
      "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": 0.6896746666666668,
      "accuracy_std_dev": 0.04960523518294776,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04329939057239058,
      "avg_duration_per_task_ms": 1345214.5555555555,
      "avg_output_tokens_per_task": 31899.07744107744,
      "last_run_timestamp": "2026-08-22T16:20:35.762Z"
    },
    {
      "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": "DeepInfra",
      "accuracy": 0.6890942823218998,
      "accuracy_std_dev": 0.1415725,
      "run_count": 2,
      "total_questions": 379,
      "avg_cost_per_task_usd": 0.020441316622691292,
      "avg_duration_per_task_ms": 430944.6226912929,
      "avg_output_tokens_per_task": 17733.345646437996,
      "last_run_timestamp": "2026-07-11T15:51:15.190Z"
    },
    {
      "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": 0.6885525,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.013761295454545455,
      "avg_duration_per_task_ms": 259532.94696969696,
      "avg_output_tokens_per_task": 32681.065656565657,
      "last_run_timestamp": "2026-09-01T18:15:18.821Z"
    },
    {
      "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": 0.6863611111111112,
      "accuracy_std_dev": 0.022997260124245744,
      "run_count": 6,
      "total_questions": 1134,
      "avg_cost_per_task_usd": 0.010325405643738976,
      "avg_duration_per_task_ms": 361986.35802469135,
      "avg_output_tokens_per_task": 46144.19135802469,
      "last_run_timestamp": "2026-09-26T08:53:22.738Z"
    },
    {
      "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": 0.6857475882352941,
      "accuracy_std_dev": 0.043243076491900706,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.014387779886148006,
      "avg_duration_per_task_ms": 244155.07653383934,
      "avg_output_tokens_per_task": 7771.313725490196,
      "last_run_timestamp": "2026-08-29T19:24:48.378Z"
    },
    {
      "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": "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.6824516220136518,
      "accuracy_std_dev": 0.08182193124974196,
      "run_count": 8,
      "total_questions": 1172,
      "avg_cost_per_task_usd": 0.03734705034129693,
      "avg_duration_per_task_ms": 243375.2090443686,
      "avg_output_tokens_per_task": 14147.406996587031,
      "last_run_timestamp": "2026-08-29T19:20:11.540Z"
    },
    {
      "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": 0.681700944587629,
      "accuracy_std_dev": 0.10151777573533613,
      "run_count": 4,
      "total_questions": 776,
      "avg_cost_per_task_usd": 0.061980561855670094,
      "avg_duration_per_task_ms": 552656.2654639175,
      "avg_output_tokens_per_task": 25817.090206185567,
      "last_run_timestamp": "2026-08-15T10:07:37.662Z"
    },
    {
      "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": "Friendli",
      "accuracy": 0.681187,
      "accuracy_std_dev": 0.0554640690582651,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.005700236111111112,
      "avg_duration_per_task_ms": 258962.23358585857,
      "avg_output_tokens_per_task": 6935.555555555556,
      "last_run_timestamp": "2026-08-01T17:52:16.799Z"
    },
    {
      "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": 0.678451,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0025699040404040404,
      "avg_duration_per_task_ms": 57532.25252525252,
      "avg_output_tokens_per_task": 2314.656565656566,
      "last_run_timestamp": "2026-09-01T13:44:51.554Z"
    },
    {
      "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": 0.678451,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01359871212121212,
      "avg_duration_per_task_ms": 222801.71212121213,
      "avg_output_tokens_per_task": 33900.580808080806,
      "last_run_timestamp": "2026-09-01T18:09:36.992Z"
    },
    {
      "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": 0.675926,
      "accuracy_std_dev": 0.009259000000000017,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0023789772727272726,
      "avg_duration_per_task_ms": 1250422.4444444445,
      "avg_output_tokens_per_task": 14190.916666666666,
      "last_run_timestamp": "2026-07-18T09:58:41.490Z"
    },
    {
      "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": 0.6757497120822621,
      "accuracy_std_dev": 0.20642627550360443,
      "run_count": 10,
      "total_questions": 1945,
      "avg_cost_per_task_usd": 0.014975519280205657,
      "avg_duration_per_task_ms": 371086.62262210797,
      "avg_output_tokens_per_task": 18143.37994858612,
      "last_run_timestamp": "2026-09-26T08:53:03.343Z"
    },
    {
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "WandB",
      "accuracy": 0.6717173333333334,
      "accuracy_std_dev": 0.009621770673264316,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.009016287878787879,
      "avg_duration_per_task_ms": 610741.1986531987,
      "avg_output_tokens_per_task": 69134.77777777778,
      "last_run_timestamp": "2026-07-18T17:04:25.343Z"
    },
    {
      "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": 0.6708755,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.014901020202020204,
      "avg_duration_per_task_ms": 14005.674242424242,
      "avg_output_tokens_per_task": 1670.229797979798,
      "last_run_timestamp": "2026-09-01T13:21:57.235Z"
    },
    {
      "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": 0.6700393558178753,
      "accuracy_std_dev": 0.03269417355990866,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.0023676661045531197,
      "avg_duration_per_task_ms": 206054.30691399664,
      "avg_output_tokens_per_task": 5153.2276559865095,
      "last_run_timestamp": "2026-08-29T17:07:45.402Z"
    },
    {
      "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": 0.6651728911651728,
      "accuracy_std_dev": 0.015545660212339019,
      "run_count": 8,
      "total_questions": 1562,
      "avg_cost_per_task_usd": 0.007795919974391807,
      "avg_duration_per_task_ms": 518511.552496799,
      "avg_output_tokens_per_task": 55598.74967989757,
      "last_run_timestamp": "2026-08-29T21:19:16.003Z"
    },
    {
      "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": 0.664983,
      "accuracy_std_dev": 0.045452036199346076,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.008639549663299663,
      "avg_duration_per_task_ms": 254582.55808080808,
      "avg_output_tokens_per_task": 42899.02777777778,
      "last_run_timestamp": "2026-09-26T09:38:02.897Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6613758571428571,
      "accuracy_std_dev": 0.010355400129243448,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.002499241702741702,
      "avg_duration_per_task_ms": 96661.21500721501,
      "avg_output_tokens_per_task": 2316.605339105339,
      "last_run_timestamp": "2026-08-29T20:54:16.286Z"
    },
    {
      "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": 0.6609084779270633,
      "accuracy_std_dev": 0.011670726056569915,
      "run_count": 8,
      "total_questions": 1563,
      "avg_cost_per_task_usd": 0.00226812412028151,
      "avg_duration_per_task_ms": 204254.08637236085,
      "avg_output_tokens_per_task": 3106.995521433141,
      "last_run_timestamp": "2026-08-29T18:47:21.578Z"
    },
    {
      "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": 0.02789490909090909,
      "avg_duration_per_task_ms": 918713.1623376623,
      "avg_output_tokens_per_task": 19192.68759018759,
      "last_run_timestamp": "2026-08-29T18:31:43.917Z"
    },
    {
      "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": 0.6583050357142856,
      "accuracy_std_dev": 0.015770411714311358,
      "run_count": 12,
      "total_questions": 2352,
      "avg_cost_per_task_usd": 0.005204377125850341,
      "avg_duration_per_task_ms": 107952.35246598639,
      "avg_output_tokens_per_task": 9805.454931972788,
      "last_run_timestamp": "2026-09-26T08:51:07.739Z"
    },
    {
      "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": 0.6582523611650485,
      "accuracy_std_dev": 0.011234936657031045,
      "run_count": 8,
      "total_questions": 1545,
      "avg_cost_per_task_usd": 0.002745144336569579,
      "avg_duration_per_task_ms": 78708.58122977347,
      "avg_output_tokens_per_task": 2915.294498381877,
      "last_run_timestamp": "2026-08-29T19:49:56.943Z"
    },
    {
      "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": "Wafer",
      "accuracy": 0.6578945157894737,
      "accuracy_std_dev": 0.06863835015499714,
      "run_count": 3,
      "total_questions": 570,
      "avg_cost_per_task_usd": 0.011776470175438595,
      "avg_duration_per_task_ms": 345542.28596491227,
      "avg_output_tokens_per_task": 29245.92456140351,
      "last_run_timestamp": "2026-07-18T08:04:42.524Z"
    },
    {
      "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": 0.6572561285289749,
      "accuracy_std_dev": 0.17360433671530034,
      "run_count": 8,
      "total_questions": 1346,
      "avg_cost_per_task_usd": 0.005431277860326895,
      "avg_duration_per_task_ms": 197404.2719167905,
      "avg_output_tokens_per_task": 6718.176820208024,
      "last_run_timestamp": "2026-08-29T19:47:54.565Z"
    },
    {
      "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": 0.6571234233128835,
      "accuracy_std_dev": 0.018221993227964927,
      "run_count": 15,
      "total_questions": 2934,
      "avg_cost_per_task_usd": 0.009085255623721884,
      "avg_duration_per_task_ms": 351031.9914792093,
      "avg_output_tokens_per_task": 60257.48977505112,
      "last_run_timestamp": "2026-08-29T22:02:09.071Z"
    },
    {
      "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": 0.6539546483050847,
      "accuracy_std_dev": 0.06028980821738355,
      "run_count": 3,
      "total_questions": 236,
      "avg_cost_per_task_usd": 0.09281549576271186,
      "avg_duration_per_task_ms": 1220256.3305084745,
      "avg_output_tokens_per_task": 32723.118644067796,
      "last_run_timestamp": "2026-08-01T18:16:39.158Z"
    },
    {
      "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": 0.653757604130809,
      "accuracy_std_dev": 0.015316614138466363,
      "run_count": 3,
      "total_questions": 581,
      "avg_cost_per_task_usd": 0.01179434079173838,
      "avg_duration_per_task_ms": 1792911.3149741825,
      "avg_output_tokens_per_task": 50132.19449225473,
      "last_run_timestamp": "2026-08-29T20:13:41.890Z"
    },
    {
      "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": 0.6529002334801761,
      "accuracy_std_dev": 0.09643724771974778,
      "run_count": 5,
      "total_questions": 908,
      "avg_cost_per_task_usd": 0.007793224669603524,
      "avg_duration_per_task_ms": 161069.71255506607,
      "avg_output_tokens_per_task": 9664.236784140969,
      "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-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": 0.6517065189340814,
      "accuracy_std_dev": 0.01961235432447518,
      "run_count": 11,
      "total_questions": 2139,
      "avg_cost_per_task_usd": 0.0068649700794763905,
      "avg_duration_per_task_ms": 223412.84338475924,
      "avg_output_tokens_per_task": 11118.252454417952,
      "last_run_timestamp": "2026-09-26T08:55:02.346Z"
    },
    {
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.6500593861209965,
      "accuracy_std_dev": 0.05876659054376805,
      "run_count": 6,
      "total_questions": 1124,
      "avg_cost_per_task_usd": 0.04040380338078292,
      "avg_duration_per_task_ms": 414098.88790035585,
      "avg_output_tokens_per_task": 20106.83896797153,
      "last_run_timestamp": "2026-08-22T10:29:25.897Z"
    },
    {
      "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": 0.6489900000000001,
      "accuracy_std_dev": 0.01599299999999998,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.015694313131313132,
      "avg_duration_per_task_ms": 13831.43686868687,
      "avg_output_tokens_per_task": 1769.391414141414,
      "last_run_timestamp": "2026-09-01T13:20:53.165Z"
    },
    {
      "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": 0.6478542063492063,
      "accuracy_std_dev": 0.03133768881430511,
      "run_count": 3,
      "total_questions": 567,
      "avg_cost_per_task_usd": 0.025561564373897706,
      "avg_duration_per_task_ms": 917973.6243386243,
      "avg_output_tokens_per_task": 30424.42680776014,
      "last_run_timestamp": "2026-08-01T17:59:38.872Z"
    },
    {
      "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": 0.6477275,
      "accuracy_std_dev": 0.015583769481418798,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.01589866666666667,
      "avg_duration_per_task_ms": 19920.939393939392,
      "avg_output_tokens_per_task": 1699.0972222222222,
      "last_run_timestamp": "2026-09-01T13:27:17.676Z"
    },
    {
      "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": 0.6475168750000001,
      "accuracy_std_dev": 0.1144747818369154,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.005135561868686868,
      "avg_duration_per_task_ms": 346231.56313131313,
      "avg_output_tokens_per_task": 5662.786616161616,
      "last_run_timestamp": "2026-08-29T19:11:19.710Z"
    },
    {
      "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": 0.6471702265758092,
      "accuracy_std_dev": 0.024540547409049426,
      "run_count": 9,
      "total_questions": 1761,
      "avg_cost_per_task_usd": 0.01087929528676888,
      "avg_duration_per_task_ms": 696427.9392390687,
      "avg_output_tokens_per_task": 56280.081771720616,
      "last_run_timestamp": "2026-08-29T21:14:10.017Z"
    },
    {
      "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": "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": 0.6461538384615385,
      "accuracy_std_dev": 0.010277671231242019,
      "run_count": 8,
      "total_questions": 1560,
      "avg_cost_per_task_usd": 0.0025117333333333335,
      "avg_duration_per_task_ms": 66425.87692307692,
      "avg_output_tokens_per_task": 2714.4269230769232,
      "last_run_timestamp": "2026-08-29T17:03:33.297Z"
    },
    {
      "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": 0.6459835714285715,
      "accuracy_std_dev": 0.011408025080832253,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0032850678210678214,
      "avg_duration_per_task_ms": 40715.905483405484,
      "avg_output_tokens_per_task": 3121.487012987013,
      "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": "DeepInfra",
      "accuracy": 0.6450217142857143,
      "accuracy_std_dev": 0.013022733909966664,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0023176284271284268,
      "avg_duration_per_task_ms": 69183.5432900433,
      "avg_output_tokens_per_task": 2807.774891774892,
      "last_run_timestamp": "2026-08-29T16:11:23.388Z"
    },
    {
      "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": 0.6446141374045801,
      "accuracy_std_dev": 0.005257999999999985,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.0025083384223918577,
      "avg_duration_per_task_ms": 104571.39694656488,
      "avg_output_tokens_per_task": 8089.860050890586,
      "last_run_timestamp": "2026-07-18T08:33:31.052Z"
    },
    {
      "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": 0.643518425,
      "accuracy_std_dev": 0.06758775688688005,
      "run_count": 4,
      "total_questions": 720,
      "avg_cost_per_task_usd": 0.028923294444444445,
      "avg_duration_per_task_ms": 727470.7069444444,
      "avg_output_tokens_per_task": 35275.89166666667,
      "last_run_timestamp": "2026-08-01T18:07:49.346Z"
    },
    {
      "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": 0.643390096687556,
      "accuracy_std_dev": 0.16884591783010003,
      "run_count": 6,
      "total_questions": 1117,
      "avg_cost_per_task_usd": 0.08375195344673231,
      "avg_duration_per_task_ms": 1334122.5317815577,
      "avg_output_tokens_per_task": 37138.017009847805,
      "last_run_timestamp": "2026-08-22T17:02:13.071Z"
    },
    {
      "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": 0.6431472852791877,
      "accuracy_std_dev": 0.05705437174474449,
      "run_count": 8,
      "total_questions": 985,
      "avg_cost_per_task_usd": 0.00601549847715736,
      "avg_duration_per_task_ms": 640442.6040609137,
      "avg_output_tokens_per_task": 40019.25786802031,
      "last_run_timestamp": "2026-08-29T17:24:11.557Z"
    },
    {
      "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": 0.6430122421874999,
      "accuracy_std_dev": 0.010735540414902267,
      "run_count": 8,
      "total_questions": 1536,
      "avg_cost_per_task_usd": 0.0024762552083333333,
      "avg_duration_per_task_ms": 124311.69205729167,
      "avg_output_tokens_per_task": 8736.557942708334,
      "last_run_timestamp": "2026-08-29T19:32:59.441Z"
    },
    {
      "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": 0.6414145,
      "accuracy_std_dev": 0.001683499999999949,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.004025537878787879,
      "avg_duration_per_task_ms": 31194.16414141414,
      "avg_output_tokens_per_task": 2323.563131313131,
      "last_run_timestamp": "2026-09-01T17:49:44.421Z"
    },
    {
      "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": 0.641414,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.004276166666666667,
      "avg_duration_per_task_ms": 30535.468434343435,
      "avg_output_tokens_per_task": 2347.165404040404,
      "last_run_timestamp": "2026-09-01T17:50:18.482Z"
    },
    {
      "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": 0.6410628652173913,
      "accuracy_std_dev": 0.012160856360823096,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.0027363427536231887,
      "avg_duration_per_task_ms": 152044.62463768115,
      "avg_output_tokens_per_task": 8852.080434782609,
      "last_run_timestamp": "2026-08-29T18:23:33.062Z"
    },
    {
      "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": 0.6410423139534884,
      "accuracy_std_dev": 0.013444699531204668,
      "run_count": 8,
      "total_questions": 1548,
      "avg_cost_per_task_usd": 0.004495651162790697,
      "avg_duration_per_task_ms": 73088.3649870801,
      "avg_output_tokens_per_task": 8450.452196382428,
      "last_run_timestamp": "2026-08-29T18:43:21.525Z"
    },
    {
      "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": 0.6408923801169591,
      "accuracy_std_dev": 0.018618251481661085,
      "run_count": 8,
      "total_questions": 1539,
      "avg_cost_per_task_usd": 0.00244093502274204,
      "avg_duration_per_task_ms": 207367.10331384017,
      "avg_output_tokens_per_task": 7914.494476933073,
      "last_run_timestamp": "2026-08-29T18:40:25.942Z"
    },
    {
      "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": 0.6408530000000001,
      "accuracy_std_dev": 0.00731669260435796,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.00720101430976431,
      "avg_duration_per_task_ms": 742062.9217171717,
      "avg_output_tokens_per_task": 55987.864478114476,
      "last_run_timestamp": "2026-08-29T20:02:26.710Z"
    },
    {
      "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": 0.6397305,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.004182681818181818,
      "avg_duration_per_task_ms": 33306.957070707074,
      "avg_output_tokens_per_task": 2424.3232323232323,
      "last_run_timestamp": "2026-09-01T17:50:13.815Z"
    },
    {
      "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": 0.6394904285714286,
      "accuracy_std_dev": 0.01660321769035958,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0016073823953823953,
      "avg_duration_per_task_ms": 115775.33910533911,
      "avg_output_tokens_per_task": 8123.611832611833,
      "last_run_timestamp": "2026-08-29T19:02:06.557Z"
    },
    {
      "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": 0.637803,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 179,
      "avg_cost_per_task_usd": 0.002030703910614525,
      "avg_duration_per_task_ms": 92260.82681564246,
      "avg_output_tokens_per_task": 6534.692737430168,
      "last_run_timestamp": "2026-07-18T09:17:11.788Z"
    },
    {
      "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": 0.6359762020460358,
      "accuracy_std_dev": 0.023468612837149117,
      "run_count": 4,
      "total_questions": 782,
      "avg_cost_per_task_usd": 0.006740489769820971,
      "avg_duration_per_task_ms": 885605.4833759591,
      "avg_output_tokens_per_task": 37192.53964194373,
      "last_run_timestamp": "2026-09-26T08:26:39.364Z"
    },
    {
      "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": 0.6352413333333333,
      "accuracy_std_dev": 0.011527714209195522,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.007043973063973064,
      "avg_duration_per_task_ms": 845264.6161616162,
      "avg_output_tokens_per_task": 50357.55892255892,
      "last_run_timestamp": "2026-08-29T22:11:02.681Z"
    },
    {
      "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": "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": 0.6337451904761905,
      "accuracy_std_dev": 0.08567913487347233,
      "run_count": 3,
      "total_questions": 567,
      "avg_cost_per_task_usd": 0.03312850793650794,
      "avg_duration_per_task_ms": 835240.3333333334,
      "avg_output_tokens_per_task": 32942.40211640212,
      "last_run_timestamp": "2026-08-01T18:05:37.080Z"
    },
    {
      "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": 0.6334175,
      "accuracy_std_dev": 0.007154500000000008,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.001171128787878788,
      "avg_duration_per_task_ms": 257331.09848484848,
      "avg_output_tokens_per_task": 11528.555555555555,
      "last_run_timestamp": "2026-08-29T19:38:59.849Z"
    },
    {
      "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": 0.6332621918976546,
      "accuracy_std_dev": 0.028783664609624945,
      "run_count": 8,
      "total_questions": 1407,
      "avg_cost_per_task_usd": 0.022304811656005688,
      "avg_duration_per_task_ms": 1028113.3994314143,
      "avg_output_tokens_per_task": 92736.27292110874,
      "last_run_timestamp": "2026-09-26T16:32:03.216Z"
    },
    {
      "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": 0.6330053363286265,
      "accuracy_std_dev": 0.03165786628011624,
      "run_count": 12,
      "total_questions": 2337,
      "avg_cost_per_task_usd": 0.0070095879332477545,
      "avg_duration_per_task_ms": 114460.44458707745,
      "avg_output_tokens_per_task": 9781.493367565254,
      "last_run_timestamp": "2026-09-26T09:22:12.174Z"
    },
    {
      "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": 0.6321081114302126,
      "accuracy_std_dev": 0.011032621401446443,
      "run_count": 9,
      "total_questions": 1741,
      "avg_cost_per_task_usd": 0.017073134979896612,
      "avg_duration_per_task_ms": 470536.2297530155,
      "avg_output_tokens_per_task": 85368.48765077542,
      "last_run_timestamp": "2026-09-26T16:13:51.663Z"
    },
    {
      "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": 0.6317296185567011,
      "accuracy_std_dev": 0.03997713513997718,
      "run_count": 3,
      "total_questions": 582,
      "avg_cost_per_task_usd": 0.02984942783505154,
      "avg_duration_per_task_ms": 817292.206185567,
      "avg_output_tokens_per_task": 36396.65120274914,
      "last_run_timestamp": "2026-08-01T18:25:56.196Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Alibaba",
      "accuracy": 0.6313133333333334,
      "accuracy_std_dev": 0.009621770673264316,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.012725252525252525,
      "avg_duration_per_task_ms": 428447.4292929293,
      "avg_output_tokens_per_task": 30647.838383838385,
      "last_run_timestamp": "2026-07-18T16:20:00.500Z"
    },
    {
      "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": "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": 0.6300085652173913,
      "accuracy_std_dev": 0.019291258106544862,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.013243481670929242,
      "avg_duration_per_task_ms": 617828.7595907928,
      "avg_output_tokens_per_task": 14030.67348678602,
      "last_run_timestamp": "2026-09-19T10:52:13.550Z"
    },
    {
      "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": 0.6297115344283006,
      "accuracy_std_dev": 0.021927190654171258,
      "run_count": 8,
      "total_questions": 1583,
      "avg_cost_per_task_usd": 0.017034094756790904,
      "avg_duration_per_task_ms": 638036.0972836387,
      "avg_output_tokens_per_task": 83517.71572962729,
      "last_run_timestamp": "2026-09-26T16:16:58.507Z"
    },
    {
      "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": 0.6294425555555555,
      "accuracy_std_dev": 0.01594050097575434,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0016517962962962962,
      "avg_duration_per_task_ms": 198352.56228956228,
      "avg_output_tokens_per_task": 9040.280583613918,
      "last_run_timestamp": "2026-08-29T17:04:59.562Z"
    },
    {
      "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": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.628788,
      "accuracy_std_dev": 0.01430999999999999,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017457550505050507,
      "avg_duration_per_task_ms": 1341424.7676767677,
      "avg_output_tokens_per_task": 31206.515151515152,
      "last_run_timestamp": "2026-07-18T17:46:02.370Z"
    },
    {
      "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": 0.626263,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.005995020202020202,
      "avg_duration_per_task_ms": 131491.52525252526,
      "avg_output_tokens_per_task": 48474.92929292929,
      "last_run_timestamp": "2026-09-28T18:12:26.736Z"
    },
    {
      "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": 0.6250000954545455,
      "accuracy_std_dev": 0.23227730753610482,
      "run_count": 3,
      "total_questions": 440,
      "avg_cost_per_task_usd": 0.01408210681818182,
      "avg_duration_per_task_ms": 344190.18863636366,
      "avg_output_tokens_per_task": 10377.513636363636,
      "last_run_timestamp": "2026-07-18T16:47:35.300Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Groq",
      "accuracy": 0.6248440193508115,
      "accuracy_std_dev": 0.023396789803541672,
      "run_count": 9,
      "total_questions": 1602,
      "avg_cost_per_task_usd": 0.007693858926342072,
      "avg_duration_per_task_ms": 37290.070536828964,
      "avg_output_tokens_per_task": 12740.69850187266,
      "last_run_timestamp": "2026-08-29T16:23:45.123Z"
    },
    {
      "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": 0.6248197142857143,
      "accuracy_std_dev": 0.11942661157585356,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.6020819559884559,
      "avg_duration_per_task_ms": 624611.841991342,
      "avg_output_tokens_per_task": 40107.266955266954,
      "last_run_timestamp": "2026-09-26T08:54:32.678Z"
    },
    {
      "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": 0.6235651617933724,
      "accuracy_std_dev": 0.15897481174775574,
      "run_count": 8,
      "total_questions": 1539,
      "avg_cost_per_task_usd": 0.14709008641975307,
      "avg_duration_per_task_ms": 1168622.465237167,
      "avg_output_tokens_per_task": 80206.91033138402,
      "last_run_timestamp": "2026-08-29T19:10:33.460Z"
    },
    {
      "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": 0.6218904017412935,
      "accuracy_std_dev": 0.10844110444806741,
      "run_count": 6,
      "total_questions": 804,
      "avg_cost_per_task_usd": 0.010471062189054726,
      "avg_duration_per_task_ms": 43908.85696517413,
      "avg_output_tokens_per_task": 6777.06592039801,
      "last_run_timestamp": "2026-08-22T17:34:08.243Z"
    },
    {
      "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": 0.6212123333333334,
      "accuracy_std_dev": 0.009622003960829715,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.00826737037037037,
      "avg_duration_per_task_ms": 370134.4276094276,
      "avg_output_tokens_per_task": 15814.858585858587,
      "last_run_timestamp": "2026-08-01T18:08:49.019Z"
    },
    {
      "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": 0.6180857142857143,
      "accuracy_std_dev": 0.07026707458326266,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.004998886002886003,
      "avg_duration_per_task_ms": 573044.753968254,
      "avg_output_tokens_per_task": 18652.590909090908,
      "last_run_timestamp": "2026-08-22T22:18:10.299Z"
    },
    {
      "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": 0.6170035,
      "accuracy_std_dev": 0.020976392200519146,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.005440873737373737,
      "avg_duration_per_task_ms": 588685.2001262626,
      "avg_output_tokens_per_task": 18061.924873737375,
      "last_run_timestamp": "2026-08-29T19:06:01.882Z"
    },
    {
      "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": 0.6166527916666666,
      "accuracy_std_dev": 0.12145770743554425,
      "run_count": 24,
      "total_questions": 4752,
      "avg_cost_per_task_usd": 0.6553831144781144,
      "avg_duration_per_task_ms": 484995.4349747475,
      "avg_output_tokens_per_task": 40791.21506734007,
      "last_run_timestamp": "2026-09-26T09:11:13.767Z"
    },
    {
      "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": 0.6165192925559605,
      "accuracy_std_dev": 0.018521943327847644,
      "run_count": 10,
      "total_questions": 1921,
      "avg_cost_per_task_usd": 0.00823595002602811,
      "avg_duration_per_task_ms": 513809.8833940656,
      "avg_output_tokens_per_task": 13523.150963040083,
      "last_run_timestamp": "2026-09-26T09:22:23.002Z"
    },
    {
      "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": 0.614478375,
      "accuracy_std_dev": 0.12366286539715297,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.6154596931818181,
      "avg_duration_per_task_ms": 632046.8207070708,
      "avg_output_tokens_per_task": 40881.84659090909,
      "last_run_timestamp": "2026-09-26T09:34:52.233Z"
    },
    {
      "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": 0.6139495138993238,
      "accuracy_std_dev": 0.029024384950147974,
      "run_count": 8,
      "total_questions": 1331,
      "avg_cost_per_task_usd": 0.02793150338091661,
      "avg_duration_per_task_ms": 583500.4891059354,
      "avg_output_tokens_per_task": 29316.752817430504,
      "last_run_timestamp": "2026-09-26T16:19:07.764Z"
    },
    {
      "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": 0.6129581867007673,
      "accuracy_std_dev": 0.0015394999999999714,
      "run_count": 2,
      "total_questions": 391,
      "avg_cost_per_task_usd": 0.0032793452685421997,
      "avg_duration_per_task_ms": 346994.51406649617,
      "avg_output_tokens_per_task": 13256.76726342711,
      "last_run_timestamp": "2026-09-01T13:34:17.711Z"
    },
    {
      "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": 0.6126859920896506,
      "accuracy_std_dev": 0.1240297456369234,
      "run_count": 23,
      "total_questions": 4551,
      "avg_cost_per_task_usd": 0.6596944438584925,
      "avg_duration_per_task_ms": 597097.6884201275,
      "avg_output_tokens_per_task": 41139.12920237311,
      "last_run_timestamp": "2026-09-26T09:32:36.461Z"
    },
    {
      "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": 0.6111113258426966,
      "accuracy_std_dev": 0.003763737002265482,
      "run_count": 3,
      "total_questions": 534,
      "avg_cost_per_task_usd": 0.009120874531835207,
      "avg_duration_per_task_ms": 904452.6797752809,
      "avg_output_tokens_per_task": 16552.734082397004,
      "last_run_timestamp": "2026-07-18T08:03:56.103Z"
    },
    {
      "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": 0.6104652558139534,
      "accuracy_std_dev": 0.010142144016429655,
      "run_count": 8,
      "total_questions": 1548,
      "avg_cost_per_task_usd": 0.0036393624031007747,
      "avg_duration_per_task_ms": 34858.98837209302,
      "avg_output_tokens_per_task": 2148.659560723514,
      "last_run_timestamp": "2026-08-29T18:49:22.541Z"
    },
    {
      "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": 0.6103109080932785,
      "accuracy_std_dev": 0.006004573355992887,
      "run_count": 8,
      "total_questions": 1458,
      "avg_cost_per_task_usd": 0.00404280109739369,
      "avg_duration_per_task_ms": 46573.90877914952,
      "avg_output_tokens_per_task": 2150.791495198903,
      "last_run_timestamp": "2026-08-29T18:43:09.218Z"
    },
    {
      "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": 0.609428,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0023919898989898988,
      "avg_duration_per_task_ms": 73719.80808080808,
      "avg_output_tokens_per_task": 2205.409090909091,
      "last_run_timestamp": "2026-09-01T13:27:58.636Z"
    },
    {
      "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": 0.608871193548387,
      "accuracy_std_dev": 0.0048793077890209365,
      "run_count": 4,
      "total_questions": 744,
      "avg_cost_per_task_usd": 0.007782237903225806,
      "avg_duration_per_task_ms": 296799.64919354836,
      "avg_output_tokens_per_task": 15367.672043010753,
      "last_run_timestamp": "2026-08-01T18:16:00.374Z"
    },
    {
      "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": 0.6087965,
      "accuracy_std_dev": 0.12831918473770787,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.6162571098484848,
      "avg_duration_per_task_ms": 633013.7184343435,
      "avg_output_tokens_per_task": 41027.7380050505,
      "last_run_timestamp": "2026-09-26T09:34:56.525Z"
    },
    {
      "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": 0.6085801047835991,
      "accuracy_std_dev": 0.25326376299160264,
      "run_count": 3,
      "total_questions": 439,
      "avg_cost_per_task_usd": 0.01201555353075171,
      "avg_duration_per_task_ms": 304422.4123006834,
      "avg_output_tokens_per_task": 10169.57630979499,
      "last_run_timestamp": "2026-07-18T16:53:38.018Z"
    },
    {
      "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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "auto-routing",
      "accuracy": 0.6065783472485767,
      "accuracy_std_dev": 0.011150535787995776,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.00814792915876028,
      "avg_duration_per_task_ms": 295371.79759645794,
      "avg_output_tokens_per_task": 17556.894370651487,
      "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": "Alibaba",
      "accuracy": 0.6060606666666666,
      "accuracy_std_dev": 0.0013745720142000956,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.008301892255892255,
      "avg_duration_per_task_ms": 153838.79292929292,
      "avg_output_tokens_per_task": 15775.223905723906,
      "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-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Venice",
      "accuracy": 0.605886368522073,
      "accuracy_std_dev": 0.02016173504399793,
      "run_count": 8,
      "total_questions": 1563,
      "avg_cost_per_task_usd": 0.07002656621880997,
      "avg_duration_per_task_ms": 1050392.2853486885,
      "avg_output_tokens_per_task": 26862.438899552144,
      "last_run_timestamp": "2026-08-30T00:22:07.120Z"
    },
    {
      "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": 0.6058202261904763,
      "accuracy_std_dev": 0.04086605084220276,
      "run_count": 8,
      "total_questions": 1512,
      "avg_cost_per_task_usd": 0.022986974867724868,
      "avg_duration_per_task_ms": 1010656.5641534391,
      "avg_output_tokens_per_task": 59799.07671957672,
      "last_run_timestamp": "2026-08-30T00:32:50.853Z"
    },
    {
      "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": 0.6053392857142857,
      "accuracy_std_dev": 0.004486622487678077,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.01102401875901876,
      "avg_duration_per_task_ms": 34344.32323232323,
      "avg_output_tokens_per_task": 2100.313131313131,
      "last_run_timestamp": "2026-08-29T18:32:24.616Z"
    },
    {
      "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": 0.6047306974789916,
      "accuracy_std_dev": 0.2173110611308259,
      "run_count": 6,
      "total_questions": 1071,
      "avg_cost_per_task_usd": 0.059062111111111114,
      "avg_duration_per_task_ms": 1034544.568627451,
      "avg_output_tokens_per_task": 19772.375350140057,
      "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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "WandB",
      "accuracy": 0.604539935483871,
      "accuracy_std_dev": 0.0028258301828351473,
      "run_count": 3,
      "total_questions": 558,
      "avg_cost_per_task_usd": 0.004098256272401434,
      "avg_duration_per_task_ms": 59405.28315412186,
      "avg_output_tokens_per_task": 2189.059139784946,
      "last_run_timestamp": "2026-07-18T08:26:39.830Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "NextBit",
      "accuracy": 0.6035355000000001,
      "accuracy_std_dev": 0.004208499999999948,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.002990320707070707,
      "avg_duration_per_task_ms": 555382.6085858586,
      "avg_output_tokens_per_task": 13221.70707070707,
      "last_run_timestamp": "2026-09-01T13:04:03.946Z"
    },
    {
      "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": 0.6026938333333334,
      "accuracy_std_dev": 0.05444530556260618,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.011141401515151514,
      "avg_duration_per_task_ms": 1165344.244949495,
      "avg_output_tokens_per_task": 32945.469696969696,
      "last_run_timestamp": "2026-08-29T19:15:54.652Z"
    },
    {
      "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": 0.6025891961165049,
      "accuracy_std_dev": 0.006394638535523308,
      "run_count": 8,
      "total_questions": 1545,
      "avg_cost_per_task_usd": 0.0033181844660194176,
      "avg_duration_per_task_ms": 55680.37928802589,
      "avg_output_tokens_per_task": 2110.4776699029126,
      "last_run_timestamp": "2026-08-29T18:47:46.000Z"
    },
    {
      "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": 0.6014871513560806,
      "accuracy_std_dev": 0.0142573356060427,
      "run_count": 6,
      "total_questions": 1143,
      "avg_cost_per_task_usd": 0.002170363954505687,
      "avg_duration_per_task_ms": 45118.168853893265,
      "avg_output_tokens_per_task": 2030.7734033245845,
      "last_run_timestamp": "2026-08-29T18:37:17.171Z"
    },
    {
      "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": 0.6005190778210118,
      "accuracy_std_dev": 0.014559222441784965,
      "run_count": 8,
      "total_questions": 1542,
      "avg_cost_per_task_usd": 0.00329739364461738,
      "avg_duration_per_task_ms": 57092.898184176396,
      "avg_output_tokens_per_task": 2093.910505836576,
      "last_run_timestamp": "2026-08-29T19:00:24.068Z"
    },
    {
      "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": 0.6004493333333333,
      "accuracy_std_dev": 0.0031744380430067634,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.014951090909090909,
      "avg_duration_per_task_ms": 287967.1666666667,
      "avg_output_tokens_per_task": 12268.760942760942,
      "last_run_timestamp": "2026-08-15T15:26:03.892Z"
    },
    {
      "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": 0.5946250426829269,
      "accuracy_std_dev": 0.04090406951159473,
      "run_count": 8,
      "total_questions": 1476,
      "avg_cost_per_task_usd": 0.09460423780487806,
      "avg_duration_per_task_ms": 840149.4261517616,
      "avg_output_tokens_per_task": 30270.114498644987,
      "last_run_timestamp": "2026-08-29T19:11:06.196Z"
    },
    {
      "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": 0.594276,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.002775611111111111,
      "avg_duration_per_task_ms": 509237.26767676766,
      "avg_output_tokens_per_task": 10972.9898989899,
      "last_run_timestamp": "2026-09-01T13:31:26.819Z"
    },
    {
      "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": 0.5922561000000001,
      "accuracy_std_dev": 0.11721268776753647,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.6146375166666666,
      "avg_duration_per_task_ms": 629521.5186868687,
      "avg_output_tokens_per_task": 40864.73838383838,
      "last_run_timestamp": "2026-09-26T09:23:13.151Z"
    },
    {
      "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": 0.5900675,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017035757575757576,
      "avg_duration_per_task_ms": 123031.02272727272,
      "avg_output_tokens_per_task": 42470.74747474748,
      "last_run_timestamp": "2026-09-01T18:10:49.087Z"
    },
    {
      "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": 0.5832802341650671,
      "accuracy_std_dev": 0.015118863604368572,
      "run_count": 8,
      "total_questions": 1563,
      "avg_cost_per_task_usd": 0.006713769673704416,
      "avg_duration_per_task_ms": 723452,
      "avg_output_tokens_per_task": 44499.71785028791,
      "last_run_timestamp": "2026-08-29T16:53:28.602Z"
    },
    {
      "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": 0.5778790920852359,
      "accuracy_std_dev": 0.14127804644250064,
      "run_count": 8,
      "total_questions": 1314,
      "avg_cost_per_task_usd": 0.014400366057838659,
      "avg_duration_per_task_ms": 93590.71461187214,
      "avg_output_tokens_per_task": 47817.760273972606,
      "last_run_timestamp": "2026-08-29T19:48:18.157Z"
    },
    {
      "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": 0.577862,
      "accuracy_std_dev": 0.03661599999999998,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.006390335858585858,
      "avg_duration_per_task_ms": 667369.1111111111,
      "avg_output_tokens_per_task": 32604.280303030304,
      "last_run_timestamp": "2026-09-26T08:13:27.134Z"
    },
    {
      "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": 0.5765865295404814,
      "accuracy_std_dev": 0.016216367287404426,
      "run_count": 3,
      "total_questions": 457,
      "avg_cost_per_task_usd": 0.0037329715536105032,
      "avg_duration_per_task_ms": 2186448.089715536,
      "avg_output_tokens_per_task": 36169.22100656455,
      "last_run_timestamp": "2026-07-18T16:00:09.483Z"
    },
    {
      "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": 0.5763891250000002,
      "accuracy_std_dev": 0.016770417924410086,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.010578772727272727,
      "avg_duration_per_task_ms": 12384.83143939394,
      "avg_output_tokens_per_task": 13650.558712121212,
      "last_run_timestamp": "2026-08-29T18:50:56.950Z"
    },
    {
      "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": 0.5749159999999999,
      "accuracy_std_dev": 0.02739257190371141,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0045310864898989904,
      "avg_duration_per_task_ms": 339529.7127525252,
      "avg_output_tokens_per_task": 15954.368055555555,
      "last_run_timestamp": "2026-08-29T19:16:45.297Z"
    },
    {
      "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": 0.5646441153846153,
      "accuracy_std_dev": 0.09153255016108605,
      "run_count": 11,
      "total_questions": 1846,
      "avg_cost_per_task_usd": 0.009133297399783316,
      "avg_duration_per_task_ms": 547859.6289274106,
      "avg_output_tokens_per_task": 45474.855904658725,
      "last_run_timestamp": "2026-09-26T16:17:38.509Z"
    },
    {
      "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": 0.5637339187817259,
      "accuracy_std_dev": 0.03764953294383463,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.025736050761421318,
      "avg_duration_per_task_ms": 1231284.9746192894,
      "avg_output_tokens_per_task": 65120.38747884941,
      "last_run_timestamp": "2026-08-30T00:49:28.653Z"
    },
    {
      "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": 0.55303,
      "accuracy_std_dev": 0.061447999999999975,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.006801247474747475,
      "avg_duration_per_task_ms": 137063.60858585857,
      "avg_output_tokens_per_task": 7268.926767676768,
      "last_run_timestamp": "2026-09-01T13:21:33.153Z"
    },
    {
      "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": 0.5521092952853598,
      "accuracy_std_dev": 0.2509952963755828,
      "run_count": 3,
      "total_questions": 403,
      "avg_cost_per_task_usd": 0.04710845161290323,
      "avg_duration_per_task_ms": 665624.1687344913,
      "avg_output_tokens_per_task": 19826.736972704715,
      "last_run_timestamp": "2026-07-18T16:43:10.588Z"
    },
    {
      "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": 0.5514706253196932,
      "accuracy_std_dev": 0.20217034155010072,
      "run_count": 8,
      "total_questions": 1564,
      "avg_cost_per_task_usd": 0.02285590537084399,
      "avg_duration_per_task_ms": 745462.75,
      "avg_output_tokens_per_task": 32123.599104859335,
      "last_run_timestamp": "2026-09-26T16:19:31.961Z"
    },
    {
      "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": 0.5496635000000001,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017014479797979796,
      "avg_duration_per_task_ms": 123100.59595959596,
      "avg_output_tokens_per_task": 42507.18939393939,
      "last_run_timestamp": "2026-09-01T17:51:18.914Z"
    },
    {
      "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": 0.5478889309989701,
      "accuracy_std_dev": 0.06042322607462552,
      "run_count": 7,
      "total_questions": 971,
      "avg_cost_per_task_usd": 0.053327887744593204,
      "avg_duration_per_task_ms": 699757.6158599382,
      "avg_output_tokens_per_task": 25522.51184346035,
      "last_run_timestamp": "2026-08-29T19:07:49.385Z"
    },
    {
      "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": 0.5476354460285131,
      "accuracy_std_dev": 0.04258471863891501,
      "run_count": 9,
      "total_questions": 1473,
      "avg_cost_per_task_usd": 0.016013412763068568,
      "avg_duration_per_task_ms": 701597.7447386286,
      "avg_output_tokens_per_task": 40502.957230142565,
      "last_run_timestamp": "2026-08-30T00:11:29.373Z"
    },
    {
      "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": 0.5433505,
      "accuracy_std_dev": 0.016294800251920854,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.017506126262626265,
      "avg_duration_per_task_ms": 128742.19065656565,
      "avg_output_tokens_per_task": 43717.62373737374,
      "last_run_timestamp": "2026-09-01T18:06:03.895Z"
    },
    {
      "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": 0.5431397499999999,
      "accuracy_std_dev": 0.09309990509897151,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.00908125946969697,
      "avg_duration_per_task_ms": 307947.68244949495,
      "avg_output_tokens_per_task": 26991.902777777777,
      "last_run_timestamp": "2026-08-29T17:01:22.015Z"
    },
    {
      "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": 0.5429295000000001,
      "accuracy_std_dev": 0.010942500000000022,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.001760580808080808,
      "avg_duration_per_task_ms": 80209.29040404041,
      "avg_output_tokens_per_task": 1718.6565656565656,
      "last_run_timestamp": "2026-09-01T13:31:30.699Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.5388051381909548,
      "accuracy_std_dev": 0.02361431811919149,
      "run_count": 8,
      "total_questions": 1194,
      "avg_cost_per_task_usd": 0.09965015829145729,
      "avg_duration_per_task_ms": 888064.2537688442,
      "avg_output_tokens_per_task": 28386.65661641541,
      "last_run_timestamp": "2026-08-29T18:52:09.401Z"
    },
    {
      "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": 0.5367792126099705,
      "accuracy_std_dev": 0.08908980723622514,
      "run_count": 7,
      "total_questions": 1364,
      "avg_cost_per_task_usd": 0.02643020747800586,
      "avg_duration_per_task_ms": 2484462.840175953,
      "avg_output_tokens_per_task": 22046.382697947214,
      "last_run_timestamp": "2026-08-22T18:19:05.094Z"
    },
    {
      "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": 0.53367,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01864569696969697,
      "avg_duration_per_task_ms": 17970.575757575756,
      "avg_output_tokens_per_task": 1670.050505050505,
      "last_run_timestamp": "2026-09-01T13:20:31.881Z"
    },
    {
      "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": 0.5317058333333333,
      "accuracy_std_dev": 0.3019237221956437,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.02117137962962963,
      "avg_duration_per_task_ms": 300182.6590909091,
      "avg_output_tokens_per_task": 22447.186026936026,
      "last_run_timestamp": "2026-08-15T15:24:35.389Z"
    },
    {
      "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": 0.530724,
      "accuracy_std_dev": 0.006341146150342195,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.005484556818181818,
      "avg_duration_per_task_ms": 124126.7058080808,
      "avg_output_tokens_per_task": 11845.058080808081,
      "last_run_timestamp": "2026-08-01T17:48:17.532Z"
    },
    {
      "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": 0.521886,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.016360343434343436,
      "avg_duration_per_task_ms": 13511.040404040405,
      "avg_output_tokens_per_task": 1443.3737373737374,
      "last_run_timestamp": "2026-09-01T14:02:01.624Z"
    },
    {
      "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": "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": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "OpenAI",
      "accuracy": 0.517396,
      "accuracy_std_dev": 0.02012409515978297,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.015192781144781144,
      "avg_duration_per_task_ms": 14836.205387205388,
      "avg_output_tokens_per_task": 1327.0218855218855,
      "last_run_timestamp": "2026-09-01T13:22:30.260Z"
    },
    {
      "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": 0.516835,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.018182969696969697,
      "avg_duration_per_task_ms": 17464.777777777777,
      "avg_output_tokens_per_task": 1627.171717171717,
      "last_run_timestamp": "2026-09-01T13:20:49.450Z"
    },
    {
      "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": "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": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "Azure",
      "accuracy": 0.511785,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01733130303030303,
      "avg_duration_per_task_ms": 13903.823232323231,
      "avg_output_tokens_per_task": 1536.9141414141413,
      "last_run_timestamp": "2026-09-01T13:47:27.116Z"
    },
    {
      "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": 0.50968,
      "accuracy_std_dev": 0.006561008459070931,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.00660445202020202,
      "avg_duration_per_task_ms": 165597.95580808082,
      "avg_output_tokens_per_task": 13339.318181818182,
      "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": "DeepInfra",
      "accuracy": 0.5095547757619738,
      "accuracy_std_dev": 0.07224117651877167,
      "run_count": 7,
      "total_questions": 1378,
      "avg_cost_per_task_usd": 0.003949532656023222,
      "avg_duration_per_task_ms": 149830.31567489114,
      "avg_output_tokens_per_task": 4238.851959361394,
      "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-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": 0.5092595,
      "accuracy_std_dev": 0.009259499999999976,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0008195479797979797,
      "avg_duration_per_task_ms": 19948.4797979798,
      "avg_output_tokens_per_task": 1859.0176767676767,
      "last_run_timestamp": "2026-09-01T17:50:32.282Z"
    },
    {
      "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-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": "Azure",
      "accuracy": 0.5033673333333334,
      "accuracy_std_dev": 0.009622237242739144,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0008440841750841751,
      "avg_duration_per_task_ms": 14338.025252525253,
      "avg_output_tokens_per_task": 1851.915824915825,
      "last_run_timestamp": "2026-09-01T17:51:18.903Z"
    },
    {
      "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": 0.503367,
      "accuracy_std_dev": 0.008418000000000037,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.013204560606060607,
      "avg_duration_per_task_ms": 13084.876262626263,
      "avg_output_tokens_per_task": 1127.6338383838383,
      "last_run_timestamp": "2026-09-01T13:20:22.396Z"
    },
    {
      "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": 0.4964458888888889,
      "accuracy_std_dev": 0.015601016017087856,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.00236331088664422,
      "avg_duration_per_task_ms": 11415.718294051627,
      "avg_output_tokens_per_task": 2344.6705948372614,
      "last_run_timestamp": "2026-08-29T19:15:06.404Z"
    },
    {
      "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": 0.4882155000000001,
      "accuracy_std_dev": 0.016835500000000003,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0008046262626262627,
      "avg_duration_per_task_ms": 20667.484848484848,
      "avg_output_tokens_per_task": 1820.4444444444443,
      "last_run_timestamp": "2026-09-01T17:50:43.769Z"
    },
    {
      "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": 0.4863289203860072,
      "accuracy_std_dev": 0.11980845067942446,
      "run_count": 7,
      "total_questions": 829,
      "avg_cost_per_task_usd": 0.031845395657418576,
      "avg_duration_per_task_ms": 57849.05307599517,
      "avg_output_tokens_per_task": 6810.792521109771,
      "last_run_timestamp": "2026-08-29T16:54:06.286Z"
    },
    {
      "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": 0.48477567101648344,
      "accuracy_std_dev": 0.036899227161410524,
      "run_count": 8,
      "total_questions": 1456,
      "avg_cost_per_task_usd": 0.011051652472527472,
      "avg_duration_per_task_ms": 444726.33997252746,
      "avg_output_tokens_per_task": 25860.356456043955,
      "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": 0.4834454444444444,
      "accuracy_std_dev": 0.005804812220354784,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0023795224466891134,
      "avg_duration_per_task_ms": 13393.97418630752,
      "avg_output_tokens_per_task": 2241.2485970819303,
      "last_run_timestamp": "2026-08-29T16:43:47.747Z"
    },
    {
      "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": 0.48292587155963307,
      "accuracy_std_dev": 0.12738947709449225,
      "run_count": 7,
      "total_questions": 1308,
      "avg_cost_per_task_usd": 0.0038326207951070333,
      "avg_duration_per_task_ms": 105139.82721712538,
      "avg_output_tokens_per_task": 3476.9533639143733,
      "last_run_timestamp": "2026-08-29T20:57:58.443Z"
    },
    {
      "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": 0.481481,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.001115070707070707,
      "avg_duration_per_task_ms": 120135.48484848485,
      "avg_output_tokens_per_task": 2056.4848484848485,
      "last_run_timestamp": "2026-09-01T14:05:19.193Z"
    },
    {
      "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": 0.48106050000000006,
      "accuracy_std_dev": 0.018107337511351573,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0010583617424242422,
      "avg_duration_per_task_ms": 63968.31502525252,
      "avg_output_tokens_per_task": 2267.661616161616,
      "last_run_timestamp": "2026-08-29T19:21:45.914Z"
    },
    {
      "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": 0.4789527698717949,
      "accuracy_std_dev": 0.04728636391114779,
      "run_count": 8,
      "total_questions": 1560,
      "avg_cost_per_task_usd": 0.010833255128205129,
      "avg_duration_per_task_ms": 439559.2346153846,
      "avg_output_tokens_per_task": 27024.191025641026,
      "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": "Together",
      "accuracy": 0.47661788888888884,
      "accuracy_std_dev": 0.012408007588689965,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0031713282828282827,
      "avg_duration_per_task_ms": 53718.556116722786,
      "avg_output_tokens_per_task": 2243.5089786756453,
      "last_run_timestamp": "2026-08-29T18:45:20.686Z"
    },
    {
      "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": 0.476434512295082,
      "accuracy_std_dev": 0.3072418769536471,
      "run_count": 3,
      "total_questions": 488,
      "avg_cost_per_task_usd": 0.05052364344262295,
      "avg_duration_per_task_ms": 618195.0081967213,
      "avg_output_tokens_per_task": 20701.57786885246,
      "last_run_timestamp": "2026-07-18T16:15:33.587Z"
    },
    {
      "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": 0.47643079999999993,
      "accuracy_std_dev": 0.007896357575490103,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0023637252525252526,
      "avg_duration_per_task_ms": 39731.30202020202,
      "avg_output_tokens_per_task": 2520.320202020202,
      "last_run_timestamp": "2026-08-29T18:16:38.845Z"
    },
    {
      "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": 0.4758696666666666,
      "accuracy_std_dev": 0.013768379457615502,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0023374680134680133,
      "avg_duration_per_task_ms": 37003.887205387204,
      "avg_output_tokens_per_task": 2485.9175084175085,
      "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": "auto-routing",
      "accuracy": 0.4757993750000001,
      "accuracy_std_dev": 0.012652503437437788,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0014489943181818182,
      "avg_duration_per_task_ms": 65459.88446969697,
      "avg_output_tokens_per_task": 2312.2613636363635,
      "last_run_timestamp": "2026-08-29T18:24:23.944Z"
    },
    {
      "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": 0.47530849999999997,
      "accuracy_std_dev": 0.00867533989247684,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.04888047474747474,
      "avg_duration_per_task_ms": 45790.32744107744,
      "avg_output_tokens_per_task": 17137.850168350167,
      "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": "Parasail",
      "accuracy": 0.4748365805432723,
      "accuracy_std_dev": 0.012294356500036914,
      "run_count": 8,
      "total_questions": 1583,
      "avg_cost_per_task_usd": 0.001331595072646873,
      "avg_duration_per_task_ms": 48334.044219835756,
      "avg_output_tokens_per_task": 2397.6392924826278,
      "last_run_timestamp": "2026-08-29T17:00:17.733Z"
    },
    {
      "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": 0.47447243417721524,
      "accuracy_std_dev": 0.01660465778733108,
      "run_count": 8,
      "total_questions": 1580,
      "avg_cost_per_task_usd": 0.0012851841772151898,
      "avg_duration_per_task_ms": 95695.09050632911,
      "avg_output_tokens_per_task": 2940.8867088607594,
      "last_run_timestamp": "2026-08-29T19:00:36.673Z"
    },
    {
      "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": 0.47327425,
      "accuracy_std_dev": 0.012469075374200769,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0010748301767676768,
      "avg_duration_per_task_ms": 178643.0707070707,
      "avg_output_tokens_per_task": 2430.4804292929293,
      "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": "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": "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": 0.46569850000000007,
      "accuracy_std_dev": 0.0211435505355652,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0009726685606060605,
      "avg_duration_per_task_ms": 187606.375,
      "avg_output_tokens_per_task": 2789.0145202020203,
      "last_run_timestamp": "2026-08-29T18:48:55.759Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.4461280000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0009776919191919191,
      "avg_duration_per_task_ms": 90556.41414141415,
      "avg_output_tokens_per_task": 1759.8131313131314,
      "last_run_timestamp": "2026-09-01T14:15:49.249Z"
    },
    {
      "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": 0.4335015,
      "accuracy_std_dev": 0.009259500000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0010084242424242425,
      "avg_duration_per_task_ms": 21177.93181818182,
      "avg_output_tokens_per_task": 1486.2676767676767,
      "last_run_timestamp": "2026-09-01T17:49:46.133Z"
    },
    {
      "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": 0.4301345,
      "accuracy_std_dev": 0.009259500000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0010125732323232322,
      "avg_duration_per_task_ms": 17946.272727272728,
      "avg_output_tokens_per_task": 1495.5707070707072,
      "last_run_timestamp": "2026-09-01T17:55:31.492Z"
    },
    {
      "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": 0.4297135,
      "accuracy_std_dev": 0.007469839539775939,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0010784545454545454,
      "avg_duration_per_task_ms": 21933.41792929293,
      "avg_output_tokens_per_task": 1515.1376262626263,
      "last_run_timestamp": "2026-09-01T17:31:07.952Z"
    },
    {
      "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": "Google",
      "accuracy": 0.39295421951219517,
      "accuracy_std_dev": 0.1609075,
      "run_count": 2,
      "total_questions": 369,
      "avg_cost_per_task_usd": 0.004463872628726287,
      "avg_duration_per_task_ms": 406770.0298102981,
      "avg_output_tokens_per_task": 25760.539295392955,
      "last_run_timestamp": "2026-07-11T15:34:14.755Z"
    },
    {
      "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": 0.3860826666666667,
      "accuracy_std_dev": 0.005722799567422292,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.00015977104377104374,
      "avg_duration_per_task_ms": 6254.590909090909,
      "avg_output_tokens_per_task": 953.9612794612794,
      "last_run_timestamp": "2026-09-01T13:54:02.609Z"
    },
    {
      "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": 0.38548754421768705,
      "accuracy_std_dev": 0.12183173346965978,
      "run_count": 3,
      "total_questions": 441,
      "avg_cost_per_task_usd": 0.009109496598639456,
      "avg_duration_per_task_ms": 405162.566893424,
      "avg_output_tokens_per_task": 14046.952380952382,
      "last_run_timestamp": "2026-08-01T18:13:08.069Z"
    },
    {
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.3526882276497696,
      "accuracy_std_dev": 0.11263472264421345,
      "run_count": 6,
      "total_questions": 1085,
      "avg_cost_per_task_usd": 0.013297605529953918,
      "avg_duration_per_task_ms": 2750242.6949308757,
      "avg_output_tokens_per_task": 14171.58801843318,
      "last_run_timestamp": "2026-08-22T17:37:26.215Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Nebius",
      "accuracy": 0.338384,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.06806519191919191,
      "avg_duration_per_task_ms": 227007.78787878787,
      "avg_output_tokens_per_task": 16827.81313131313,
      "last_run_timestamp": "2026-07-11T08:07:17.820Z"
    },
    {
      "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": "Mistral",
      "accuracy": 0.3251026825396825,
      "accuracy_std_dev": 0.01882667006019798,
      "run_count": 3,
      "total_questions": 567,
      "avg_cost_per_task_usd": 0.0001578130511463845,
      "avg_duration_per_task_ms": 9924.358024691359,
      "avg_output_tokens_per_task": 577.2874779541446,
      "last_run_timestamp": "2026-07-18T09:09:45.839Z"
    },
    {
      "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": 0.324373,
      "accuracy_std_dev": 0.011389673110127261,
      "run_count": 3,
      "total_questions": 558,
      "avg_cost_per_task_usd": 0.00011326881720430107,
      "avg_duration_per_task_ms": 11535.883512544802,
      "avg_output_tokens_per_task": 458.0806451612903,
      "last_run_timestamp": "2026-07-18T09:08:52.364Z"
    },
    {
      "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": 0.3174145461538462,
      "accuracy_std_dev": 0.06193726613628906,
      "run_count": 8,
      "total_questions": 1560,
      "avg_cost_per_task_usd": 0.0030096519230769234,
      "avg_duration_per_task_ms": 33381.006410256414,
      "avg_output_tokens_per_task": 1258.1852564102564,
      "last_run_timestamp": "2026-08-29T17:41:08.059Z"
    },
    {
      "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": 0.3152173260869565,
      "accuracy_std_dev": 0.010785089831182077,
      "run_count": 3,
      "total_questions": 552,
      "avg_cost_per_task_usd": 0.000030884057971014495,
      "avg_duration_per_task_ms": 13871.666666666666,
      "avg_output_tokens_per_task": 431.25,
      "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": "Novita",
      "accuracy": 0.3119177802607076,
      "accuracy_std_dev": 0.024779336369016997,
      "run_count": 3,
      "total_questions": 537,
      "avg_cost_per_task_usd": 0.000113,
      "avg_duration_per_task_ms": 43675.3947858473,
      "avg_output_tokens_per_task": 480.8175046554935,
      "last_run_timestamp": "2026-07-18T08:00:15.379Z"
    },
    {
      "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": 0.29040400000000005,
      "accuracy_std_dev": 0.004209000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.000060893939393939394,
      "avg_duration_per_task_ms": 9271.39393939394,
      "avg_output_tokens_per_task": 573.0479797979798,
      "last_run_timestamp": "2026-09-01T13:00:51.348Z"
    },
    {
      "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": 0.2866762857142857,
      "accuracy_std_dev": 0.018435693808589945,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0014998318903318908,
      "avg_duration_per_task_ms": 64419.20274170274,
      "avg_output_tokens_per_task": 19046.510822510823,
      "last_run_timestamp": "2026-08-29T18:21:00.168Z"
    },
    {
      "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": 0.2780711647307286,
      "accuracy_std_dev": 0.014628725522067875,
      "run_count": 5,
      "total_questions": 947,
      "avg_cost_per_task_usd": 0.0063300253431890175,
      "avg_duration_per_task_ms": 245873.9429778247,
      "avg_output_tokens_per_task": 28094.863780359028,
      "last_run_timestamp": "2026-08-29T18:18:26.937Z"
    },
    {
      "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": 0.27607364417177915,
      "accuracy_std_dev": 0.013048869410369953,
      "run_count": 3,
      "total_questions": 489,
      "avg_cost_per_task_usd": 0.000023339468302658487,
      "avg_duration_per_task_ms": 14320.529652351739,
      "avg_output_tokens_per_task": 284.7852760736196,
      "last_run_timestamp": "2026-07-18T08:04:03.449Z"
    },
    {
      "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": "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": 0.2722029671280277,
      "accuracy_std_dev": 0.017795645653423824,
      "run_count": 9,
      "total_questions": 1734,
      "avg_cost_per_task_usd": 0.0002858858131487889,
      "avg_duration_per_task_ms": 345712.0732410611,
      "avg_output_tokens_per_task": 6992.540369088812,
      "last_run_timestamp": "2026-08-29T16:57:41.646Z"
    },
    {
      "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": 0.27185566666666666,
      "accuracy_std_dev": 0.008747009330178078,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.00006098307952622674,
      "avg_duration_per_task_ms": 9307.97461928934,
      "avg_output_tokens_per_task": 612.7817258883249,
      "last_run_timestamp": "2026-09-01T12:41:31.862Z"
    },
    {
      "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": 0.2700818571428571,
      "accuracy_std_dev": 0.012593595701541222,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0017664740259740262,
      "avg_duration_per_task_ms": 42604.03391053391,
      "avg_output_tokens_per_task": 22219.9329004329,
      "last_run_timestamp": "2026-08-29T18:45:59.793Z"
    },
    {
      "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": 0.26936,
      "accuracy_std_dev": 0.00673399999999999,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0000632449494949495,
      "avg_duration_per_task_ms": 9455.388888888889,
      "avg_output_tokens_per_task": 592.1616161616162,
      "last_run_timestamp": "2026-09-01T13:21:46.100Z"
    },
    {
      "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": 0.268788525198939,
      "accuracy_std_dev": 0.004147499999999998,
      "run_count": 2,
      "total_questions": 377,
      "avg_cost_per_task_usd": 0.0000679681697612732,
      "avg_duration_per_task_ms": 9180.114058355437,
      "avg_output_tokens_per_task": 568.5623342175066,
      "last_run_timestamp": "2026-09-01T13:04:23.041Z"
    },
    {
      "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": 0.258137,
      "accuracy_std_dev": 0.0432133818895644,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.00887828367003367,
      "avg_duration_per_task_ms": 545067.4250841751,
      "avg_output_tokens_per_task": 21968.73063973064,
      "last_run_timestamp": "2026-08-29T19:48:22.407Z"
    },
    {
      "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": 0.25591029787234043,
      "accuracy_std_dev": 0.0038087773191231088,
      "run_count": 3,
      "total_questions": 564,
      "avg_cost_per_task_usd": 0.0013007553191489362,
      "avg_duration_per_task_ms": 38112.132978723406,
      "avg_output_tokens_per_task": 5103.014184397163,
      "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": "Novita",
      "accuracy": 0.24990534129692835,
      "accuracy_std_dev": 0.019806748083779954,
      "run_count": 9,
      "total_questions": 1758,
      "avg_cost_per_task_usd": 0.00038077246871444826,
      "avg_duration_per_task_ms": 63806.37883959044,
      "avg_output_tokens_per_task": 7291.984072810012,
      "last_run_timestamp": "2026-08-29T16:55:14.828Z"
    },
    {
      "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": 0.23833585714285715,
      "accuracy_std_dev": 0.09800656715959478,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.001642277056277056,
      "avg_duration_per_task_ms": 571156.3896103896,
      "avg_output_tokens_per_task": 5738.018037518037,
      "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": "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": 0.19696999999999998,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0190694696969697,
      "avg_duration_per_task_ms": 2493159.101010101,
      "avg_output_tokens_per_task": 105666.86363636363,
      "last_run_timestamp": "2026-09-03T21:14:17.509Z"
    },
    {
      "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": 0.16687712500000002,
      "accuracy_std_dev": 0.10885759932641073,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0013347329545454543,
      "avg_duration_per_task_ms": 43166.87626262626,
      "avg_output_tokens_per_task": 1152.546085858586,
      "last_run_timestamp": "2026-08-29T18:43:09.114Z"
    },
    {
      "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": 0.114478,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.017837661616161616,
      "avg_duration_per_task_ms": 71618.32323232324,
      "avg_output_tokens_per_task": 11767.727272727272,
      "last_run_timestamp": "2026-09-01T12:56:59.733Z"
    },
    {
      "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": 0.10606100000000002,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.017886560606060607,
      "avg_duration_per_task_ms": 21072.777777777777,
      "avg_output_tokens_per_task": 11801.328282828283,
      "last_run_timestamp": "2026-09-01T13:00:01.945Z"
    },
    {
      "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": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.018059262626262628,
      "avg_duration_per_task_ms": 21777.424242424244,
      "avg_output_tokens_per_task": 11808.9898989899,
      "last_run_timestamp": "2026-09-01T13:21:37.702Z"
    },
    {
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Together",
      "accuracy": 0.005051,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00002468181818181818,
      "avg_duration_per_task_ms": 955.2929292929293,
      "avg_output_tokens_per_task": 122.03030303030303,
      "last_run_timestamp": "2026-07-04T16:00:23.384Z"
    },
    {
      "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": 0.002806333333333333,
      "accuracy_std_dev": 0.0015872190215034036,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0035528451178451176,
      "avg_duration_per_task_ms": 168324.14646464647,
      "avg_output_tokens_per_task": 3352.3703703703704,
      "last_run_timestamp": "2026-08-15T14:54:58.071Z"
    },
    {
      "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": 0.002187456692913386,
      "accuracy_std_dev": 0.0006963147187156109,
      "run_count": 4,
      "total_questions": 762,
      "avg_cost_per_task_usd": 0.0035705761154855645,
      "avg_duration_per_task_ms": 176293.88188976378,
      "avg_output_tokens_per_task": 3370.4238845144355,
      "last_run_timestamp": "2026-08-22T17:29:06.480Z"
    },
    {
      "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": "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": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google",
      "accuracy": 0.805556,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.07709047916666667,
      "avg_duration_per_task_ms": 121300.85416666667,
      "avg_output_tokens_per_task": 14694.6875,
      "last_run_timestamp": "2026-08-15T00:15:15.845Z"
    },
    {
      "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": 0.804166625,
      "accuracy_std_dev": 0.021065488167483215,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.8715094075,
      "avg_duration_per_task_ms": 137048.4925,
      "avg_output_tokens_per_task": 5786.43,
      "last_run_timestamp": "2026-09-26T08:21:11.883Z"
    },
    {
      "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": "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": "auto-routing",
      "accuracy": 0.795833375,
      "accuracy_std_dev": 0.02846769733811245,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.9313334725,
      "avg_duration_per_task_ms": 130292.9175,
      "avg_output_tokens_per_task": 5748.73,
      "last_run_timestamp": "2026-09-26T08:18:08.440Z"
    },
    {
      "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": 0.7933333333333333,
      "accuracy_std_dev": 0.014401440259764123,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.6872690333333334,
      "avg_duration_per_task_ms": 114143.74666666667,
      "avg_output_tokens_per_task": 6925.7266666666665,
      "last_run_timestamp": "2026-09-26T08:22:07.944Z"
    },
    {
      "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": 0.793333,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.10986884,
      "avg_duration_per_task_ms": 261740.51,
      "avg_output_tokens_per_task": 11828.99,
      "last_run_timestamp": "2026-08-29T16:51:02.854Z"
    },
    {
      "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": 0.7911111666666667,
      "accuracy_std_dev": 0.019116304205369344,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.8711063166666666,
      "avg_duration_per_task_ms": 151211.67333333334,
      "avg_output_tokens_per_task": 5696.93,
      "last_run_timestamp": "2026-09-26T08:09:55.578Z"
    },
    {
      "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": 0.790833375,
      "accuracy_std_dev": 0.01630862714897777,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.8205446474999999,
      "avg_duration_per_task_ms": 138181.955,
      "avg_output_tokens_per_task": 5731.4725,
      "last_run_timestamp": "2026-09-26T08:05:12.739Z"
    },
    {
      "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": 0.788889,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08521802666666665,
      "avg_duration_per_task_ms": 39816.48,
      "avg_output_tokens_per_task": 4684.933333333333,
      "last_run_timestamp": "2026-09-26T08:13:10.629Z"
    },
    {
      "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": 0.78833325,
      "accuracy_std_dev": 0.009860203404975974,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.8196274600000001,
      "avg_duration_per_task_ms": 183041.66,
      "avg_output_tokens_per_task": 7098.595,
      "last_run_timestamp": "2026-09-26T08:15:33.263Z"
    },
    {
      "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": 0.7883064274193549,
      "accuracy_std_dev": 0.013624850889459287,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.07264496370967742,
      "avg_duration_per_task_ms": 503105.06048387097,
      "avg_output_tokens_per_task": 9382.189516129032,
      "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": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Google",
      "accuracy": 0.7877778666666666,
      "accuracy_std_dev": 0.02049995242552745,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.9492467306666665,
      "avg_duration_per_task_ms": 98954.73466666667,
      "avg_output_tokens_per_task": 5602.965333333334,
      "last_run_timestamp": "2026-09-26T08:12:28.628Z"
    },
    {
      "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": 0.7875,
      "accuracy_std_dev": 0.024593980747329203,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.48769549000000006,
      "avg_duration_per_task_ms": 129773.455,
      "avg_output_tokens_per_task": 7130.5275,
      "last_run_timestamp": "2026-09-26T08:22:26.398Z"
    },
    {
      "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": 0.7870992438478748,
      "accuracy_std_dev": 0.008959550928450943,
      "run_count": 9,
      "total_questions": 447,
      "avg_cost_per_task_usd": 0.11607393959731543,
      "avg_duration_per_task_ms": 143483.8120805369,
      "avg_output_tokens_per_task": 11288.324384787473,
      "last_run_timestamp": "2026-09-26T08:27:30.005Z"
    },
    {
      "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": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.14780512,
      "avg_duration_per_task_ms": 461958.93,
      "avg_output_tokens_per_task": 11472.92,
      "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": "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": 0.786667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 1.8471384599999998,
      "avg_duration_per_task_ms": 331793.46,
      "avg_output_tokens_per_task": 22741.94,
      "last_run_timestamp": "2026-08-01T08:06:01.602Z"
    },
    {
      "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": "Morph",
      "accuracy": 0.786667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.17120372,
      "avg_duration_per_task_ms": 466039.6,
      "avg_output_tokens_per_task": 22362.94,
      "last_run_timestamp": "2026-07-04T16:00:23.544Z"
    },
    {
      "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": 0.7866669999999999,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 1.2795196199999999,
      "avg_duration_per_task_ms": 104647.27,
      "avg_output_tokens_per_task": 5686.75,
      "last_run_timestamp": "2026-09-01T13:03:28.251Z"
    },
    {
      "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": 0.7866665,
      "accuracy_std_dev": 0.026666499999999982,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12536731,
      "avg_duration_per_task_ms": 118841.15,
      "avg_output_tokens_per_task": 19578.58,
      "last_run_timestamp": "2026-09-01T17:30:58.594Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "BaseTen",
      "accuracy": 0.7857143728222997,
      "accuracy_std_dev": 0.022636034639976627,
      "run_count": 6,
      "total_questions": 287,
      "avg_cost_per_task_usd": 0.11757857142857142,
      "avg_duration_per_task_ms": 73705.68641114983,
      "avg_output_tokens_per_task": 8241.627177700348,
      "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-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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "StreamLake",
      "accuracy": 0.7837037777777778,
      "accuracy_std_dev": 0.023117842088731067,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.04375683111111111,
      "avg_duration_per_task_ms": 211952.00222222222,
      "avg_output_tokens_per_task": 9463.306666666667,
      "last_run_timestamp": "2026-08-29T17:11:32.467Z"
    },
    {
      "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": 0.7833334999999999,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02978465,
      "avg_duration_per_task_ms": 411185.87,
      "avg_output_tokens_per_task": 14644.16,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "ModelRun",
      "accuracy": 0.7833334999999999,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06834454,
      "avg_duration_per_task_ms": 91367.71,
      "avg_output_tokens_per_task": 11815.81,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.78166675,
      "accuracy_std_dev": 0.018484295136344808,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1845917575,
      "avg_duration_per_task_ms": 205801.49,
      "avg_output_tokens_per_task": 9205.5075,
      "last_run_timestamp": "2026-08-29T16:12:45.134Z"
    },
    {
      "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": 0.7813334,
      "accuracy_std_dev": 0.022861647653657863,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.016082664,
      "avg_duration_per_task_ms": 178213.516,
      "avg_output_tokens_per_task": 11864.648,
      "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-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Azure",
      "accuracy": 0.78041675,
      "accuracy_std_dev": 0.02409111033789228,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.4685391825,
      "avg_duration_per_task_ms": 111064.53375,
      "avg_output_tokens_per_task": 7216.75875,
      "last_run_timestamp": "2026-09-26T08:13:10.617Z"
    },
    {
      "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": 0.78000025,
      "accuracy_std_dev": 0.008164965810553043,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.7813667750000001,
      "avg_duration_per_task_ms": 113768.35,
      "avg_output_tokens_per_task": 6852.055,
      "last_run_timestamp": "2026-09-26T08:13:59.150Z"
    },
    {
      "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": 0.78,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.19113603999999998,
      "avg_duration_per_task_ms": 79138.93,
      "avg_output_tokens_per_task": 6244.06,
      "last_run_timestamp": "2026-09-29T12:00:07.887Z"
    },
    {
      "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": 0.78,
      "accuracy_std_dev": 0.033993485215716136,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.04603017555555555,
      "avg_duration_per_task_ms": 205800.68,
      "avg_output_tokens_per_task": 9387.753333333334,
      "last_run_timestamp": "2026-08-29T16:11:46.640Z"
    },
    {
      "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-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7788889166666667,
      "accuracy_std_dev": 0.02298676618897032,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.4834922383333334,
      "avg_duration_per_task_ms": 127667.31666666667,
      "avg_output_tokens_per_task": 7153.7075,
      "last_run_timestamp": "2026-09-26T08:12:13.435Z"
    },
    {
      "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": 0.7783335,
      "accuracy_std_dev": 0.028037782049406136,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.099593165,
      "avg_duration_per_task_ms": 81195.23,
      "avg_output_tokens_per_task": 4770.2,
      "last_run_timestamp": "2026-09-26T08:13:27.135Z"
    },
    {
      "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": 0.77833325,
      "accuracy_std_dev": 0.026404708125398767,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.807941035,
      "avg_duration_per_task_ms": 183353.72,
      "avg_output_tokens_per_task": 6938.485,
      "last_run_timestamp": "2026-09-26T08:11:40.896Z"
    },
    {
      "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": 0.77813332,
      "accuracy_std_dev": 0.02505255531991896,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.5623465103999999,
      "avg_duration_per_task_ms": 118279.7304,
      "avg_output_tokens_per_task": 7235.6968,
      "last_run_timestamp": "2026-09-26T08:16:10.311Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.7771428571428571,
      "accuracy_std_dev": 0.03204321345321288,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.027308874285714286,
      "avg_duration_per_task_ms": 297009.5657142857,
      "avg_output_tokens_per_task": 12274.965714285714,
      "last_run_timestamp": "2026-08-29T16:33:11.754Z"
    },
    {
      "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": 0.7770961820040899,
      "accuracy_std_dev": 0.01908827114120082,
      "run_count": 10,
      "total_questions": 489,
      "avg_cost_per_task_usd": 0.07789318609406953,
      "avg_duration_per_task_ms": 97210.93251533742,
      "avg_output_tokens_per_task": 4258.01226993865,
      "last_run_timestamp": "2026-09-26T08:11:43.280Z"
    },
    {
      "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": 0.7770371111111111,
      "accuracy_std_dev": 0.02224700963397825,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.4357714733333334,
      "avg_duration_per_task_ms": 133682.32888888888,
      "avg_output_tokens_per_task": 7264.266666666666,
      "last_run_timestamp": "2026-09-26T08:17:01.715Z"
    },
    {
      "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": "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": 0.776,
      "accuracy_std_dev": 0.022939553421982742,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.007782092000000001,
      "avg_duration_per_task_ms": 211248.992,
      "avg_output_tokens_per_task": 4235.592,
      "last_run_timestamp": "2026-09-26T08:27:53.093Z"
    },
    {
      "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": 0.7755555000000001,
      "accuracy_std_dev": 0.0236225357991756,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.4783927666666667,
      "avg_duration_per_task_ms": 128906.20333333334,
      "avg_output_tokens_per_task": 7843.846666666666,
      "last_run_timestamp": "2026-09-26T08:22:36.885Z"
    },
    {
      "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": 0.7750001249999999,
      "accuracy_std_dev": 0.025331099134450842,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0651128975,
      "avg_duration_per_task_ms": 279584.37,
      "avg_output_tokens_per_task": 14735.76,
      "last_run_timestamp": "2026-08-29T17:42:47.827Z"
    },
    {
      "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": 0.775,
      "accuracy_std_dev": 0.029011578149938697,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.09548250750000001,
      "avg_duration_per_task_ms": 233151.12,
      "avg_output_tokens_per_task": 11897.7625,
      "last_run_timestamp": "2026-08-29T16:59:11.598Z"
    },
    {
      "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": 0.7746666,
      "accuracy_std_dev": 0.02902873978387625,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.15693796000000002,
      "avg_duration_per_task_ms": 605970.636,
      "avg_output_tokens_per_task": 12427.38,
      "last_run_timestamp": "2026-09-26T08:23:03.870Z"
    },
    {
      "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": 0.7746666,
      "accuracy_std_dev": 0.022861927596770998,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.080879576,
      "avg_duration_per_task_ms": 126701.14,
      "avg_output_tokens_per_task": 4510.112,
      "last_run_timestamp": "2026-09-26T08:11:43.279Z"
    },
    {
      "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": 0.7746666,
      "accuracy_std_dev": 0.020395973461445754,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.120308296,
      "avg_duration_per_task_ms": 80105.144,
      "avg_output_tokens_per_task": 4494.34,
      "last_run_timestamp": "2026-09-26T08:09:25.342Z"
    },
    {
      "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": 0.7746664000000001,
      "accuracy_std_dev": 0.016546172434735443,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.09735419199999999,
      "avg_duration_per_task_ms": 130067.476,
      "avg_output_tokens_per_task": 4561.928,
      "last_run_timestamp": "2026-09-26T08:10:40.765Z"
    },
    {
      "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-sonnet-5.5-20260928",
      "model_slug": "anthropic/claude-sonnet-5.5",
      "model_name": "Anthropic: Claude Sonnet 5.5",
      "provider_name": "Google",
      "accuracy": 0.7733333333333333,
      "accuracy_std_dev": 0.014401440259764123,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.18828580666666667,
      "avg_duration_per_task_ms": 67452,
      "avg_output_tokens_per_task": 6335.053333333333,
      "last_run_timestamp": "2026-09-29T12:00:08.126Z"
    },
    {
      "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": 0.7733333333333333,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02996874,
      "avg_duration_per_task_ms": 113373.20666666667,
      "avg_output_tokens_per_task": 12010.053333333333,
      "last_run_timestamp": "2026-08-29T16:46:02.125Z"
    },
    {
      "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": 0.7733332727272727,
      "accuracy_std_dev": 0.018422746153456893,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.4337744636363636,
      "avg_duration_per_task_ms": 96869.80363636363,
      "avg_output_tokens_per_task": 5661.309090909091,
      "last_run_timestamp": "2026-09-26T08:11:40.823Z"
    },
    {
      "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": 0.7733332,
      "accuracy_std_dev": 0.004216264811417815,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.007032796,
      "avg_duration_per_task_ms": 161320.244,
      "avg_output_tokens_per_task": 16830.132,
      "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": "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": 0.773333,
      "accuracy_std_dev": 0.019999999999999962,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12487065999999998,
      "avg_duration_per_task_ms": 149585.85,
      "avg_output_tokens_per_task": 20477.48,
      "last_run_timestamp": "2026-09-01T17:31:27.257Z"
    },
    {
      "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": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "Novita",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06535734,
      "avg_duration_per_task_ms": 368327.08,
      "avg_output_tokens_per_task": 13147.48,
      "last_run_timestamp": "2026-09-23T15:00:14.137Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Alibaba",
      "accuracy": 0.7726666,
      "accuracy_std_dev": 0.03641728405359191,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.113975304,
      "avg_duration_per_task_ms": 177276.2,
      "avg_output_tokens_per_task": 9278.056,
      "last_run_timestamp": "2026-08-29T16:15:12.658Z"
    },
    {
      "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": 0.7725924444444445,
      "accuracy_std_dev": 0.024433234243142,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.07833540444444444,
      "avg_duration_per_task_ms": 224489.25555555554,
      "avg_output_tokens_per_task": 9337.3,
      "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": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Anthropic",
      "accuracy": 0.7725,
      "accuracy_std_dev": 0.030810865688909155,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.47031593,
      "avg_duration_per_task_ms": 117109.07,
      "avg_output_tokens_per_task": 7210.535,
      "last_run_timestamp": "2026-09-26T08:12:30.655Z"
    },
    {
      "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": 0.7725,
      "accuracy_std_dev": 0.03302967521033168,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.22805288999999998,
      "avg_duration_per_task_ms": 411518.45,
      "avg_output_tokens_per_task": 13622.02,
      "last_run_timestamp": "2026-09-26T08:05:12.789Z"
    },
    {
      "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": 0.7722223333333333,
      "accuracy_std_dev": 0.03890480468042762,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07840497666666667,
      "avg_duration_per_task_ms": 265427.81666666665,
      "avg_output_tokens_per_task": 11547.233333333334,
      "last_run_timestamp": "2026-09-26T08:13:10.636Z"
    },
    {
      "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": 0.77166675,
      "accuracy_std_dev": 0.023033966466666143,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.5286243174999999,
      "avg_duration_per_task_ms": 174531.3525,
      "avg_output_tokens_per_task": 10507.29,
      "last_run_timestamp": "2026-09-26T08:05:12.739Z"
    },
    {
      "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": 0.7715355505617978,
      "accuracy_std_dev": 0.03676049559271777,
      "run_count": 9,
      "total_questions": 445,
      "avg_cost_per_task_usd": 0.08241994157303371,
      "avg_duration_per_task_ms": 1113393.984269663,
      "avg_output_tokens_per_task": 27003.040449438202,
      "last_run_timestamp": "2026-09-26T08:24:45.038Z"
    },
    {
      "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": 0.7715151818181818,
      "accuracy_std_dev": 0.029417231634840356,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.5523802672727273,
      "avg_duration_per_task_ms": 183979.0109090909,
      "avg_output_tokens_per_task": 10526.685454545455,
      "last_run_timestamp": "2026-09-26T08:05:12.761Z"
    },
    {
      "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": 0.7710438282828282,
      "accuracy_std_dev": 0.01667057037866429,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.1615302070707071,
      "avg_duration_per_task_ms": 338222.0707070707,
      "avg_output_tokens_per_task": 20443.252525252527,
      "last_run_timestamp": "2026-09-26T08:27:30.395Z"
    },
    {
      "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": 0.7708335,
      "accuracy_std_dev": 0.022345177208963923,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.17567465999999998,
      "avg_duration_per_task_ms": 650826.8375,
      "avg_output_tokens_per_task": 27562.8475,
      "last_run_timestamp": "2026-08-29T17:01:22.022Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.7706668,
      "accuracy_std_dev": 0.023702035544653124,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.074997112,
      "avg_duration_per_task_ms": 263018.86,
      "avg_output_tokens_per_task": 12780.056,
      "last_run_timestamp": "2026-09-26T08:10:10.216Z"
    },
    {
      "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": 0.7706668,
      "accuracy_std_dev": 0.01818435958069462,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.071001692,
      "avg_duration_per_task_ms": 34525.74,
      "avg_output_tokens_per_task": 4762.176,
      "last_run_timestamp": "2026-09-26T08:16:16.877Z"
    },
    {
      "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": 0.770303,
      "accuracy_std_dev": 0.018282610525752703,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.4931648818181818,
      "avg_duration_per_task_ms": 129914.10363636364,
      "avg_output_tokens_per_task": 8068.774545454546,
      "last_run_timestamp": "2026-09-26T08:08:19.694Z"
    },
    {
      "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": 0.7702221333333333,
      "accuracy_std_dev": 0.02970127552786618,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.5012201893333333,
      "avg_duration_per_task_ms": 129604.79333333333,
      "avg_output_tokens_per_task": 7921.828,
      "last_run_timestamp": "2026-09-26T08:23:02.726Z"
    },
    {
      "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": 0.77,
      "accuracy_std_dev": 0.03341663757396906,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.4393913750000001,
      "avg_duration_per_task_ms": 81234.3425,
      "avg_output_tokens_per_task": 5692.71875,
      "last_run_timestamp": "2026-09-26T08:16:17.700Z"
    },
    {
      "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": 0.77,
      "accuracy_std_dev": 0.023332999999999993,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.35428131,
      "avg_duration_per_task_ms": 261243.27,
      "avg_output_tokens_per_task": 8792.41,
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Parasail",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.007453485392754197,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.081881785,
      "avg_duration_per_task_ms": 99238.51,
      "avg_output_tokens_per_task": 4510.805,
      "last_run_timestamp": "2026-09-26T08:25:16.354Z"
    },
    {
      "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": 0.77,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.073975635,
      "avg_duration_per_task_ms": 45899.64,
      "avg_output_tokens_per_task": 4662.84,
      "last_run_timestamp": "2026-09-26T08:21:33.013Z"
    },
    {
      "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": "Alibaba",
      "accuracy": 0.7688888888888888,
      "accuracy_std_dev": 0.01662949930056987,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.1780795511111111,
      "avg_duration_per_task_ms": 402300.5511111111,
      "avg_output_tokens_per_task": 27328.78888888889,
      "last_run_timestamp": "2026-08-29T16:25:47.585Z"
    },
    {
      "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": 0.7688888333333334,
      "accuracy_std_dev": 0.014229173651535601,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.10475028333333332,
      "avg_duration_per_task_ms": 333223.94666666666,
      "avg_output_tokens_per_task": 20225.06333333333,
      "last_run_timestamp": "2026-09-26T08:21:56.720Z"
    },
    {
      "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": "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": 0.7687072040816327,
      "accuracy_std_dev": 0.0156945,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 1.7723489081632653,
      "avg_duration_per_task_ms": 313060.2448979592,
      "avg_output_tokens_per_task": 23077.85714285714,
      "last_run_timestamp": "2026-08-01T08:06:07.759Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "BaseTen",
      "accuracy": 0.768421252631579,
      "accuracy_std_dev": 0.0038080000000000336,
      "run_count": 2,
      "total_questions": 95,
      "avg_cost_per_task_usd": 0.06421121052631579,
      "avg_duration_per_task_ms": 69460.44210526315,
      "avg_output_tokens_per_task": 5883.873684210526,
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Wafer",
      "accuracy": 0.7683335,
      "accuracy_std_dev": 0.010929191289843909,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.09395701500000002,
      "avg_duration_per_task_ms": 284778.19,
      "avg_output_tokens_per_task": 21938.115,
      "last_run_timestamp": "2026-09-26T08:23:03.305Z"
    },
    {
      "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": 0.76833325,
      "accuracy_std_dev": 0.02048026828284971,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.4290680700000001,
      "avg_duration_per_task_ms": 101484.8275,
      "avg_output_tokens_per_task": 5562.66,
      "last_run_timestamp": "2026-09-26T08:11:43.281Z"
    },
    {
      "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": 0.76833325,
      "accuracy_std_dev": 0.02074978581064154,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.062534375,
      "avg_duration_per_task_ms": 51740.385,
      "avg_output_tokens_per_task": 4651.035,
      "last_run_timestamp": "2026-09-26T08:14:03.992Z"
    },
    {
      "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": 0.7675,
      "accuracy_std_dev": 0.038505122315089464,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.11618410500000001,
      "avg_duration_per_task_ms": 295967.115,
      "avg_output_tokens_per_task": 12435.425,
      "last_run_timestamp": "2026-08-29T17:05:17.192Z"
    },
    {
      "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": 0.7671155218855219,
      "accuracy_std_dev": 0.020370521752577246,
      "run_count": 6,
      "total_questions": 297,
      "avg_cost_per_task_usd": 0.04689751515151515,
      "avg_duration_per_task_ms": 143803.91919191918,
      "avg_output_tokens_per_task": 4429.346801346801,
      "last_run_timestamp": "2026-09-26T08:24:40.664Z"
    },
    {
      "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": 0.7671130089285715,
      "accuracy_std_dev": 0.016761300652263098,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.05441544419642857,
      "avg_duration_per_task_ms": 217468.82142857142,
      "avg_output_tokens_per_task": 9416.404017857143,
      "last_run_timestamp": "2026-08-29T16:11:58.048Z"
    },
    {
      "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": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.18363575999999998,
      "avg_duration_per_task_ms": 79763.32,
      "avg_output_tokens_per_task": 6162.44,
      "last_run_timestamp": "2026-09-29T12:00:08.100Z"
    },
    {
      "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": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.18092455999999998,
      "avg_duration_per_task_ms": 72930.62,
      "avg_output_tokens_per_task": 6117.32,
      "last_run_timestamp": "2026-09-29T12:00:06.622Z"
    },
    {
      "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": "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": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 1.86423468,
      "avg_duration_per_task_ms": 335268.18,
      "avg_output_tokens_per_task": 22931.04,
      "last_run_timestamp": "2026-08-01T08:05:35.224Z"
    },
    {
      "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": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0119651,
      "avg_duration_per_task_ms": 540203.32,
      "avg_output_tokens_per_task": 23288.26,
      "last_run_timestamp": "2026-09-29T03:00:05.982Z"
    },
    {
      "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-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Alibaba",
      "accuracy": 0.7666666666666667,
      "accuracy_std_dev": 0.020366930350175255,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.1362619733333333,
      "avg_duration_per_task_ms": 318797.34,
      "avg_output_tokens_per_task": 19441.493333333332,
      "last_run_timestamp": "2026-09-26T08:05:12.735Z"
    },
    {
      "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": 0.7666666666666667,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12935825333333334,
      "avg_duration_per_task_ms": 58122.44,
      "avg_output_tokens_per_task": 4752.84,
      "last_run_timestamp": "2026-09-19T08:15:57.323Z"
    },
    {
      "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": 0.766666625,
      "accuracy_std_dev": 0.025819792151262073,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.03997693,
      "avg_duration_per_task_ms": 243468.475,
      "avg_output_tokens_per_task": 12205.0675,
      "last_run_timestamp": "2026-08-29T16:25:15.306Z"
    },
    {
      "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": 0.7666665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.21096532,
      "avg_duration_per_task_ms": 124417.44,
      "avg_output_tokens_per_task": 5140.16,
      "last_run_timestamp": "2026-09-01T12:59:37.360Z"
    },
    {
      "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": 0.7663316432160805,
      "accuracy_std_dev": 0.01611815722591683,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.08850016080402008,
      "avg_duration_per_task_ms": 191747.41206030152,
      "avg_output_tokens_per_task": 11737.070351758794,
      "last_run_timestamp": "2026-08-29T16:41:08.022Z"
    },
    {
      "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": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7659570000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 47,
      "avg_cost_per_task_usd": 0.006887468085106384,
      "avg_duration_per_task_ms": 773525.340425532,
      "avg_output_tokens_per_task": 9139.936170212766,
      "last_run_timestamp": "2026-09-22T21:00:04.649Z"
    },
    {
      "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": 0.76583325,
      "accuracy_std_dev": 0.03239305266855072,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1614212075,
      "avg_duration_per_task_ms": 219712.16,
      "avg_output_tokens_per_task": 9777.065,
      "last_run_timestamp": "2026-08-29T17:16:15.105Z"
    },
    {
      "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": "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": 0.765217347826087,
      "accuracy_std_dev": 0.020893702830421756,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.5270942373913046,
      "avg_duration_per_task_ms": 123051.67565217392,
      "avg_output_tokens_per_task": 8091.90347826087,
      "last_run_timestamp": "2026-09-26T08:26:39.210Z"
    },
    {
      "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": 0.7651820931174089,
      "accuracy_std_dev": 0.03553283679302851,
      "run_count": 5,
      "total_questions": 247,
      "avg_cost_per_task_usd": 0.09236648582995953,
      "avg_duration_per_task_ms": 316304.95546558703,
      "avg_output_tokens_per_task": 9844.623481781377,
      "last_run_timestamp": "2026-08-29T16:35:10.931Z"
    },
    {
      "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": 0.7651515555555556,
      "accuracy_std_dev": 0.04160471364201867,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02325603535353535,
      "avg_duration_per_task_ms": 506742.0202020202,
      "avg_output_tokens_per_task": 13406.39393939394,
      "last_run_timestamp": "2026-08-29T16:33:19.761Z"
    },
    {
      "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": 0.7648808660714286,
      "accuracy_std_dev": 0.04130098452896405,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.1113698392857143,
      "avg_duration_per_task_ms": 925496.1897321428,
      "avg_output_tokens_per_task": 10117.198660714286,
      "last_run_timestamp": "2026-08-29T16:54:14.369Z"
    },
    {
      "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": 0.7644446666666667,
      "accuracy_std_dev": 0.01911642045177101,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.13677892666666666,
      "avg_duration_per_task_ms": 241869.39333333334,
      "avg_output_tokens_per_task": 11667.466666666667,
      "last_run_timestamp": "2026-08-08T08:29:31.005Z"
    },
    {
      "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": 0.764166625,
      "accuracy_std_dev": 0.030081035231676705,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.56641688625,
      "avg_duration_per_task_ms": 181084.34625,
      "avg_output_tokens_per_task": 10579.67875,
      "last_run_timestamp": "2026-09-26T08:14:04.078Z"
    },
    {
      "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": 0.7639931779448622,
      "accuracy_std_dev": 0.027369852484074517,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.0707108245614035,
      "avg_duration_per_task_ms": 77146.8395989975,
      "avg_output_tokens_per_task": 13251.315789473685,
      "last_run_timestamp": "2026-08-29T17:41:16.842Z"
    },
    {
      "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": 0.7633335,
      "accuracy_std_dev": 0.015275361427802623,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.028111930000000004,
      "avg_duration_per_task_ms": 550867.43,
      "avg_output_tokens_per_task": 16398.445,
      "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": "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": 0.7633335,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.34846716,
      "avg_duration_per_task_ms": 136967.1,
      "avg_output_tokens_per_task": 12613.18,
      "last_run_timestamp": "2026-09-01T17:30:16.052Z"
    },
    {
      "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": 0.763333375,
      "accuracy_std_dev": 0.031091303715578986,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0139452675,
      "avg_duration_per_task_ms": 180694.13,
      "avg_output_tokens_per_task": 14861.8225,
      "last_run_timestamp": "2026-08-29T17:17:18.830Z"
    },
    {
      "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": 0.763333375,
      "accuracy_std_dev": 0.022607772755943365,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.040371300000000006,
      "avg_duration_per_task_ms": 175751.1025,
      "avg_output_tokens_per_task": 11405.795,
      "last_run_timestamp": "2026-08-29T17:07:19.233Z"
    },
    {
      "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": 0.763333375,
      "accuracy_std_dev": 0.02708005108441221,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1274939625,
      "avg_duration_per_task_ms": 269624.8775,
      "avg_output_tokens_per_task": 11653.1325,
      "last_run_timestamp": "2026-08-29T16:48:46.161Z"
    },
    {
      "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": "SiliconFlow",
      "accuracy": 0.7633332500000001,
      "accuracy_std_dev": 0.017743016363840187,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.018957181666666663,
      "avg_duration_per_task_ms": 239454.035,
      "avg_output_tokens_per_task": 17272.2,
      "last_run_timestamp": "2026-09-26T08:11:43.368Z"
    },
    {
      "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": 0.7633332500000001,
      "accuracy_std_dev": 0.005773358354328977,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.039311875,
      "avg_duration_per_task_ms": 88152.73,
      "avg_output_tokens_per_task": 4482.07,
      "last_run_timestamp": "2026-09-26T08:07:18.978Z"
    },
    {
      "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": 0.7633331666666667,
      "accuracy_std_dev": 0.02631724543853748,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.5436579066666667,
      "avg_duration_per_task_ms": 181008.31666666668,
      "avg_output_tokens_per_task": 7997.5633333333335,
      "last_run_timestamp": "2026-09-01T13:02:58.914Z"
    },
    {
      "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": 0.763333,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.10827190999999999,
      "avg_duration_per_task_ms": 51051.39,
      "avg_output_tokens_per_task": 12753.93,
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "StreamLake",
      "accuracy": 0.7626666,
      "accuracy_std_dev": 0.016110821761784848,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.09137509199999999,
      "avg_duration_per_task_ms": 332031.4,
      "avg_output_tokens_per_task": 19933.684,
      "last_run_timestamp": "2026-09-26T08:21:56.175Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.7626262929292928,
      "accuracy_std_dev": 0.02553140404849484,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.10055969696969695,
      "avg_duration_per_task_ms": 35430.38888888889,
      "avg_output_tokens_per_task": 3961.580808080808,
      "last_run_timestamp": "2026-09-26T08:30:51.947Z"
    },
    {
      "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": 0.7622223333333333,
      "accuracy_std_dev": 0.025141417307348132,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08369306,
      "avg_duration_per_task_ms": 261571.12,
      "avg_output_tokens_per_task": 19541.486666666668,
      "last_run_timestamp": "2026-09-26T08:16:17.045Z"
    },
    {
      "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": 0.7622223333333333,
      "accuracy_std_dev": 0.013698571612479268,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06797058666666667,
      "avg_duration_per_task_ms": 364277.74,
      "avg_output_tokens_per_task": 20919.446666666667,
      "last_run_timestamp": "2026-09-26T08:05:25.554Z"
    },
    {
      "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": 0.7622223333333333,
      "accuracy_std_dev": 0.006285707880227596,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.1522076066666667,
      "avg_duration_per_task_ms": 298029.9066666667,
      "avg_output_tokens_per_task": 19805.7,
      "last_run_timestamp": "2026-09-26T08:05:12.730Z"
    },
    {
      "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": 0.7622223333333333,
      "accuracy_std_dev": 0.013698571612479268,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09941037999999999,
      "avg_duration_per_task_ms": 137232.76666666666,
      "avg_output_tokens_per_task": 5643.006666666667,
      "last_run_timestamp": "2026-09-26T08:22:37.050Z"
    },
    {
      "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": 0.76222225,
      "accuracy_std_dev": 0.015234723173970049,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.016681795000000003,
      "avg_duration_per_task_ms": 202557.275,
      "avg_output_tokens_per_task": 19520.191666666666,
      "last_run_timestamp": "2026-09-26T08:22:26.866Z"
    },
    {
      "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": 0.7622221666666666,
      "accuracy_std_dev": 0.021659565788327558,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.026935736666666668,
      "avg_duration_per_task_ms": 147567.38333333333,
      "avg_output_tokens_per_task": 9119.013333333334,
      "last_run_timestamp": "2026-09-26T16:00:25.687Z"
    },
    {
      "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": 0.7620424450867053,
      "accuracy_std_dev": 0.05069435998456729,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.0817946936416185,
      "avg_duration_per_task_ms": 410762.72543352604,
      "avg_output_tokens_per_task": 12839.21387283237,
      "last_run_timestamp": "2026-08-29T17:06:46.422Z"
    },
    {
      "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": 0.7619999000000001,
      "accuracy_std_dev": 0.014313938818159046,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.008990316,
      "avg_duration_per_task_ms": 356282.446,
      "avg_output_tokens_per_task": 14140.9,
      "last_run_timestamp": "2026-09-26T08:22:54.418Z"
    },
    {
      "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": 0.7619998,
      "accuracy_std_dev": 0.02599996410305214,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.171845166,
      "avg_duration_per_task_ms": 631744.54,
      "avg_output_tokens_per_task": 40127.626,
      "last_run_timestamp": "2026-09-26T08:01:18.093Z"
    },
    {
      "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-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Novita",
      "accuracy": 0.7619047142857143,
      "accuracy_std_dev": 0.01942484091285683,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07395918571428572,
      "avg_duration_per_task_ms": 290755.1314285714,
      "avg_output_tokens_per_task": 14557.031428571428,
      "last_run_timestamp": "2026-08-29T17:02:01.051Z"
    },
    {
      "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": 0.7619047142857143,
      "accuracy_std_dev": 0.02487158128532752,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03577183142857143,
      "avg_duration_per_task_ms": 462004.73142857145,
      "avg_output_tokens_per_task": 13726.342857142858,
      "last_run_timestamp": "2026-09-26T08:07:18.969Z"
    },
    {
      "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": 0.76166675,
      "accuracy_std_dev": 0.03312451676911681,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.01217459,
      "avg_duration_per_task_ms": 159488.2175,
      "avg_output_tokens_per_task": 14638.645,
      "last_run_timestamp": "2026-08-29T16:53:19.610Z"
    },
    {
      "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": 0.76166675,
      "accuracy_std_dev": 0.024209737481796473,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.16434805749999998,
      "avg_duration_per_task_ms": 662935.715,
      "avg_output_tokens_per_task": 26007.7175,
      "last_run_timestamp": "2026-08-29T16:14:58.275Z"
    },
    {
      "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": 0.76166675,
      "accuracy_std_dev": 0.028431218171008805,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.051254930000000004,
      "avg_duration_per_task_ms": 163368.405,
      "avg_output_tokens_per_task": 4534.255,
      "last_run_timestamp": "2026-09-26T08:24:09.390Z"
    },
    {
      "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": 0.76166675,
      "accuracy_std_dev": 0.025110926131616508,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.06906778,
      "avg_duration_per_task_ms": 159996.095,
      "avg_output_tokens_per_task": 6808.63,
      "last_run_timestamp": "2026-09-26T08:22:24.923Z"
    },
    {
      "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": 0.761666625,
      "accuracy_std_dev": 0.020749836011023693,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.4161910325,
      "avg_duration_per_task_ms": 98455.6425,
      "avg_output_tokens_per_task": 5629.3925,
      "last_run_timestamp": "2026-09-26T08:10:40.762Z"
    },
    {
      "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": 0.7614156780821918,
      "accuracy_std_dev": 0.0228848139730987,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.024887965753424657,
      "avg_duration_per_task_ms": 458119.4315068493,
      "avg_output_tokens_per_task": 13217.239726027397,
      "last_run_timestamp": "2026-08-29T16:59:36.100Z"
    },
    {
      "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": 0.7612120909090909,
      "accuracy_std_dev": 0.02825828791202695,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.5180262836363636,
      "avg_duration_per_task_ms": 222126.0781818182,
      "avg_output_tokens_per_task": 10227.196363636363,
      "last_run_timestamp": "2026-09-26T08:22:25.160Z"
    },
    {
      "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": 0.7609524285714285,
      "accuracy_std_dev": 0.03369861083129663,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07208951428571428,
      "avg_duration_per_task_ms": 276051.22,
      "avg_output_tokens_per_task": 11448.085714285715,
      "last_run_timestamp": "2026-08-29T16:45:57.840Z"
    },
    {
      "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": 0.7609522857142857,
      "accuracy_std_dev": 0.017969598920035443,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08491382571428571,
      "avg_duration_per_task_ms": 105368.18285714286,
      "avg_output_tokens_per_task": 24364.49142857143,
      "last_run_timestamp": "2026-09-01T17:31:46.348Z"
    },
    {
      "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": "Cohere",
      "accuracy": 0.7609424343434344,
      "accuracy_std_dev": 0.012517,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.02651870707070707,
      "avg_duration_per_task_ms": 84229.25252525252,
      "avg_output_tokens_per_task": 17462.434343434343,
      "last_run_timestamp": "2026-09-26T08:10:56.841Z"
    },
    {
      "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": 0.760833625,
      "accuracy_std_dev": 0.018390020316312206,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.4986317424999999,
      "avg_duration_per_task_ms": 132839.435,
      "avg_output_tokens_per_task": 8114.2475,
      "last_run_timestamp": "2026-09-26T08:16:12.134Z"
    },
    {
      "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": 0.7608334999999999,
      "accuracy_std_dev": 0.02119686166629392,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.16130322249999998,
      "avg_duration_per_task_ms": 442219.4275,
      "avg_output_tokens_per_task": 9533.43,
      "last_run_timestamp": "2026-08-29T16:32:10.657Z"
    },
    {
      "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": 0.7605556666666666,
      "accuracy_std_dev": 0.013391060222734018,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.12239992333333333,
      "avg_duration_per_task_ms": 458072.06,
      "avg_output_tokens_per_task": 22242.516666666666,
      "last_run_timestamp": "2026-09-26T08:22:24.688Z"
    },
    {
      "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": 0.7604689296482412,
      "accuracy_std_dev": 0.02230978235970715,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.10629371356783922,
      "avg_duration_per_task_ms": 377132.17587939696,
      "avg_output_tokens_per_task": 11770.236180904523,
      "last_run_timestamp": "2026-08-29T16:33:36.999Z"
    },
    {
      "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": 0.7604167291666667,
      "accuracy_std_dev": 0.017826499999999967,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.2660930625,
      "avg_duration_per_task_ms": 146365.54166666666,
      "avg_output_tokens_per_task": 6163.291666666667,
      "last_run_timestamp": "2026-09-26T08:05:12.774Z"
    },
    {
      "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": 0.7603703333333334,
      "accuracy_std_dev": 0.028822460188771784,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.056350817777777784,
      "avg_duration_per_task_ms": 98825.95333333334,
      "avg_output_tokens_per_task": 4242.228888888889,
      "last_run_timestamp": "2026-09-26T08:32:07.081Z"
    },
    {
      "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": 0.760080620967742,
      "accuracy_std_dev": 0.02198073466379138,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.03671067338709678,
      "avg_duration_per_task_ms": 531209.4012096775,
      "avg_output_tokens_per_task": 12968.175403225807,
      "last_run_timestamp": "2026-08-29T17:21:09.754Z"
    },
    {
      "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-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "DeepInfra",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.034318637633216124,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.03080999,
      "avg_duration_per_task_ms": 262760.8,
      "avg_output_tokens_per_task": 18663.095,
      "last_run_timestamp": "2026-09-26T08:09:25.342Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05756704,
      "avg_duration_per_task_ms": 256268.94,
      "avg_output_tokens_per_task": 10406.48,
      "last_run_timestamp": "2026-09-29T02:27:33.734Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0.033333,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.11355728,
      "avg_duration_per_task_ms": 147828.6,
      "avg_output_tokens_per_task": 17720.73,
      "last_run_timestamp": "2026-09-01T17:30:58.271Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "AkashML",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.020727419665747104,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.0731778,
      "avg_duration_per_task_ms": 258357.41,
      "avg_output_tokens_per_task": 11852.263333333334,
      "last_run_timestamp": "2026-09-26T08:07:46.003Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0.038682027601113875,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07101471333333334,
      "avg_duration_per_task_ms": 318099.35333333333,
      "avg_output_tokens_per_task": 12493.926666666666,
      "last_run_timestamp": "2026-09-26T08:05:12.731Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.35068044,
      "avg_duration_per_task_ms": 248200.34,
      "avg_output_tokens_per_task": 11571.02,
      "last_run_timestamp": "2026-09-01T12:41:31.788Z"
    },
    {
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "WandB",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12663513999999998,
      "avg_duration_per_task_ms": 73292.09,
      "avg_output_tokens_per_task": 9837.08,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Relace",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.027619090000000002,
      "avg_duration_per_task_ms": 42870.03,
      "avg_output_tokens_per_task": 4628.08,
      "last_run_timestamp": "2026-09-26T09:28:27.720Z"
    },
    {
      "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": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04223252,
      "avg_duration_per_task_ms": 69942.84,
      "avg_output_tokens_per_task": 4621.51,
      "last_run_timestamp": "2026-09-26T08:05:12.737Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0.027371610829992666,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.009037317142857142,
      "avg_duration_per_task_ms": 270209.39714285714,
      "avg_output_tokens_per_task": 6316.0257142857145,
      "last_run_timestamp": "2026-09-26T08:08:19.666Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0.026666999999999996,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.014824000000000002,
      "avg_duration_per_task_ms": 335920.77,
      "avg_output_tokens_per_task": 4448.51,
      "last_run_timestamp": "2026-09-12T08:11:26.740Z"
    },
    {
      "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": 0.759166625,
      "accuracy_std_dev": 0.02319920826201565,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0076381949999999995,
      "avg_duration_per_task_ms": 261034.725,
      "avg_output_tokens_per_task": 15531.6525,
      "last_run_timestamp": "2026-09-26T08:24:41.493Z"
    },
    {
      "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": 0.758888875,
      "accuracy_std_dev": 0.0209643399476995,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.6800024841666666,
      "avg_duration_per_task_ms": 74777.63666666667,
      "avg_output_tokens_per_task": 5635.594166666667,
      "last_run_timestamp": "2026-09-26T08:23:22.613Z"
    },
    {
      "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": 0.7588888333333333,
      "accuracy_std_dev": 0.03578052612067383,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.4306518199999999,
      "avg_duration_per_task_ms": 94862.28333333334,
      "avg_output_tokens_per_task": 5554.29,
      "last_run_timestamp": "2026-09-26T08:25:11.484Z"
    },
    {
      "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": 0.7588406086956522,
      "accuracy_std_dev": 0.021414040391536566,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.59698552,
      "avg_duration_per_task_ms": 123337.09304347826,
      "avg_output_tokens_per_task": 8161.509565217391,
      "last_run_timestamp": "2026-09-26T08:22:57.486Z"
    },
    {
      "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": 0.7586668,
      "accuracy_std_dev": 0.01654634971708262,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.27091952399999997,
      "avg_duration_per_task_ms": 154225.472,
      "avg_output_tokens_per_task": 23085.08,
      "last_run_timestamp": "2026-09-01T17:32:40.487Z"
    },
    {
      "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": 0.7586668,
      "accuracy_std_dev": 0.012927091279943853,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.08354813200000001,
      "avg_duration_per_task_ms": 248803.664,
      "avg_output_tokens_per_task": 12064.656,
      "last_run_timestamp": "2026-09-26T08:08:19.674Z"
    },
    {
      "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": 0.7585185555555556,
      "accuracy_std_dev": 0.01743499783711622,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.07257207555555555,
      "avg_duration_per_task_ms": 58870.41333333333,
      "avg_output_tokens_per_task": 4565.811111111111,
      "last_run_timestamp": "2026-09-26T08:26:14.308Z"
    },
    {
      "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": 0.7583332500000001,
      "accuracy_std_dev": 0.015184156584660873,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.163021,
      "avg_duration_per_task_ms": 227043.59,
      "avg_output_tokens_per_task": 11333.075,
      "last_run_timestamp": "2026-08-29T16:55:33.128Z"
    },
    {
      "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": 0.7583332500000001,
      "accuracy_std_dev": 0.022298623526744884,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.025212460000000003,
      "avg_duration_per_task_ms": 129050.295,
      "avg_output_tokens_per_task": 16100.67,
      "last_run_timestamp": "2026-09-26T08:22:17.628Z"
    },
    {
      "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": "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.7578125491071429,
      "accuracy_std_dev": 0.037360230019845994,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.18927937723214286,
      "avg_duration_per_task_ms": 471319.05133928574,
      "avg_output_tokens_per_task": 27760.915178571428,
      "last_run_timestamp": "2026-09-26T08:12:00.291Z"
    },
    {
      "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": 0.7577778333333334,
      "accuracy_std_dev": 0.037646776132079215,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08977579333333333,
      "avg_duration_per_task_ms": 61527.54,
      "avg_output_tokens_per_task": 12239.113333333333,
      "last_run_timestamp": "2026-08-29T16:25:14.949Z"
    },
    {
      "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": 0.7577776666666667,
      "accuracy_std_dev": 0.0342828683423336,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.08650762222222223,
      "avg_duration_per_task_ms": 226626.72,
      "avg_output_tokens_per_task": 9806.055555555555,
      "last_run_timestamp": "2026-08-29T16:32:40.387Z"
    },
    {
      "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": 0.7576270508474576,
      "accuracy_std_dev": 0.017540967140446465,
      "run_count": 6,
      "total_questions": 295,
      "avg_cost_per_task_usd": 0.11473522711864406,
      "avg_duration_per_task_ms": 198579.0440677966,
      "avg_output_tokens_per_task": 7218.661016949152,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Novita",
      "accuracy": 0.7575758636363636,
      "accuracy_std_dev": 0.02000269988924933,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.09427858838383839,
      "avg_duration_per_task_ms": 374504.8080808081,
      "avg_output_tokens_per_task": 13062.229797979799,
      "last_run_timestamp": "2026-08-29T16:33:12.146Z"
    },
    {
      "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": 0.7575,
      "accuracy_std_dev": 0.023318483371565994,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.34416930749999997,
      "avg_duration_per_task_ms": 221562.1725,
      "avg_output_tokens_per_task": 11914.215,
      "last_run_timestamp": "2026-09-26T08:22:18.965Z"
    },
    {
      "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": 0.7573335999999999,
      "accuracy_std_dev": 0.017178865569064807,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.151421216,
      "avg_duration_per_task_ms": 315167.644,
      "avg_output_tokens_per_task": 20253,
      "last_run_timestamp": "2026-09-26T08:10:25.658Z"
    },
    {
      "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": 0.7573333999999999,
      "accuracy_std_dev": 0.031439640936880967,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.081021904,
      "avg_duration_per_task_ms": 42035.076,
      "avg_output_tokens_per_task": 4841.128,
      "last_run_timestamp": "2026-09-26T08:26:16.831Z"
    },
    {
      "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": 0.7573099970760234,
      "accuracy_std_dev": 0.047450719617777116,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.10973368713450293,
      "avg_duration_per_task_ms": 2695204.277777778,
      "avg_output_tokens_per_task": 24270.733918128655,
      "last_run_timestamp": "2026-08-29T16:45:34.216Z"
    },
    {
      "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": 0.7572254739884392,
      "accuracy_std_dev": 0.020048765907361054,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.014727791907514451,
      "avg_duration_per_task_ms": 418785.12427745666,
      "avg_output_tokens_per_task": 15970.300578034683,
      "last_run_timestamp": "2026-08-29T16:29:07.425Z"
    },
    {
      "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": 0.756944,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.2442584375,
      "avg_duration_per_task_ms": 645873.5208333334,
      "avg_output_tokens_per_task": 12350.729166666666,
      "last_run_timestamp": "2026-09-26T08:23:04.076Z"
    },
    {
      "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": 0.7568027346938776,
      "accuracy_std_dev": 0.038809665680729605,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.3020268469387756,
      "avg_duration_per_task_ms": 145185.1581632653,
      "avg_output_tokens_per_task": 20600.117346938776,
      "last_run_timestamp": "2026-09-01T17:31:47.860Z"
    },
    {
      "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": 0.7567114127516779,
      "accuracy_std_dev": 0.02249334300085622,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.020425387583892618,
      "avg_duration_per_task_ms": 519535.90939597314,
      "avg_output_tokens_per_task": 17284.703020134228,
      "last_run_timestamp": "2026-09-26T08:07:18.967Z"
    },
    {
      "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": 0.7566669999999999,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01828589,
      "avg_duration_per_task_ms": 240722.54,
      "avg_output_tokens_per_task": 15147.8,
      "last_run_timestamp": "2026-09-19T08:02:39.158Z"
    },
    {
      "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": 0.7566669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05416404,
      "avg_duration_per_task_ms": 1008004.72,
      "avg_output_tokens_per_task": 38151.6,
      "last_run_timestamp": "2026-09-26T00:00:04.127Z"
    },
    {
      "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": 0.7566667499999999,
      "accuracy_std_dev": 0.013743584364622665,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.33711718500000004,
      "avg_duration_per_task_ms": 415294.135,
      "avg_output_tokens_per_task": 10214.705,
      "last_run_timestamp": "2026-09-26T08:13:27.132Z"
    },
    {
      "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": 0.7566665,
      "accuracy_std_dev": 0.02687425451214104,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.54838445,
      "avg_duration_per_task_ms": 178864.39333333334,
      "avg_output_tokens_per_task": 10795.433333333332,
      "last_run_timestamp": "2026-09-26T08:26:47.832Z"
    },
    {
      "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": 0.7566665,
      "accuracy_std_dev": 0.033720051184965114,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.010374689999999999,
      "avg_duration_per_task_ms": 242049.23333333334,
      "avg_output_tokens_per_task": 15360.75,
      "last_run_timestamp": "2026-08-29T16:54:19.718Z"
    },
    {
      "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": 0.7566665,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.10920926999999998,
      "avg_duration_per_task_ms": 101251.93,
      "avg_output_tokens_per_task": 25100.99,
      "last_run_timestamp": "2026-09-01T17:23:16.205Z"
    },
    {
      "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": 0.7566665,
      "accuracy_std_dev": 0.043333500000000025,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.42697405000000005,
      "avg_duration_per_task_ms": 522179.69,
      "avg_output_tokens_per_task": 37590.96,
      "last_run_timestamp": "2026-09-01T12:41:24.474Z"
    },
    {
      "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": 0.7566665,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.29550982,
      "avg_duration_per_task_ms": 152697.48,
      "avg_output_tokens_per_task": 10806.55,
      "last_run_timestamp": "2026-09-01T17:30:15.770Z"
    },
    {
      "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": 0.756,
      "accuracy_std_dev": 0.012365076271499492,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.0694329,
      "avg_duration_per_task_ms": 122849.852,
      "avg_output_tokens_per_task": 4849.072,
      "last_run_timestamp": "2026-09-26T08:26:14.187Z"
    },
    {
      "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": 0.7559523794642857,
      "accuracy_std_dev": 0.05190211941817757,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.07933722767857142,
      "avg_duration_per_task_ms": 255983.8861607143,
      "avg_output_tokens_per_task": 9718.457589285714,
      "last_run_timestamp": "2026-08-29T16:45:34.327Z"
    },
    {
      "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": 0.7555556666666666,
      "accuracy_std_dev": 0.012570472951413688,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05382074666666667,
      "avg_duration_per_task_ms": 90716.92666666667,
      "avg_output_tokens_per_task": 14804.206666666667,
      "last_run_timestamp": "2026-09-26T08:22:43.308Z"
    },
    {
      "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": 0.7555556666666666,
      "accuracy_std_dev": 0.03823241466202323,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0327312,
      "avg_duration_per_task_ms": 153585.12,
      "avg_output_tokens_per_task": 15537.24,
      "last_run_timestamp": "2026-09-26T08:17:02.789Z"
    },
    {
      "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": 0.7555556666666666,
      "accuracy_std_dev": 0.011331263428624758,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05866996,
      "avg_duration_per_task_ms": 356562.5266666667,
      "avg_output_tokens_per_task": 13311.113333333333,
      "last_run_timestamp": "2026-09-26T08:12:56.927Z"
    },
    {
      "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": 0.7555556666666666,
      "accuracy_std_dev": 0.012570472951413688,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06505471333333333,
      "avg_duration_per_task_ms": 157291.16,
      "avg_output_tokens_per_task": 14450.066666666668,
      "last_run_timestamp": "2026-09-26T08:05:12.747Z"
    },
    {
      "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": 0.7554439597989949,
      "accuracy_std_dev": 0.00981512802705599,
      "run_count": 4,
      "total_questions": 199,
      "avg_cost_per_task_usd": 0.047420793969849245,
      "avg_duration_per_task_ms": 374713.7035175879,
      "avg_output_tokens_per_task": 7841.51256281407,
      "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": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7553623043478261,
      "accuracy_std_dev": 0.032979826821380405,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.19905160782608697,
      "avg_duration_per_task_ms": 122704.35391304348,
      "avg_output_tokens_per_task": 7790.118260869565,
      "last_run_timestamp": "2026-09-26T08:11:40.828Z"
    },
    {
      "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": 0.7552381428571429,
      "accuracy_std_dev": 0.009897295726302106,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03504835714285714,
      "avg_duration_per_task_ms": 271800.3542857143,
      "avg_output_tokens_per_task": 13667.517142857143,
      "last_run_timestamp": "2026-09-26T08:10:40.746Z"
    },
    {
      "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": 0.7551515454545454,
      "accuracy_std_dev": 0.023968159933786534,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.024316309090909087,
      "avg_duration_per_task_ms": 165689.23272727273,
      "avg_output_tokens_per_task": 18155.507272727275,
      "last_run_timestamp": "2026-09-26T08:08:49.576Z"
    },
    {
      "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": 0.755,
      "accuracy_std_dev": 0.02279120200866991,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.346774905,
      "avg_duration_per_task_ms": 239775.925,
      "avg_output_tokens_per_task": 12209.34,
      "last_run_timestamp": "2026-09-29T03:00:06.005Z"
    },
    {
      "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": 0.755,
      "accuracy_std_dev": 0.017240198505237704,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.055826839999999996,
      "avg_duration_per_task_ms": 122326.96,
      "avg_output_tokens_per_task": 4308.39,
      "last_run_timestamp": "2026-09-26T08:26:39.685Z"
    },
    {
      "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": 0.7546666,
      "accuracy_std_dev": 0.034615490431886144,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.201248028,
      "avg_duration_per_task_ms": 612127.586,
      "avg_output_tokens_per_task": 28603.458,
      "last_run_timestamp": "2026-09-26T08:13:27.139Z"
    },
    {
      "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": 0.7546664000000001,
      "accuracy_std_dev": 0.04509740371950474,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.17806932799999997,
      "avg_duration_per_task_ms": 249010.144,
      "avg_output_tokens_per_task": 9041.408,
      "last_run_timestamp": "2026-09-26T08:26:47.380Z"
    },
    {
      "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": 0.7544445,
      "accuracy_std_dev": 0.025939233358948773,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.19473402666666667,
      "avg_duration_per_task_ms": 127608.55,
      "avg_output_tokens_per_task": 7689.33,
      "last_run_timestamp": "2026-09-26T08:22:38.150Z"
    },
    {
      "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": 0.7543103448275862,
      "accuracy_std_dev": 0.016399321098441732,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.04285599137931035,
      "avg_duration_per_task_ms": 524163.9051724138,
      "avg_output_tokens_per_task": 22101.97988505747,
      "last_run_timestamp": "2026-09-26T16:00:25.681Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "StreamLake",
      "accuracy": 0.7542857142857143,
      "accuracy_std_dev": 0.02838017896205673,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07299003142857143,
      "avg_duration_per_task_ms": 266110.34571428574,
      "avg_output_tokens_per_task": 9872.694285714286,
      "last_run_timestamp": "2026-08-29T17:22:46.177Z"
    },
    {
      "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": 0.754166625,
      "accuracy_std_dev": 0.02895150279319496,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1985393725,
      "avg_duration_per_task_ms": 132141.125,
      "avg_output_tokens_per_task": 7833.895,
      "last_run_timestamp": "2026-09-26T08:08:19.638Z"
    },
    {
      "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": 0.754,
      "accuracy_std_dev": 0.02048284919633984,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.08895154,
      "avg_duration_per_task_ms": 301180.94,
      "avg_output_tokens_per_task": 9633.452,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cloudflare",
      "accuracy": 0.7539393636363636,
      "accuracy_std_dev": 0.01999089692049736,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.024043370909090908,
      "avg_duration_per_task_ms": 299110.4909090909,
      "avg_output_tokens_per_task": 17382.84909090909,
      "last_run_timestamp": "2026-09-26T08:12:56.930Z"
    },
    {
      "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": 0.7539393636363636,
      "accuracy_std_dev": 0.014342041050845328,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.05402639454545455,
      "avg_duration_per_task_ms": 296059.4763636364,
      "avg_output_tokens_per_task": 31833.65090909091,
      "last_run_timestamp": "2026-09-26T16:00:20.167Z"
    },
    {
      "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": 0.7538536705202312,
      "accuracy_std_dev": 0.022320967608497434,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.20567977745664742,
      "avg_duration_per_task_ms": 493851.28612716764,
      "avg_output_tokens_per_task": 20258.49710982659,
      "last_run_timestamp": "2026-09-26T08:08:00.526Z"
    },
    {
      "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": 0.7533668080808081,
      "accuracy_std_dev": 0.00967513458239728,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0209760505050505,
      "avg_duration_per_task_ms": 99025.76767676767,
      "avg_output_tokens_per_task": 14135.353535353535,
      "last_run_timestamp": "2026-09-26T08:05:09.043Z"
    },
    {
      "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": 0.7533630986547086,
      "accuracy_std_dev": 0.021783295924476934,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.07844548430493273,
      "avg_duration_per_task_ms": 256047.19506726458,
      "avg_output_tokens_per_task": 9877.659192825113,
      "last_run_timestamp": "2026-08-29T16:07:02.345Z"
    },
    {
      "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": 0.7533335,
      "accuracy_std_dev": 0.026666500000000037,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.5044717400000001,
      "avg_duration_per_task_ms": 153665.59,
      "avg_output_tokens_per_task": 8278.2,
      "last_run_timestamp": "2026-09-01T12:41:31.807Z"
    },
    {
      "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": 0.7533335,
      "accuracy_std_dev": 0.014142371326973414,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.19823152,
      "avg_duration_per_task_ms": 359205.05,
      "avg_output_tokens_per_task": 12583.985,
      "last_run_timestamp": "2026-09-26T08:10:43.550Z"
    },
    {
      "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": 0.7533333999999999,
      "accuracy_std_dev": 0.01738441192102859,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.019357872,
      "avg_duration_per_task_ms": 107397.768,
      "avg_output_tokens_per_task": 16129.496,
      "last_run_timestamp": "2026-09-26T08:24:48.615Z"
    },
    {
      "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": 0.7533333999999999,
      "accuracy_std_dev": 0.0193218356643462,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.037210308,
      "avg_duration_per_task_ms": 202748.952,
      "avg_output_tokens_per_task": 17567.136,
      "last_run_timestamp": "2026-09-26T08:23:54.132Z"
    },
    {
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Venice",
      "accuracy": 0.7533332857142857,
      "accuracy_std_dev": 0.027371645624698595,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.022015817142857143,
      "avg_duration_per_task_ms": 446424.36,
      "avg_output_tokens_per_task": 15560.805714285714,
      "last_run_timestamp": "2026-08-29T16:47:28.160Z"
    },
    {
      "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": 0.7533331428571429,
      "accuracy_std_dev": 0.023904532346270185,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11555654857142857,
      "avg_duration_per_task_ms": 316732.6342857143,
      "avg_output_tokens_per_task": 9759.925714285715,
      "last_run_timestamp": "2026-08-29T16:44:49.542Z"
    },
    {
      "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": 0.753333,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.36418883999999996,
      "avg_duration_per_task_ms": 138377.24,
      "avg_output_tokens_per_task": 12769.81,
      "last_run_timestamp": "2026-09-01T17:31:47.883Z"
    },
    {
      "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": 0.753333,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.41880896,
      "avg_duration_per_task_ms": 447303.41,
      "avg_output_tokens_per_task": 36551.01,
      "last_run_timestamp": "2026-09-01T12:59:50.177Z"
    },
    {
      "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": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.6857741799999999,
      "avg_duration_per_task_ms": 105619.36,
      "avg_output_tokens_per_task": 7597.08,
      "last_run_timestamp": "2026-09-01T13:02:16.941Z"
    },
    {
      "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": "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Chutes",
      "accuracy": 0.7529762410714286,
      "accuracy_std_dev": 0.03645787970917796,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.10114379464285715,
      "avg_duration_per_task_ms": 414862.9285714286,
      "avg_output_tokens_per_task": 9685.97767857143,
      "last_run_timestamp": "2026-08-29T16:31:41.307Z"
    },
    {
      "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": 0.752923926900585,
      "accuracy_std_dev": 0.02558177145020674,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.006783909356725145,
      "avg_duration_per_task_ms": 266624.59064327483,
      "avg_output_tokens_per_task": 11932.982456140351,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Wafer",
      "accuracy": 0.7525925555555556,
      "accuracy_std_dev": 0.02463451135257706,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021416855555555555,
      "avg_duration_per_task_ms": 123014.89111111111,
      "avg_output_tokens_per_task": 15302.471111111112,
      "last_run_timestamp": "2026-09-26T08:21:23.410Z"
    },
    {
      "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": 0.752499875,
      "accuracy_std_dev": 0.04128993437400179,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.19582053000000002,
      "avg_duration_per_task_ms": 106281.8925,
      "avg_output_tokens_per_task": 7612.88,
      "last_run_timestamp": "2026-09-26T08:23:00.199Z"
    },
    {
      "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": 0.7524445333333333,
      "accuracy_std_dev": 0.03730686310384309,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.19597836800000001,
      "avg_duration_per_task_ms": 104969.73733333334,
      "avg_output_tokens_per_task": 7646.229333333334,
      "last_run_timestamp": "2026-09-26T08:12:13.439Z"
    },
    {
      "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": 0.752,
      "accuracy_std_dev": 0.015434478002187193,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.036405735999999994,
      "avg_duration_per_task_ms": 202702.796,
      "avg_output_tokens_per_task": 18338.232,
      "last_run_timestamp": "2026-09-26T08:01:18.110Z"
    },
    {
      "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": 0.7516667499999999,
      "accuracy_std_dev": 0.009860203404975986,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07850260499999999,
      "avg_duration_per_task_ms": 143951.215,
      "avg_output_tokens_per_task": 4587.75,
      "last_run_timestamp": "2026-09-26T08:22:24.392Z"
    },
    {
      "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": 0.7516667499999999,
      "accuracy_std_dev": 0.0172401743375031,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.14700563,
      "avg_duration_per_task_ms": 224982.975,
      "avg_output_tokens_per_task": 8550.835,
      "last_run_timestamp": "2026-09-26T08:08:19.643Z"
    },
    {
      "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": 0.751666625,
      "accuracy_std_dev": 0.02598077013358873,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.54754274,
      "avg_duration_per_task_ms": 191361.1975,
      "avg_output_tokens_per_task": 10587.185,
      "last_run_timestamp": "2026-09-26T08:16:15.812Z"
    },
    {
      "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": 0.7516665,
      "accuracy_std_dev": 0.03378863221188454,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.025036024999999996,
      "avg_duration_per_task_ms": 332301.835,
      "avg_output_tokens_per_task": 13337.575,
      "last_run_timestamp": "2026-08-29T17:19:05.764Z"
    },
    {
      "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": 0.7516665,
      "accuracy_std_dev": 0.03484088434942489,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.017483123333333336,
      "avg_duration_per_task_ms": 78588.8,
      "avg_output_tokens_per_task": 4208.683333333333,
      "last_run_timestamp": "2026-09-26T08:25:00.207Z"
    },
    {
      "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": 0.7509090909090909,
      "accuracy_std_dev": 0.04884723887507879,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.08554166363636365,
      "avg_duration_per_task_ms": 195446.1018181818,
      "avg_output_tokens_per_task": 7986.087272727273,
      "last_run_timestamp": "2026-09-26T08:09:25.337Z"
    },
    {
      "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": 0.75083325,
      "accuracy_std_dev": 0.025152283418956218,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.014382082499999999,
      "avg_duration_per_task_ms": 226130.345,
      "avg_output_tokens_per_task": 15484.9525,
      "last_run_timestamp": "2026-08-29T16:48:14.107Z"
    },
    {
      "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": 0.7506668,
      "accuracy_std_dev": 0.027840878020637227,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.08836137599999999,
      "avg_duration_per_task_ms": 290555.676,
      "avg_output_tokens_per_task": 12011.224,
      "last_run_timestamp": "2026-09-26T08:10:43.592Z"
    },
    {
      "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": 0.7506666,
      "accuracy_std_dev": 0.029090015088342618,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.008165608,
      "avg_duration_per_task_ms": 191715.892,
      "avg_output_tokens_per_task": 4404.68,
      "last_run_timestamp": "2026-09-26T08:23:40.133Z"
    },
    {
      "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": 0.7506265313283209,
      "accuracy_std_dev": 0.029800800965536136,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.06770081704260651,
      "avg_duration_per_task_ms": 350619.045112782,
      "avg_output_tokens_per_task": 10643.185463659147,
      "last_run_timestamp": "2026-08-29T17:16:16.996Z"
    },
    {
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Venice",
      "accuracy": 0.750303,
      "accuracy_std_dev": 0.04506919157755715,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.11684560363636365,
      "avg_duration_per_task_ms": 153202.71454545454,
      "avg_output_tokens_per_task": 8545.56,
      "last_run_timestamp": "2026-09-26T08:22:30.723Z"
    },
    {
      "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": 0.7500002567567566,
      "accuracy_std_dev": 0.007287030686241284,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.009324743243243245,
      "avg_duration_per_task_ms": 215748.13513513515,
      "avg_output_tokens_per_task": 3777.3783783783783,
      "last_run_timestamp": "2026-09-26T08:06:43.931Z"
    },
    {
      "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": 0.75,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 1.2253275,
      "avg_duration_per_task_ms": 98359.26,
      "avg_output_tokens_per_task": 5434.86,
      "last_run_timestamp": "2026-09-01T13:03:25.552Z"
    },
    {
      "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": 0.75,
      "accuracy_std_dev": 0.027688836500474354,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.5338496725,
      "avg_duration_per_task_ms": 189310.245,
      "avg_output_tokens_per_task": 10232.6975,
      "last_run_timestamp": "2026-09-26T08:09:25.333Z"
    },
    {
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Z.AI",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.03766314634582015,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04774656,
      "avg_duration_per_task_ms": 476512.7733333333,
      "avg_output_tokens_per_task": 9148.09,
      "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": "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": 0.75,
      "accuracy_std_dev": 0.020275820303997587,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.0059813,
      "avg_duration_per_task_ms": 146386.56,
      "avg_output_tokens_per_task": 4218.225,
      "last_run_timestamp": "2026-09-26T08:22:40.372Z"
    },
    {
      "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": 0.7499999583333333,
      "accuracy_std_dev": 0.0368179040090995,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.2266885883333333,
      "avg_duration_per_task_ms": 100830.78416666666,
      "avg_output_tokens_per_task": 7512.6875,
      "last_run_timestamp": "2026-09-26T08:24:10.457Z"
    },
    {
      "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": 0.7496296666666666,
      "accuracy_std_dev": 0.01780858133971749,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021206060000000002,
      "avg_duration_per_task_ms": 122947.83777777778,
      "avg_output_tokens_per_task": 16271.093333333334,
      "last_run_timestamp": "2026-09-26T08:08:19.683Z"
    },
    {
      "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": 0.7493332,
      "accuracy_std_dev": 0.01162353169049754,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.172436168,
      "avg_duration_per_task_ms": 308861.672,
      "avg_output_tokens_per_task": 10652.584,
      "last_run_timestamp": "2026-09-26T08:22:58.738Z"
    },
    {
      "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": 0.748889,
      "accuracy_std_dev": 0.01257094435593444,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05290565333333333,
      "avg_duration_per_task_ms": 112293.75333333333,
      "avg_output_tokens_per_task": 4801.486666666667,
      "last_run_timestamp": "2026-09-26T08:17:02.746Z"
    },
    {
      "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": 0.748889,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.01121336,
      "avg_duration_per_task_ms": 201242.51333333334,
      "avg_output_tokens_per_task": 4454.833333333333,
      "last_run_timestamp": "2026-09-26T08:24:26.075Z"
    },
    {
      "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": 0.7488888888888889,
      "accuracy_std_dev": 0.06409870860277306,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.07423276,
      "avg_duration_per_task_ms": 214855.35777777777,
      "avg_output_tokens_per_task": 9268.002222222221,
      "last_run_timestamp": "2026-08-29T16:58:32.929Z"
    },
    {
      "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": 0.7485715714285714,
      "accuracy_std_dev": 0.030386743739697415,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.27195099142857143,
      "avg_duration_per_task_ms": 529335.9085714286,
      "avg_output_tokens_per_task": 9553.417142857143,
      "last_run_timestamp": "2026-08-29T16:51:55.567Z"
    },
    {
      "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": 0.748333375,
      "accuracy_std_dev": 0.027233626580651628,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.488940975,
      "avg_duration_per_task_ms": 194699.49,
      "avg_output_tokens_per_task": 10000.255,
      "last_run_timestamp": "2026-09-26T08:11:43.276Z"
    },
    {
      "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": 0.7483332500000001,
      "accuracy_std_dev": 0.012802094015726486,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.17867437000000003,
      "avg_duration_per_task_ms": 156388.555,
      "avg_output_tokens_per_task": 12772.235,
      "last_run_timestamp": "2026-09-26T08:22:52.562Z"
    },
    {
      "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": "ModelRun",
      "accuracy": 0.7482992653061223,
      "accuracy_std_dev": 0.011944500000000025,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.033130377551020405,
      "avg_duration_per_task_ms": 119867.04081632652,
      "avg_output_tokens_per_task": 8697.826530612245,
      "last_run_timestamp": "2026-07-11T08:05:22.288Z"
    },
    {
      "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": 0.7481202957393484,
      "accuracy_std_dev": 0.01771080940943976,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.01360957142857143,
      "avg_duration_per_task_ms": 218299.9649122807,
      "avg_output_tokens_per_task": 16333.967418546366,
      "last_run_timestamp": "2026-09-26T08:20:57.091Z"
    },
    {
      "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": 0.748116,
      "accuracy_std_dev": 0.022837863998193896,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.37004660695652175,
      "avg_duration_per_task_ms": 167935.71652173914,
      "avg_output_tokens_per_task": 11884.72,
      "last_run_timestamp": "2026-09-26T08:22:24.476Z"
    },
    {
      "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": 0.7480843850574713,
      "accuracy_std_dev": 0.025275715672864952,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.006891692528735631,
      "avg_duration_per_task_ms": 300685.9626436782,
      "avg_output_tokens_per_task": 13202.612068965518,
      "last_run_timestamp": "2026-08-29T17:00:01.341Z"
    },
    {
      "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": 0.7480002,
      "accuracy_std_dev": 0.014236085605249795,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.14549154800000003,
      "avg_duration_per_task_ms": 139795.724,
      "avg_output_tokens_per_task": 19960.988,
      "last_run_timestamp": "2026-09-26T08:09:25.342Z"
    },
    {
      "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": 0.748,
      "accuracy_std_dev": 0.015434218839967281,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.011429584,
      "avg_duration_per_task_ms": 130555.472,
      "avg_output_tokens_per_task": 4185.616,
      "last_run_timestamp": "2026-09-26T08:24:45.322Z"
    },
    {
      "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": 0.7479998000000001,
      "accuracy_std_dev": 0.023247453386554,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.025147304000000002,
      "avg_duration_per_task_ms": 339787.244,
      "avg_output_tokens_per_task": 4397.436,
      "last_run_timestamp": "2026-09-26T08:13:28.807Z"
    },
    {
      "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": 0.747916625,
      "accuracy_std_dev": 0.029459786032986997,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.56374661625,
      "avg_duration_per_task_ms": 203094.54,
      "avg_output_tokens_per_task": 9850.0775,
      "last_run_timestamp": "2026-09-26T08:23:07.274Z"
    },
    {
      "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": 0.7478787272727273,
      "accuracy_std_dev": 0.029242020409898036,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.33650350909090904,
      "avg_duration_per_task_ms": 209916.4818181818,
      "avg_output_tokens_per_task": 11777.303636363637,
      "last_run_timestamp": "2026-09-26T08:09:55.575Z"
    },
    {
      "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": 0.7477778333333334,
      "accuracy_std_dev": 0.0267822656461116,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07834820999999999,
      "avg_duration_per_task_ms": 157847.65,
      "avg_output_tokens_per_task": 8613.366666666667,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepInfra",
      "accuracy": 0.7477703913043479,
      "accuracy_std_dev": 0.01921375665647317,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.0067899197324414705,
      "avg_duration_per_task_ms": 182508.58026755851,
      "avg_output_tokens_per_task": 12210.97491638796,
      "last_run_timestamp": "2026-09-26T08:06:10.867Z"
    },
    {
      "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": 0.7475668503649635,
      "accuracy_std_dev": 0.021546232673712428,
      "run_count": 11,
      "total_questions": 548,
      "avg_cost_per_task_usd": 0.04062066605839416,
      "avg_duration_per_task_ms": 299491.20802919706,
      "avg_output_tokens_per_task": 26277.279197080294,
      "last_run_timestamp": "2026-09-26T16:00:20.166Z"
    },
    {
      "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": 0.747500125,
      "accuracy_std_dev": 0.03323948977661021,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0216457225,
      "avg_duration_per_task_ms": 167096.8075,
      "avg_output_tokens_per_task": 13839.2125,
      "last_run_timestamp": "2026-09-26T08:10:10.199Z"
    },
    {
      "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": 0.7475,
      "accuracy_std_dev": 0.028565091361660317,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.012148102500000002,
      "avg_duration_per_task_ms": 152199.3025,
      "avg_output_tokens_per_task": 16425.225,
      "last_run_timestamp": "2026-09-26T08:04:53.554Z"
    },
    {
      "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": 0.7474747407407407,
      "accuracy_std_dev": 0.061334593171164214,
      "run_count": 6,
      "total_questions": 297,
      "avg_cost_per_task_usd": 0.07039795622895623,
      "avg_duration_per_task_ms": 90683.95286195286,
      "avg_output_tokens_per_task": 12893.693602693602,
      "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": "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": 0.7466670000000001,
      "accuracy_std_dev": 0.03999999999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.44994164999999997,
      "avg_duration_per_task_ms": 128787.73,
      "avg_output_tokens_per_task": 20836.09,
      "last_run_timestamp": "2026-09-01T17:32:23.560Z"
    },
    {
      "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": 0.746667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.18056328,
      "avg_duration_per_task_ms": 84967.36,
      "avg_output_tokens_per_task": 6058.98,
      "last_run_timestamp": "2026-09-29T12:00:06.611Z"
    },
    {
      "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": 0.746667,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.3847332,
      "avg_duration_per_task_ms": 155324.72,
      "avg_output_tokens_per_task": 13079.53,
      "last_run_timestamp": "2026-09-01T17:31:47.956Z"
    },
    {
      "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": "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": 0.7466667499999999,
      "accuracy_std_dev": 0.03349941956493424,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.54667351,
      "avg_duration_per_task_ms": 183079.6925,
      "avg_output_tokens_per_task": 10525.565,
      "last_run_timestamp": "2026-09-26T08:07:18.989Z"
    },
    {
      "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": 0.7466667499999999,
      "accuracy_std_dev": 0.014907082584714572,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.12188539999999999,
      "avg_duration_per_task_ms": 449678.28,
      "avg_output_tokens_per_task": 19415.325,
      "last_run_timestamp": "2026-09-26T08:23:24.023Z"
    },
    {
      "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": 0.7466667499999999,
      "accuracy_std_dev": 0.012018481141038592,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0546571325,
      "avg_duration_per_task_ms": 218124.0125,
      "avg_output_tokens_per_task": 9184.585,
      "last_run_timestamp": "2026-08-22T16:12:21.444Z"
    },
    {
      "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": 0.7466667499999999,
      "accuracy_std_dev": 0.02494415986132826,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.0097815,
      "avg_duration_per_task_ms": 206039.625,
      "avg_output_tokens_per_task": 4442.965,
      "last_run_timestamp": "2026-09-26T08:26:29.786Z"
    },
    {
      "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": 0.7466666666666667,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.2640489466666667,
      "avg_duration_per_task_ms": 261052.74666666667,
      "avg_output_tokens_per_task": 16845.666666666668,
      "last_run_timestamp": "2026-09-26T08:22:21.004Z"
    },
    {
      "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": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.42183384,
      "avg_duration_per_task_ms": 170456.6,
      "avg_output_tokens_per_task": 28298.04,
      "last_run_timestamp": "2026-09-01T17:30:52.780Z"
    },
    {
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "GMICloud",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.024366269999999995,
      "avg_duration_per_task_ms": 118191.69,
      "avg_output_tokens_per_task": 9023.81,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "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": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04198429,
      "avg_duration_per_task_ms": 312370.3,
      "avg_output_tokens_per_task": 24895.75,
      "last_run_timestamp": "2026-09-26T16:00:24.612Z"
    },
    {
      "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": 0.7466665,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03717445,
      "avg_duration_per_task_ms": 350241.71,
      "avg_output_tokens_per_task": 12501.07,
      "last_run_timestamp": "2026-09-26T08:10:43.553Z"
    },
    {
      "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/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Novita",
      "accuracy": 0.7466664000000001,
      "accuracy_std_dev": 0.03126950946593184,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.172299032,
      "avg_duration_per_task_ms": 325786.904,
      "avg_output_tokens_per_task": 11139.324,
      "last_run_timestamp": "2026-09-26T08:08:19.641Z"
    },
    {
      "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": 0.7466330707070707,
      "accuracy_std_dev": 0.035448714834779266,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07755478535353535,
      "avg_duration_per_task_ms": 198637.78535353535,
      "avg_output_tokens_per_task": 10041.954545454546,
      "last_run_timestamp": "2026-08-29T16:44:10.304Z"
    },
    {
      "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": 0.7457142857142857,
      "accuracy_std_dev": 0.01866272796407924,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.041890954285714294,
      "avg_duration_per_task_ms": 192572.89428571428,
      "avg_output_tokens_per_task": 12803.885714285714,
      "last_run_timestamp": "2026-09-26T08:05:12.818Z"
    },
    {
      "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": 0.7455555,
      "accuracy_std_dev": 0.027599586239350292,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.11502652333333335,
      "avg_duration_per_task_ms": 272510.09,
      "avg_output_tokens_per_task": 12128.78,
      "last_run_timestamp": "2026-09-26T08:22:52.642Z"
    },
    {
      "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": "Anthropic",
      "accuracy": 0.7454546363636363,
      "accuracy_std_dev": 0.026180138836609183,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.35241780545454543,
      "avg_duration_per_task_ms": 231797.27818181817,
      "avg_output_tokens_per_task": 10791.974545454545,
      "last_run_timestamp": "2026-09-26T08:23:01.142Z"
    },
    {
      "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": 0.745416625,
      "accuracy_std_dev": 0.01870356936481312,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.03490152,
      "avg_duration_per_task_ms": 436262.0025,
      "avg_output_tokens_per_task": 9745.805,
      "last_run_timestamp": "2026-08-22T16:09:51.985Z"
    },
    {
      "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": 0.745,
      "accuracy_std_dev": 0.021015867023513466,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.008970824999999998,
      "avg_duration_per_task_ms": 224202.575,
      "avg_output_tokens_per_task": 13976.12,
      "last_run_timestamp": "2026-08-29T16:33:30.698Z"
    },
    {
      "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": 0.745,
      "accuracy_std_dev": 0.009860245661239886,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.45607220499999995,
      "avg_duration_per_task_ms": 481202.88,
      "avg_output_tokens_per_task": 36078.37,
      "last_run_timestamp": "2026-09-01T12:59:55.717Z"
    },
    {
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Wafer",
      "accuracy": 0.7447619999999999,
      "accuracy_std_dev": 0.020381773461320103,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.2414585142857143,
      "avg_duration_per_task_ms": 282563.52571428573,
      "avg_output_tokens_per_task": 10218.734285714285,
      "last_run_timestamp": "2026-09-29T21:00:05.975Z"
    },
    {
      "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": 0.7447618571428571,
      "accuracy_std_dev": 0.035405124090119165,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08685644000000001,
      "avg_duration_per_task_ms": 256492.48857142858,
      "avg_output_tokens_per_task": 12140.245714285715,
      "last_run_timestamp": "2026-09-26T08:23:17.900Z"
    },
    {
      "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": "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": "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": 0.74416675,
      "accuracy_std_dev": 0.021586450870569266,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.32272964499999995,
      "avg_duration_per_task_ms": 199123.9275,
      "avg_output_tokens_per_task": 11446.9275,
      "last_run_timestamp": "2026-09-26T08:10:43.577Z"
    },
    {
      "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": 0.7441077474747474,
      "accuracy_std_dev": 0.03625850000000003,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.012172767676767676,
      "avg_duration_per_task_ms": 161557.24242424243,
      "avg_output_tokens_per_task": 3853.3636363636365,
      "last_run_timestamp": "2026-09-26T08:10:43.660Z"
    },
    {
      "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": 0.7440306848137535,
      "accuracy_std_dev": 0.019647453058182128,
      "run_count": 7,
      "total_questions": 349,
      "avg_cost_per_task_usd": 0.028864432664756447,
      "avg_duration_per_task_ms": 650673.9828080229,
      "avg_output_tokens_per_task": 21865.07736389685,
      "last_run_timestamp": "2026-09-26T16:00:24.615Z"
    },
    {
      "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": 0.744,
      "accuracy_std_dev": 0.019137165704461043,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.006426796,
      "avg_duration_per_task_ms": 131863.972,
      "avg_output_tokens_per_task": 4262.824,
      "last_run_timestamp": "2026-09-26T08:07:18.986Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Phala",
      "accuracy": 0.7438097142857142,
      "accuracy_std_dev": 0.031240303349150428,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11824320857142855,
      "avg_duration_per_task_ms": 449478.22285714286,
      "avg_output_tokens_per_task": 11476.202857142856,
      "last_run_timestamp": "2026-08-29T16:43:42.334Z"
    },
    {
      "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": 0.74338923575419,
      "accuracy_std_dev": 0.032960776721153426,
      "run_count": 18,
      "total_questions": 895,
      "avg_cost_per_task_usd": 0.11096166927374301,
      "avg_duration_per_task_ms": 107590.68156424581,
      "avg_output_tokens_per_task": 8138.489385474861,
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AkashML",
      "accuracy": 0.7433335,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02746015,
      "avg_duration_per_task_ms": 758204.27,
      "avg_output_tokens_per_task": 13329.06,
      "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": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7433335,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.0032184600000000002,
      "avg_duration_per_task_ms": 64941.97,
      "avg_output_tokens_per_task": 11846.7,
      "last_run_timestamp": "2026-08-19T18:00:02.940Z"
    },
    {
      "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": 0.7433332500000001,
      "accuracy_std_dev": 0.022852157692163335,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.07602055999999999,
      "avg_duration_per_task_ms": 242604.3475,
      "avg_output_tokens_per_task": 11060.65,
      "last_run_timestamp": "2026-08-29T17:19:05.156Z"
    },
    {
      "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": 0.7433330000000001,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.34578612,
      "avg_duration_per_task_ms": 121381.44,
      "avg_output_tokens_per_task": 6958.96,
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "DeepInfra",
      "accuracy": 0.7432883288590605,
      "accuracy_std_dev": 0.0384259139109129,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.058029342281879205,
      "avg_duration_per_task_ms": 234207.55369127516,
      "avg_output_tokens_per_task": 7397.003355704698,
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Venice",
      "accuracy": 0.7432794959677419,
      "accuracy_std_dev": 0.019165579787994942,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.14048193346774193,
      "avg_duration_per_task_ms": 566039.3870967742,
      "avg_output_tokens_per_task": 27554.729838709678,
      "last_run_timestamp": "2026-09-26T08:23:04.265Z"
    },
    {
      "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": 0.7429854713804713,
      "accuracy_std_dev": 0.03977251990661671,
      "run_count": 12,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.022424602693602694,
      "avg_duration_per_task_ms": 322809.1447811448,
      "avg_output_tokens_per_task": 12708.414141414141,
      "last_run_timestamp": "2026-09-26T08:22:59.882Z"
    },
    {
      "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": 0.7428572857142857,
      "accuracy_std_dev": 0.03184420977515506,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07498379428571428,
      "avg_duration_per_task_ms": 408111.76,
      "avg_output_tokens_per_task": 11455.622857142856,
      "last_run_timestamp": "2026-08-29T16:53:19.306Z"
    },
    {
      "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": 0.7427776666666667,
      "accuracy_std_dev": 0.036434422742175517,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3671939166666667,
      "avg_duration_per_task_ms": 137596.83,
      "avg_output_tokens_per_task": 5212.336666666667,
      "last_run_timestamp": "2026-09-01T13:00:30.911Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "GMICloud",
      "accuracy": 0.7426666000000001,
      "accuracy_std_dev": 0.01236473120775379,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.07381416,
      "avg_duration_per_task_ms": 152125.172,
      "avg_output_tokens_per_task": 4690.728,
      "last_run_timestamp": "2026-09-26T08:27:29.717Z"
    },
    {
      "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": 0.7425,
      "accuracy_std_dev": 0.029988405167664416,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.5165171675,
      "avg_duration_per_task_ms": 228220.77,
      "avg_output_tokens_per_task": 10000.7275,
      "last_run_timestamp": "2026-09-26T08:11:12.012Z"
    },
    {
      "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": 0.742499875,
      "accuracy_std_dev": 0.021327474782762628,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0092532975,
      "avg_duration_per_task_ms": 294444.1,
      "avg_output_tokens_per_task": 4179.7575,
      "last_run_timestamp": "2026-09-26T08:24:45.222Z"
    },
    {
      "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": 0.7422221666666666,
      "accuracy_std_dev": 0.009937943254964244,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.34229832999999993,
      "avg_duration_per_task_ms": 115494.89,
      "avg_output_tokens_per_task": 7015.326666666667,
      "last_run_timestamp": "2026-09-26T08:16:17.866Z"
    },
    {
      "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": 0.7422221666666666,
      "accuracy_std_dev": 0.029731460662720387,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.34941253333333333,
      "avg_duration_per_task_ms": 221865.17666666667,
      "avg_output_tokens_per_task": 10845.116666666667,
      "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": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Azure",
      "accuracy": 0.7422221666666666,
      "accuracy_std_dev": 0.01791636261835036,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3279842533333333,
      "avg_duration_per_task_ms": 196889.34333333332,
      "avg_output_tokens_per_task": 10978.62,
      "last_run_timestamp": "2026-09-01T17:31:47.053Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Venice",
      "accuracy": 0.7419047142857143,
      "accuracy_std_dev": 0.03059492644080661,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.1254379457142857,
      "avg_duration_per_task_ms": 439339.8028571429,
      "avg_output_tokens_per_task": 9897.565714285714,
      "last_run_timestamp": "2026-08-29T16:44:15.194Z"
    },
    {
      "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": 0.7418180909090909,
      "accuracy_std_dev": 0.04264882202669216,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.09875620727272726,
      "avg_duration_per_task_ms": 153243.34545454546,
      "avg_output_tokens_per_task": 8615.118181818181,
      "last_run_timestamp": "2026-09-26T08:22:32.817Z"
    },
    {
      "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": 0.7416666666666667,
      "accuracy_std_dev": 0.0371060392419108,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.15347501333333333,
      "avg_duration_per_task_ms": 347171.24333333335,
      "avg_output_tokens_per_task": 21806.65,
      "last_run_timestamp": "2026-09-26T08:23:03.682Z"
    },
    {
      "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": 0.741666625,
      "accuracy_std_dev": 0.021794446924259765,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0756606775,
      "avg_duration_per_task_ms": 327680.4725,
      "avg_output_tokens_per_task": 12938.9875,
      "last_run_timestamp": "2026-08-29T16:16:50.461Z"
    },
    {
      "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": 0.7414969591836733,
      "accuracy_std_dev": 0.023315275936604284,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.01141882142857143,
      "avg_duration_per_task_ms": 212540.27551020408,
      "avg_output_tokens_per_task": 4189.09693877551,
      "last_run_timestamp": "2026-09-26T08:26:46.238Z"
    },
    {
      "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": 0.7413600133779265,
      "accuracy_std_dev": 0.021148441080582975,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.06112101672240802,
      "avg_duration_per_task_ms": 177751.98996655518,
      "avg_output_tokens_per_task": 7034.244147157191,
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.7411113333333333,
      "accuracy_std_dev": 0.028588182830595485,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.018602686666666667,
      "avg_duration_per_task_ms": 242873.4,
      "avg_output_tokens_per_task": 16988.533333333333,
      "last_run_timestamp": "2026-09-26T08:21:38.495Z"
    },
    {
      "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": 0.7411111666666667,
      "accuracy_std_dev": 0.02732958194836176,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.34793083666666663,
      "avg_duration_per_task_ms": 202714.10666666666,
      "avg_output_tokens_per_task": 11886.45,
      "last_run_timestamp": "2026-09-26T08:09:25.366Z"
    },
    {
      "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": 0.7411111666666667,
      "accuracy_std_dev": 0.025361580257130863,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.019523156666666666,
      "avg_duration_per_task_ms": 217853.99333333335,
      "avg_output_tokens_per_task": 12622.736666666666,
      "last_run_timestamp": "2026-08-29T16:56:37.261Z"
    },
    {
      "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": 0.7411111666666667,
      "accuracy_std_dev": 0.0418846402213535,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.011101129999999999,
      "avg_duration_per_task_ms": 1049014.8366666667,
      "avg_output_tokens_per_task": 11369.113333333333,
      "last_run_timestamp": "2026-08-29T16:44:33.448Z"
    },
    {
      "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": 0.7411110000000001,
      "accuracy_std_dev": 0.033425712732366185,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03801252333333333,
      "avg_duration_per_task_ms": 199405.32333333333,
      "avg_output_tokens_per_task": 13535.723333333333,
      "last_run_timestamp": "2026-09-26T08:09:22.651Z"
    },
    {
      "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": 0.740909090909091,
      "accuracy_std_dev": 0.034424213616038424,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.3756752018181817,
      "avg_duration_per_task_ms": 205007.96545454545,
      "avg_output_tokens_per_task": 11919.866363636364,
      "last_run_timestamp": "2026-09-26T08:26:37.377Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.7406667,
      "accuracy_std_dev": 0.04386339856201296,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.127795346,
      "avg_duration_per_task_ms": 805643.47,
      "avg_output_tokens_per_task": 30964.946,
      "last_run_timestamp": "2026-09-26T08:01:16.576Z"
    },
    {
      "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": 0.7406059090909092,
      "accuracy_std_dev": 0.025017968199074807,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.024014645454545454,
      "avg_duration_per_task_ms": 391382.4381818182,
      "avg_output_tokens_per_task": 12338.783636363636,
      "last_run_timestamp": "2026-09-26T08:08:19.641Z"
    },
    {
      "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": 0.7404919865771813,
      "accuracy_std_dev": 0.019703479706781643,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.13110433892617449,
      "avg_duration_per_task_ms": 475456.4194630872,
      "avg_output_tokens_per_task": 10412.023489932886,
      "last_run_timestamp": "2026-08-22T08:22:21.071Z"
    },
    {
      "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": 0.7402553306451614,
      "accuracy_std_dev": 0.04059281224847572,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.12427361491935483,
      "avg_duration_per_task_ms": 525015.5120967742,
      "avg_output_tokens_per_task": 25091.802419354837,
      "last_run_timestamp": "2026-09-26T08:23:03.773Z"
    },
    {
      "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": 0.74,
      "accuracy_std_dev": 0.05789968336207521,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.01944233714285714,
      "avg_duration_per_task_ms": 226107.22,
      "avg_output_tokens_per_task": 12391.968571428572,
      "last_run_timestamp": "2026-08-29T16:25:14.953Z"
    },
    {
      "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": "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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Darkbloom",
      "accuracy": 0.74,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.054267579999999996,
      "avg_duration_per_task_ms": 797443.04,
      "avg_output_tokens_per_task": 42406.53,
      "last_run_timestamp": "2026-08-29T16:33:37.702Z"
    },
    {
      "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": 0.74,
      "accuracy_std_dev": 0.02380486645100116,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.09328288000000001,
      "avg_duration_per_task_ms": 826540.2925,
      "avg_output_tokens_per_task": 38601.0775,
      "last_run_timestamp": "2026-09-26T16:00:20.164Z"
    },
    {
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Cerebras",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.03187268323457296,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.4139113142857143,
      "avg_duration_per_task_ms": 36529.922857142854,
      "avg_output_tokens_per_task": 9198.594285714285,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "AtlasCloud",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.03944045441548562,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1116368375,
      "avg_duration_per_task_ms": 298791.685,
      "avg_output_tokens_per_task": 9064.0575,
      "last_run_timestamp": "2026-08-29T16:12:10.433Z"
    },
    {
      "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": 0.74,
      "accuracy_std_dev": 0.01115542718142162,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.014306968,
      "avg_duration_per_task_ms": 188774.736,
      "avg_output_tokens_per_task": 4205.028,
      "last_run_timestamp": "2026-09-26T08:25:40.376Z"
    },
    {
      "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": 0.7396135507246376,
      "accuracy_std_dev": 0.02509467552355785,
      "run_count": 7,
      "total_questions": 345,
      "avg_cost_per_task_usd": 0.0175928115942029,
      "avg_duration_per_task_ms": 309043.7826086957,
      "avg_output_tokens_per_task": 14060.257971014493,
      "last_run_timestamp": "2026-09-26T08:21:25.884Z"
    },
    {
      "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": 0.7395366098654709,
      "accuracy_std_dev": 0.01837246202847226,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.11281150224215246,
      "avg_duration_per_task_ms": 100432.45515695067,
      "avg_output_tokens_per_task": 9897.378923766817,
      "last_run_timestamp": "2026-08-30T00:01:10.606Z"
    },
    {
      "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": 0.739160162601626,
      "accuracy_std_dev": 0.016384230901693223,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.02238267073170732,
      "avg_duration_per_task_ms": 118042.10162601626,
      "avg_output_tokens_per_task": 3823.4349593495936,
      "last_run_timestamp": "2026-09-26T08:17:01.644Z"
    },
    {
      "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": 0.7391112,
      "accuracy_std_dev": 0.02063675876746798,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.5042409586666666,
      "avg_duration_per_task_ms": 734521.768,
      "avg_output_tokens_per_task": 9377.424,
      "last_run_timestamp": "2026-09-26T08:10:56.862Z"
    },
    {
      "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": 0.7390475714285715,
      "accuracy_std_dev": 0.014876513626299345,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.007957537142857143,
      "avg_duration_per_task_ms": 146934.60285714286,
      "avg_output_tokens_per_task": 13560.128571428571,
      "last_run_timestamp": "2026-08-29T17:19:58.529Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Chutes",
      "accuracy": 0.7387387567567568,
      "accuracy_std_dev": 0.03311066334489372,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.1412797635135135,
      "avg_duration_per_task_ms": 662200.6148648649,
      "avg_output_tokens_per_task": 26757.62162162162,
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7386935778894472,
      "accuracy_std_dev": 0.06597081195877463,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.24299765829145728,
      "avg_duration_per_task_ms": 432313.4522613065,
      "avg_output_tokens_per_task": 26613.50502512563,
      "last_run_timestamp": "2026-08-29T16:13:54.412Z"
    },
    {
      "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": 0.7386666000000001,
      "accuracy_std_dev": 0.02777674167068556,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.008449704,
      "avg_duration_per_task_ms": 262399.948,
      "avg_output_tokens_per_task": 5465.996,
      "last_run_timestamp": "2026-09-26T08:30:17.131Z"
    },
    {
      "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": 0.7383335,
      "accuracy_std_dev": 0.007264793338423339,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.029026915000000004,
      "avg_duration_per_task_ms": 190934.755,
      "avg_output_tokens_per_task": 18429.705,
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Reka",
      "accuracy": 0.7383335,
      "accuracy_std_dev": 0.01092919128984391,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07321903,
      "avg_duration_per_task_ms": 107189.01,
      "avg_output_tokens_per_task": 4495.765,
      "last_run_timestamp": "2026-09-26T08:22:24.359Z"
    },
    {
      "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": 0.7383335,
      "accuracy_std_dev": 0.01092919128984391,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.023440455,
      "avg_duration_per_task_ms": 211636.695,
      "avg_output_tokens_per_task": 4181.36,
      "last_run_timestamp": "2026-09-26T08:23:00.240Z"
    },
    {
      "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": 0.73833325,
      "accuracy_std_dev": 0.03345812889175814,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0298002275,
      "avg_duration_per_task_ms": 280974.21,
      "avg_output_tokens_per_task": 17047.3125,
      "last_run_timestamp": "2026-09-26T08:28:05.592Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "GMICloud",
      "accuracy": 0.7380952857142857,
      "accuracy_std_dev": 0.034862897001639706,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.10877247142857141,
      "avg_duration_per_task_ms": 324815.4542857143,
      "avg_output_tokens_per_task": 18549.505714285715,
      "last_run_timestamp": "2026-09-26T08:13:27.148Z"
    },
    {
      "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": 0.7380952857142857,
      "accuracy_std_dev": 0.01771508190952629,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07855077714285713,
      "avg_duration_per_task_ms": 317722.2542857143,
      "avg_output_tokens_per_task": 13494.22,
      "last_run_timestamp": "2026-08-29T16:33:11.313Z"
    },
    {
      "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": 0.7371364697986577,
      "accuracy_std_dev": 0.02632276711591866,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.02334386577181208,
      "avg_duration_per_task_ms": 151084.5889261745,
      "avg_output_tokens_per_task": 7290.451342281879,
      "last_run_timestamp": "2026-09-26T16:00:18.484Z"
    },
    {
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02745565,
      "avg_duration_per_task_ms": 321964.94,
      "avg_output_tokens_per_task": 12731.03,
      "last_run_timestamp": "2026-09-26T08:24:06.205Z"
    },
    {
      "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": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "OpenAI",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.05666649999999995,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.5401649000000001,
      "avg_duration_per_task_ms": 161301.85,
      "avg_output_tokens_per_task": 8843.07,
      "last_run_timestamp": "2026-09-01T13:02:17.066Z"
    },
    {
      "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": 0.7364864414414414,
      "accuracy_std_dev": 0.03584140193121722,
      "run_count": 9,
      "total_questions": 444,
      "avg_cost_per_task_usd": 0.09302163738738738,
      "avg_duration_per_task_ms": 492123.9324324324,
      "avg_output_tokens_per_task": 14254.815315315316,
      "last_run_timestamp": "2026-08-29T16:54:12.073Z"
    },
    {
      "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": 0.7364862297297298,
      "accuracy_std_dev": 0.026034309298820803,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.046338074324324324,
      "avg_duration_per_task_ms": 757120.8243243244,
      "avg_output_tokens_per_task": 14026.905405405405,
      "last_run_timestamp": "2026-08-29T16:49:08.522Z"
    },
    {
      "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": 0.7361687676348548,
      "accuracy_std_dev": 0.033251517706865664,
      "run_count": 10,
      "total_questions": 482,
      "avg_cost_per_task_usd": 0.057402344398340244,
      "avg_duration_per_task_ms": 118233.5,
      "avg_output_tokens_per_task": 7620.255186721992,
      "last_run_timestamp": "2026-09-26T08:08:33.476Z"
    },
    {
      "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": 0.7359389551954242,
      "accuracy_std_dev": 0.023276928156711292,
      "run_count": 21,
      "total_questions": 1049,
      "avg_cost_per_task_usd": 0.1205376796949476,
      "avg_duration_per_task_ms": 180044.07244995233,
      "avg_output_tokens_per_task": 9034.192564346997,
      "last_run_timestamp": "2026-09-26T08:08:19.667Z"
    },
    {
      "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": 0.7356555409836066,
      "accuracy_std_dev": 0.012054742658389706,
      "run_count": 5,
      "total_questions": 244,
      "avg_cost_per_task_usd": 0.01272932786885246,
      "avg_duration_per_task_ms": 60416.67213114754,
      "avg_output_tokens_per_task": 3637.938524590164,
      "last_run_timestamp": "2026-09-29T03:00:05.978Z"
    },
    {
      "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": "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": 0.7355556666666666,
      "accuracy_std_dev": 0.012570472951413688,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.020994060000000002,
      "avg_duration_per_task_ms": 93700.16666666667,
      "avg_output_tokens_per_task": 4193.253333333333,
      "last_run_timestamp": "2026-09-26T08:26:42.017Z"
    },
    {
      "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": 0.7355553333333333,
      "accuracy_std_dev": 0.003142853940113798,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.13248254666666667,
      "avg_duration_per_task_ms": 94960.21333333333,
      "avg_output_tokens_per_task": 24879.886666666665,
      "last_run_timestamp": "2026-09-01T17:31:46.230Z"
    },
    {
      "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": 0.7355553333333333,
      "accuracy_std_dev": 0.013698571612479268,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.011050406666666665,
      "avg_duration_per_task_ms": 621248.44,
      "avg_output_tokens_per_task": 4366.6466666666665,
      "last_run_timestamp": "2026-09-26T08:07:18.968Z"
    },
    {
      "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": 0.7350002499999999,
      "accuracy_std_dev": 0.008660205926391138,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.03314667,
      "avg_duration_per_task_ms": 134389.755,
      "avg_output_tokens_per_task": 16623.845,
      "last_run_timestamp": "2026-09-26T08:05:12.731Z"
    },
    {
      "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": 0.7348993255033557,
      "accuracy_std_dev": 0.031569685085691364,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.030765446308724832,
      "avg_duration_per_task_ms": 431458.02013422817,
      "avg_output_tokens_per_task": 13334.538590604026,
      "last_run_timestamp": "2026-09-26T08:22:26.272Z"
    },
    {
      "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": "Phala",
      "accuracy": 0.7346937142857143,
      "accuracy_std_dev": 0.03541650000000002,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.1668343367346939,
      "avg_duration_per_task_ms": 415202.8163265306,
      "avg_output_tokens_per_task": 9046.959183673469,
      "last_run_timestamp": "2026-07-11T15:34:16.332Z"
    },
    {
      "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": 0.7342855714285714,
      "accuracy_std_dev": 0.03175860195671242,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.12320952285714286,
      "avg_duration_per_task_ms": 247294.2142857143,
      "avg_output_tokens_per_task": 9957.725714285714,
      "last_run_timestamp": "2026-08-29T16:46:44.782Z"
    },
    {
      "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": 0.7340742222222222,
      "accuracy_std_dev": 0.04842648211642921,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.07961510222222222,
      "avg_duration_per_task_ms": 198839.9311111111,
      "avg_output_tokens_per_task": 8248.502222222221,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Inceptron",
      "accuracy": 0.7339180701754385,
      "accuracy_std_dev": 0.02156426040350099,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.009527433583959901,
      "avg_duration_per_task_ms": 516946.6215538847,
      "avg_output_tokens_per_task": 13232.288220551378,
      "last_run_timestamp": "2026-09-26T08:07:18.969Z"
    },
    {
      "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": 0.7333335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.34495834000000003,
      "avg_duration_per_task_ms": 101333.53,
      "avg_output_tokens_per_task": 6956.68,
      "last_run_timestamp": "2026-09-26T08:22:24.747Z"
    },
    {
      "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": 0.7333335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.09979016999999998,
      "avg_duration_per_task_ms": 95640.78,
      "avg_output_tokens_per_task": 4171.89,
      "last_run_timestamp": "2026-09-01T12:57:13.483Z"
    },
    {
      "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": 0.733333375,
      "accuracy_std_dev": 0.05196139592749965,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0075224875,
      "avg_duration_per_task_ms": 325334.9425,
      "avg_output_tokens_per_task": 15635.5625,
      "last_run_timestamp": "2026-09-26T08:07:31.852Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Mancer 2",
      "accuracy": 0.73333325,
      "accuracy_std_dev": 0.004713927363409425,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.093701365,
      "avg_duration_per_task_ms": 409506.12,
      "avg_output_tokens_per_task": 12562.605,
      "last_run_timestamp": "2026-09-26T08:23:24.418Z"
    },
    {
      "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": 0.7333332,
      "accuracy_std_dev": 0.01633001326882499,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.016486864,
      "avg_duration_per_task_ms": 85790.632,
      "avg_output_tokens_per_task": 4187.072,
      "last_run_timestamp": "2026-09-26T08:05:55.055Z"
    },
    {
      "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": 0.7333331666666667,
      "accuracy_std_dev": 0.012171581962597242,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3451316866666666,
      "avg_duration_per_task_ms": 86668.75666666667,
      "avg_output_tokens_per_task": 6980.716666666666,
      "last_run_timestamp": "2026-09-26T08:17:01.560Z"
    },
    {
      "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": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.34216192,
      "avg_duration_per_task_ms": 118971.98,
      "avg_output_tokens_per_task": 7077.54,
      "last_run_timestamp": "2026-09-26T08:22:25.197Z"
    },
    {
      "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-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Relace",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.13542124,
      "avg_duration_per_task_ms": 450545.7,
      "avg_output_tokens_per_task": 22302.62,
      "last_run_timestamp": "2026-09-28T06:00:03.393Z"
    },
    {
      "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": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.046736959999999994,
      "avg_duration_per_task_ms": 211644.04,
      "avg_output_tokens_per_task": 28167.02,
      "last_run_timestamp": "2026-09-28T18:00:05.897Z"
    },
    {
      "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": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "Venice",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0268046,
      "avg_duration_per_task_ms": 1050187.54,
      "avg_output_tokens_per_task": 22343.12,
      "last_run_timestamp": "2026-09-28T18:00:05.487Z"
    },
    {
      "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": 0.7329933877551019,
      "accuracy_std_dev": 0.02075268657909138,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.05900951530612245,
      "avg_duration_per_task_ms": 460991.887755102,
      "avg_output_tokens_per_task": 12569.994897959185,
      "last_run_timestamp": "2026-09-26T08:23:24.611Z"
    },
    {
      "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": 0.7328571428571429,
      "accuracy_std_dev": 0.034013600054887995,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.0738926457142857,
      "avg_duration_per_task_ms": 257306.79142857142,
      "avg_output_tokens_per_task": 11384.062857142857,
      "last_run_timestamp": "2026-08-29T17:08:03.684Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Crusoe",
      "accuracy": 0.7326621543624161,
      "accuracy_std_dev": 0.030789543369281885,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.09465771812080537,
      "avg_duration_per_task_ms": 390427.067114094,
      "avg_output_tokens_per_task": 11230.369127516778,
      "last_run_timestamp": "2026-08-29T16:33:38.512Z"
    },
    {
      "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": 0.732381,
      "accuracy_std_dev": 0.026530205265481322,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07102172,
      "avg_duration_per_task_ms": 255607.41142857142,
      "avg_output_tokens_per_task": 7988.711428571429,
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "auto-routing",
      "accuracy": 0.7316890269058296,
      "accuracy_std_dev": 0.028135970753590752,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.03146767488789237,
      "avg_duration_per_task_ms": 485485.01569506724,
      "avg_output_tokens_per_task": 14361.15470852018,
      "last_run_timestamp": "2026-08-29T16:11:34.927Z"
    },
    {
      "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": 0.73166675,
      "accuracy_std_dev": 0.03210208110368392,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.00461813,
      "avg_duration_per_task_ms": 254807.5,
      "avg_output_tokens_per_task": 4301.06,
      "last_run_timestamp": "2026-09-26T08:25:43.420Z"
    },
    {
      "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": 0.7315434328859061,
      "accuracy_std_dev": 0.019791500921937582,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.007604364093959733,
      "avg_duration_per_task_ms": 318961.6610738255,
      "avg_output_tokens_per_task": 13646.53187919463,
      "last_run_timestamp": "2026-09-26T08:05:12.821Z"
    },
    {
      "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": 0.7314287142857142,
      "accuracy_std_dev": 0.02301532369478017,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07236491428571429,
      "avg_duration_per_task_ms": 89556.96,
      "avg_output_tokens_per_task": 10594.671428571428,
      "last_run_timestamp": "2026-09-26T16:00:24.624Z"
    },
    {
      "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": 0.7314287142857142,
      "accuracy_std_dev": 0.009897460680314293,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.39489586285714284,
      "avg_duration_per_task_ms": 142184.54857142858,
      "avg_output_tokens_per_task": 11046.07142857143,
      "last_run_timestamp": "2026-09-26T08:21:22.720Z"
    },
    {
      "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": 0.7311110000000001,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012396426666666667,
      "avg_duration_per_task_ms": 337807.9,
      "avg_output_tokens_per_task": 15691.053333333333,
      "last_run_timestamp": "2026-09-26T08:22:46.339Z"
    },
    {
      "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": 0.7311110000000001,
      "accuracy_std_dev": 0.013698950130089036,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.025903600000000002,
      "avg_duration_per_task_ms": 117172.05333333333,
      "avg_output_tokens_per_task": 15158.646666666667,
      "last_run_timestamp": "2026-09-26T08:07:18.966Z"
    },
    {
      "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": 0.7311110000000001,
      "accuracy_std_dev": 0.035416115465514666,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.3332058333333333,
      "avg_duration_per_task_ms": 117249.60666666667,
      "avg_output_tokens_per_task": 11086.966666666667,
      "last_run_timestamp": "2026-09-01T17:31:47.546Z"
    },
    {
      "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": 0.730909,
      "accuracy_std_dev": 0.02416641341351107,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.06488479090909091,
      "avg_duration_per_task_ms": 624903.3563636363,
      "avg_output_tokens_per_task": 8185.998181818181,
      "last_run_timestamp": "2026-09-26T08:22:17.866Z"
    },
    {
      "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": 0.730909,
      "accuracy_std_dev": 0.028180123769003633,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.06749359090909092,
      "avg_duration_per_task_ms": 291533.3,
      "avg_output_tokens_per_task": 9954.196363636363,
      "last_run_timestamp": "2026-09-26T08:23:16.657Z"
    },
    {
      "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": 0.7307692064777328,
      "accuracy_std_dev": 0.04749195792721122,
      "run_count": 5,
      "total_questions": 247,
      "avg_cost_per_task_usd": 0.15509178542510121,
      "avg_duration_per_task_ms": 774237.7125506073,
      "avg_output_tokens_per_task": 29578.534412955465,
      "last_run_timestamp": "2026-08-29T16:30:43.592Z"
    },
    {
      "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": 0.7305263684210526,
      "accuracy_std_dev": 0.033390586105116714,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.35496904315789474,
      "avg_duration_per_task_ms": 206890.78210526315,
      "avg_output_tokens_per_task": 10297.44,
      "last_run_timestamp": "2026-09-26T08:17:01.551Z"
    },
    {
      "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": 0.7304761428571429,
      "accuracy_std_dev": 0.040293813213262725,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.006170451428571428,
      "avg_duration_per_task_ms": 328862.59714285715,
      "avg_output_tokens_per_task": 15818.588571428572,
      "last_run_timestamp": "2026-09-26T08:13:27.139Z"
    },
    {
      "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": 0.7300306912442397,
      "accuracy_std_dev": 0.022919556859094893,
      "run_count": 9,
      "total_questions": 434,
      "avg_cost_per_task_usd": 0.018301880184331796,
      "avg_duration_per_task_ms": 93142.05069124424,
      "avg_output_tokens_per_task": 5748.317972350231,
      "last_run_timestamp": "2026-09-26T16:00:25.658Z"
    },
    {
      "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": 0.7300002,
      "accuracy_std_dev": 0.01633001326882499,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.06427955199999999,
      "avg_duration_per_task_ms": 445840.884,
      "avg_output_tokens_per_task": 8867.492,
      "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-5.2-20260616",
      "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": "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": 0.73,
      "accuracy_std_dev": 0.030489986214056124,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.49638783,
      "avg_duration_per_task_ms": 192253.77333333335,
      "avg_output_tokens_per_task": 9911.976666666667,
      "last_run_timestamp": "2026-09-26T08:22:57.969Z"
    },
    {
      "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": 0.73,
      "accuracy_std_dev": 0.0247205998233255,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.36644244,
      "avg_duration_per_task_ms": 230988.18,
      "avg_output_tokens_per_task": 10810.57375,
      "last_run_timestamp": "2026-09-26T08:22:54.473Z"
    },
    {
      "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": 0.73,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.23651355,
      "avg_duration_per_task_ms": 142691.29,
      "avg_output_tokens_per_task": 22898.27,
      "last_run_timestamp": "2026-09-01T17:31:43.526Z"
    },
    {
      "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": 0.73,
      "accuracy_std_dev": 0.026874047787782167,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.40094411999999996,
      "avg_duration_per_task_ms": 212371.995,
      "avg_output_tokens_per_task": 8595.31,
      "last_run_timestamp": "2026-09-26T08:23:04.164Z"
    },
    {
      "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": 0.73,
      "accuracy_std_dev": 0.016666999999999987,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.30719247000000005,
      "avg_duration_per_task_ms": 162813.34,
      "avg_output_tokens_per_task": 10978.12,
      "last_run_timestamp": "2026-09-01T17:32:02.405Z"
    },
    {
      "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": 0.73,
      "accuracy_std_dev": 0.01374388753228139,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.243089465,
      "avg_duration_per_task_ms": 228014.985,
      "avg_output_tokens_per_task": 12589.82,
      "last_run_timestamp": "2026-09-26T08:11:43.368Z"
    },
    {
      "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": 0.7295868703170029,
      "accuracy_std_dev": 0.031298136757478015,
      "run_count": 7,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.07363949279538905,
      "avg_duration_per_task_ms": 211055.38616714696,
      "avg_output_tokens_per_task": 8076.533141210374,
      "last_run_timestamp": "2026-09-26T08:14:46.950Z"
    },
    {
      "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": 0.7295848835341366,
      "accuracy_std_dev": 0.03127469483608115,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.04057755020080321,
      "avg_duration_per_task_ms": 229097.8734939759,
      "avg_output_tokens_per_task": 9654.106425702812,
      "last_run_timestamp": "2026-09-26T08:25:16.178Z"
    },
    {
      "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": 0.7294807939698492,
      "accuracy_std_dev": 0.09894662195850853,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.011416806532663316,
      "avg_duration_per_task_ms": 215331.50753768845,
      "avg_output_tokens_per_task": 13145.422110552763,
      "last_run_timestamp": "2026-08-29T17:11:34.630Z"
    },
    {
      "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": 0.7293334,
      "accuracy_std_dev": 0.03935030820260496,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.051760342,
      "avg_duration_per_task_ms": 152743.692,
      "avg_output_tokens_per_task": 8580.186,
      "last_run_timestamp": "2026-09-26T08:30:17.801Z"
    },
    {
      "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": "Phala",
      "accuracy": 0.729167,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.044939625000000004,
      "avg_duration_per_task_ms": 321431,
      "avg_output_tokens_per_task": 27916.854166666668,
      "last_run_timestamp": "2026-07-11T08:03:26.385Z"
    },
    {
      "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": 0.72916675,
      "accuracy_std_dev": 0.019346895388860176,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.059703584999999997,
      "avg_duration_per_task_ms": 1171969.47,
      "avg_output_tokens_per_task": 26770.295,
      "last_run_timestamp": "2026-09-26T08:10:43.579Z"
    },
    {
      "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": 0.7290909090909091,
      "accuracy_std_dev": 0.024826342686218476,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.02280834909090909,
      "avg_duration_per_task_ms": 86499.97454545455,
      "avg_output_tokens_per_task": 7271.118181818182,
      "last_run_timestamp": "2026-09-26T16:00:24.613Z"
    },
    {
      "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": 0.7288886666666667,
      "accuracy_std_dev": 0.02572412999932599,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.041228259999999996,
      "avg_duration_per_task_ms": 125095.52,
      "avg_output_tokens_per_task": 23895.02666666667,
      "last_run_timestamp": "2026-09-26T08:16:16.958Z"
    },
    {
      "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": 0.72874888252149,
      "accuracy_std_dev": 0.05217086698179985,
      "run_count": 7,
      "total_questions": 349,
      "avg_cost_per_task_usd": 0.06682716905444126,
      "avg_duration_per_task_ms": 1514284.8567335242,
      "avg_output_tokens_per_task": 8343.650429799427,
      "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": "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": 0.7286667,
      "accuracy_std_dev": 0.04099317937913574,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.106204026,
      "avg_duration_per_task_ms": 169828.326,
      "avg_output_tokens_per_task": 9261.106,
      "last_run_timestamp": "2026-09-26T08:22:58.695Z"
    },
    {
      "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": 0.7285714285714285,
      "accuracy_std_dev": 0.03504452237962912,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.04105087428571428,
      "avg_duration_per_task_ms": 218582.77142857143,
      "avg_output_tokens_per_task": 20493.322857142855,
      "last_run_timestamp": "2026-09-26T08:07:18.966Z"
    },
    {
      "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": 0.7283335,
      "accuracy_std_dev": 0.02279116544299565,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.345777245,
      "avg_duration_per_task_ms": 85393.745,
      "avg_output_tokens_per_task": 6868.74,
      "last_run_timestamp": "2026-09-26T08:25:13.689Z"
    },
    {
      "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": 0.7283270805860806,
      "accuracy_std_dev": 0.03258021639071963,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.09055074542124542,
      "avg_duration_per_task_ms": 216630.95054945056,
      "avg_output_tokens_per_task": 8399.408424908424,
      "last_run_timestamp": "2026-09-26T08:23:18.761Z"
    },
    {
      "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": 0.7281478888888889,
      "accuracy_std_dev": 0.026625571215341294,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.045681291111111105,
      "avg_duration_per_task_ms": 180676.11555555556,
      "avg_output_tokens_per_task": 8297.222222222223,
      "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": "auto-routing",
      "accuracy": 0.728,
      "accuracy_std_dev": 0.029333386364345997,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.032369690000000007,
      "avg_duration_per_task_ms": 162941.07,
      "avg_output_tokens_per_task": 8635.816,
      "last_run_timestamp": "2026-09-26T08:12:13.455Z"
    },
    {
      "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": 0.7276787589285714,
      "accuracy_std_dev": 0.023663327303270934,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.019942993303571428,
      "avg_duration_per_task_ms": 481307.24776785716,
      "avg_output_tokens_per_task": 14518.11830357143,
      "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": "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DeepInfra",
      "accuracy": 0.7273567338262477,
      "accuracy_std_dev": 0.034097962965258875,
      "run_count": 11,
      "total_questions": 541,
      "avg_cost_per_task_usd": 0.10647174121996304,
      "avg_duration_per_task_ms": 340266.1959334566,
      "avg_output_tokens_per_task": 15129.060998151572,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Inceptron",
      "accuracy": 0.7272349799498746,
      "accuracy_std_dev": 0.02832567897437543,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.07201908771929824,
      "avg_duration_per_task_ms": 227357.91979949875,
      "avg_output_tokens_per_task": 11495.062656641605,
      "last_run_timestamp": "2026-08-29T16:55:12.780Z"
    },
    {
      "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": 0.7272108979591837,
      "accuracy_std_dev": 0.019828664901097113,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.020857016326530614,
      "avg_duration_per_task_ms": 240921.06530612244,
      "avg_output_tokens_per_task": 10264.273469387756,
      "last_run_timestamp": "2026-09-26T08:13:28.724Z"
    },
    {
      "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": 0.7271692397260274,
      "accuracy_std_dev": 0.01097209078627325,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.12186630821917807,
      "avg_duration_per_task_ms": 191102.53424657535,
      "avg_output_tokens_per_task": 9680.431506849314,
      "last_run_timestamp": "2026-09-26T16:00:20.148Z"
    },
    {
      "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": 0.7271506693548386,
      "accuracy_std_dev": 0.030421326219611147,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.13253313306451614,
      "avg_duration_per_task_ms": 725346.5120967742,
      "avg_output_tokens_per_task": 29741.665322580644,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.029598939999999997,
      "avg_duration_per_task_ms": 159012.9,
      "avg_output_tokens_per_task": 16200.32,
      "last_run_timestamp": "2026-09-26T08:25:06.306Z"
    },
    {
      "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": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.3914632,
      "avg_duration_per_task_ms": 307698.14,
      "avg_output_tokens_per_task": 11120.02,
      "last_run_timestamp": "2026-09-29T12:00:07.886Z"
    },
    {
      "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": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "SiliconFlow",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06664494,
      "avg_duration_per_task_ms": 229624.28,
      "avg_output_tokens_per_task": 13080.68,
      "last_run_timestamp": "2026-09-23T06:00:08.551Z"
    },
    {
      "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": 0.72666675,
      "accuracy_std_dev": 0.01414201777284626,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.19420406499999998,
      "avg_duration_per_task_ms": 374699.79,
      "avg_output_tokens_per_task": 9514.245,
      "last_run_timestamp": "2026-09-26T08:17:01.582Z"
    },
    {
      "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": 0.7266666363636364,
      "accuracy_std_dev": 0.025897973870460233,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.028845547272727274,
      "avg_duration_per_task_ms": 194541.59636363637,
      "avg_output_tokens_per_task": 9348.309090909092,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "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": 0.7266666,
      "accuracy_std_dev": 0.028906085481088594,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.078362018,
      "avg_duration_per_task_ms": 221052.738,
      "avg_output_tokens_per_task": 8631.572,
      "last_run_timestamp": "2026-09-26T08:23:24.640Z"
    },
    {
      "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": 0.7266665454545455,
      "accuracy_std_dev": 0.021835047354544654,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.07702561272727272,
      "avg_duration_per_task_ms": 268603.45636363636,
      "avg_output_tokens_per_task": 9013.23090909091,
      "last_run_timestamp": "2026-09-26T08:23:24.553Z"
    },
    {
      "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": 0.7266665,
      "accuracy_std_dev": 0.020000166668055548,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.46161753000000005,
      "avg_duration_per_task_ms": 792082.91,
      "avg_output_tokens_per_task": 9376.13,
      "last_run_timestamp": "2026-09-26T08:24:06.679Z"
    },
    {
      "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": 0.7266665,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.20674873,
      "avg_duration_per_task_ms": 122950.24,
      "avg_output_tokens_per_task": 5001.27,
      "last_run_timestamp": "2026-09-01T13:00:51.721Z"
    },
    {
      "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": 0.7266665,
      "accuracy_std_dev": 0.022607791184677924,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.022968040000000002,
      "avg_duration_per_task_ms": 247273.5,
      "avg_output_tokens_per_task": 4817.96,
      "last_run_timestamp": "2026-09-26T08:11:43.349Z"
    },
    {
      "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": 0.7264310101010101,
      "accuracy_std_dev": 0.027838133791617557,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04808676262626262,
      "avg_duration_per_task_ms": 349379.2727272727,
      "avg_output_tokens_per_task": 10200.30303030303,
      "last_run_timestamp": "2026-08-29T16:33:11.651Z"
    },
    {
      "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": 0.7261538461538461,
      "accuracy_std_dev": 0.03891341591641658,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.04461130461538462,
      "avg_duration_per_task_ms": 117488.42923076924,
      "avg_output_tokens_per_task": 8622.856923076923,
      "last_run_timestamp": "2026-09-26T08:17:01.729Z"
    },
    {
      "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": 0.7260000999999999,
      "accuracy_std_dev": 0.02958217406293864,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.009236828,
      "avg_duration_per_task_ms": 354448.192,
      "avg_output_tokens_per_task": 13566.026,
      "last_run_timestamp": "2026-09-26T08:23:24.666Z"
    },
    {
      "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": 0.7259998000000001,
      "accuracy_std_dev": 0.02502449821994441,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.07057247200000001,
      "avg_duration_per_task_ms": 194517.306,
      "avg_output_tokens_per_task": 8403.624,
      "last_run_timestamp": "2026-09-26T08:23:24.525Z"
    },
    {
      "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": 0.72583325,
      "accuracy_std_dev": 0.022714969979894317,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.05062893500000001,
      "avg_duration_per_task_ms": 399514.3275,
      "avg_output_tokens_per_task": 14703.53,
      "last_run_timestamp": "2026-08-29T16:13:54.386Z"
    },
    {
      "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": 0.7252010563002681,
      "accuracy_std_dev": 0.03163209639632926,
      "run_count": 15,
      "total_questions": 746,
      "avg_cost_per_task_usd": 0.1270538310991957,
      "avg_duration_per_task_ms": 103835.95710455764,
      "avg_output_tokens_per_task": 9115.029490616622,
      "last_run_timestamp": "2026-09-26T08:26:14.351Z"
    },
    {
      "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": 0.725,
      "accuracy_std_dev": 0.025000000000000022,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00440039,
      "avg_duration_per_task_ms": 1062673.02,
      "avg_output_tokens_per_task": 13355.81,
      "last_run_timestamp": "2026-09-26T08:07:18.990Z"
    },
    {
      "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": 0.7248498828828829,
      "accuracy_std_dev": 0.09386597077919659,
      "run_count": 9,
      "total_questions": 444,
      "avg_cost_per_task_usd": 0.07434012837837839,
      "avg_duration_per_task_ms": 150951.06756756757,
      "avg_output_tokens_per_task": 13940.813063063064,
      "last_run_timestamp": "2026-08-29T16:57:02.331Z"
    },
    {
      "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": 0.7244446666666666,
      "accuracy_std_dev": 0.003142853940113798,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02107129333333333,
      "avg_duration_per_task_ms": 190752.33333333334,
      "avg_output_tokens_per_task": 8303.633333333333,
      "last_run_timestamp": "2026-09-26T16:00:20.149Z"
    },
    {
      "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": 0.7240537802197802,
      "accuracy_std_dev": 0.10194063410640787,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.14525350366300366,
      "avg_duration_per_task_ms": 610254.7802197802,
      "avg_output_tokens_per_task": 27583.227106227107,
      "last_run_timestamp": "2026-09-26T08:05:12.733Z"
    },
    {
      "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": 0.7238232103746397,
      "accuracy_std_dev": 0.022993657910877263,
      "run_count": 7,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.38125144668587896,
      "avg_duration_per_task_ms": 402514.99711815565,
      "avg_output_tokens_per_task": 9102.43227665706,
      "last_run_timestamp": "2026-09-26T08:24:54.432Z"
    },
    {
      "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": 0.7237036666666666,
      "accuracy_std_dev": 0.02535886411318755,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05174337333333333,
      "avg_duration_per_task_ms": 273069.79333333333,
      "avg_output_tokens_per_task": 8633.697777777777,
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Novita",
      "accuracy": 0.7236365454545454,
      "accuracy_std_dev": 0.030664462731500284,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.022053703636363636,
      "avg_duration_per_task_ms": 116837.80363636363,
      "avg_output_tokens_per_task": 6700.861818181818,
      "last_run_timestamp": "2026-09-26T16:00:25.697Z"
    },
    {
      "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": 0.7234432490842492,
      "accuracy_std_dev": 0.039259986936420826,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.03470968315018315,
      "avg_duration_per_task_ms": 178672.7087912088,
      "avg_output_tokens_per_task": 8952.778388278388,
      "last_run_timestamp": "2026-09-26T08:24:54.334Z"
    },
    {
      "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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Novita",
      "accuracy": 0.7233335,
      "accuracy_std_dev": 0.016666800000300002,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.052515885000000005,
      "avg_duration_per_task_ms": 281245.84,
      "avg_output_tokens_per_task": 8472.745,
      "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": "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": 0.7233335,
      "accuracy_std_dev": 0.06333349999999999,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02959498,
      "avg_duration_per_task_ms": 615820.2,
      "avg_output_tokens_per_task": 25536.79,
      "last_run_timestamp": "2026-09-26T08:17:01.598Z"
    },
    {
      "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": 0.7233330000000001,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.09717974,
      "avg_duration_per_task_ms": 146982.06,
      "avg_output_tokens_per_task": 51201.48,
      "last_run_timestamp": "2026-09-01T17:23:14.324Z"
    },
    {
      "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": 0.722508731958763,
      "accuracy_std_dev": 0.02841921157751388,
      "run_count": 4,
      "total_questions": 194,
      "avg_cost_per_task_usd": 0.14483782474226806,
      "avg_duration_per_task_ms": 621637.6597938144,
      "avg_output_tokens_per_task": 28411.036082474227,
      "last_run_timestamp": "2026-08-22T08:19:48.101Z"
    },
    {
      "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": 0.722499875,
      "accuracy_std_dev": 0.03558044972817764,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.29646921,
      "avg_duration_per_task_ms": 366904.835,
      "avg_output_tokens_per_task": 9370.25,
      "last_run_timestamp": "2026-09-26T08:26:14.256Z"
    },
    {
      "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": 0.7222223333333333,
      "accuracy_std_dev": 0.013698571612479227,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.1329602666666667,
      "avg_duration_per_task_ms": 430783.74,
      "avg_output_tokens_per_task": 33596.9,
      "last_run_timestamp": "2026-09-26T08:22:24.836Z"
    },
    {
      "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": "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Alibaba",
      "accuracy": 0.7220459047619048,
      "accuracy_std_dev": 0.03375849235312024,
      "run_count": 19,
      "total_questions": 945,
      "avg_cost_per_task_usd": 0.11261768253968255,
      "avg_duration_per_task_ms": 146602.55555555556,
      "avg_output_tokens_per_task": 8336.625396825397,
      "last_run_timestamp": "2026-09-26T08:14:43.050Z"
    },
    {
      "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": 0.7214815555555555,
      "accuracy_std_dev": 0.022394489084549912,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05446545555555556,
      "avg_duration_per_task_ms": 331631.70444444445,
      "avg_output_tokens_per_task": 11480.095555555556,
      "last_run_timestamp": "2026-08-29T17:03:20.133Z"
    },
    {
      "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": 0.7214813333333334,
      "accuracy_std_dev": 0.02490029941363937,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.06093124888888888,
      "avg_duration_per_task_ms": 320605.1911111111,
      "avg_output_tokens_per_task": 9726.764444444445,
      "last_run_timestamp": "2026-08-30T00:01:02.154Z"
    },
    {
      "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": 0.7211113333333333,
      "accuracy_std_dev": 0.019309173922142694,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.04666514,
      "avg_duration_per_task_ms": 136735.63333333333,
      "avg_output_tokens_per_task": 27417.226666666666,
      "last_run_timestamp": "2026-09-26T08:23:22.508Z"
    },
    {
      "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": 0.7211113333333333,
      "accuracy_std_dev": 0.011331263428624743,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012200653333333334,
      "avg_duration_per_task_ms": 38919.36666666667,
      "avg_output_tokens_per_task": 3634.8866666666668,
      "last_run_timestamp": "2026-09-26T08:23:23.813Z"
    },
    {
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.720833375,
      "accuracy_std_dev": 0.01898462645759391,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.048760994999999994,
      "avg_duration_per_task_ms": 383295.01,
      "avg_output_tokens_per_task": 9253.3225,
      "last_run_timestamp": "2026-08-29T16:11:23.421Z"
    },
    {
      "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": 0.7207407777777778,
      "accuracy_std_dev": 0.023819327953845365,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.058247844444444453,
      "avg_duration_per_task_ms": 312927.7555555556,
      "avg_output_tokens_per_task": 10357.577777777778,
      "last_run_timestamp": "2026-08-30T00:01:09.484Z"
    },
    {
      "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": 0.720606,
      "accuracy_std_dev": 0.021919006936695594,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.07318695090909091,
      "avg_duration_per_task_ms": 256701.00727272726,
      "avg_output_tokens_per_task": 8429.969090909091,
      "last_run_timestamp": "2026-09-26T08:05:12.739Z"
    },
    {
      "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": 0.72,
      "accuracy_std_dev": 0.0354337331437149,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.330923255,
      "avg_duration_per_task_ms": 203933.605,
      "avg_output_tokens_per_task": 10272.2525,
      "last_run_timestamp": "2026-09-26T08:10:43.603Z"
    },
    {
      "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": "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": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0037487199999999997,
      "avg_duration_per_task_ms": 165993.86,
      "avg_output_tokens_per_task": 15330.68,
      "last_run_timestamp": "2026-09-23T18:00:03.326Z"
    },
    {
      "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": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Azure",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.01845912401316305,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.13026846666666667,
      "avg_duration_per_task_ms": 109723.80666666667,
      "avg_output_tokens_per_task": 4025.5433333333335,
      "last_run_timestamp": "2026-09-01T13:00:52.206Z"
    },
    {
      "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": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.14489558,
      "avg_duration_per_task_ms": 248203.88,
      "avg_output_tokens_per_task": 4971.24,
      "last_run_timestamp": "2026-09-23T15:00:07.657Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7192983846153846,
      "accuracy_std_dev": 0.06419322678515234,
      "run_count": 10,
      "total_questions": 494,
      "avg_cost_per_task_usd": 0.14071399392712552,
      "avg_duration_per_task_ms": 1258098.6842105263,
      "avg_output_tokens_per_task": 26322.05870445344,
      "last_run_timestamp": "2026-09-26T08:12:28.093Z"
    },
    {
      "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": 0.719166875,
      "accuracy_std_dev": 0.03169958486493749,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0506060625,
      "avg_duration_per_task_ms": 962184.9925,
      "avg_output_tokens_per_task": 8508.9975,
      "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-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7191666250000001,
      "accuracy_std_dev": 0.03978245617711878,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.06181229,
      "avg_duration_per_task_ms": 550251.335,
      "avg_output_tokens_per_task": 13930.4,
      "last_run_timestamp": "2026-08-29T17:14:29.092Z"
    },
    {
      "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": 0.7191666250000001,
      "accuracy_std_dev": 0.020666060140587385,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0268572725,
      "avg_duration_per_task_ms": 236107.135,
      "avg_output_tokens_per_task": 10157.88,
      "last_run_timestamp": "2026-08-29T17:16:15.215Z"
    },
    {
      "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": 0.7190475714285715,
      "accuracy_std_dev": 0.02815552369991236,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.3010150485714286,
      "avg_duration_per_task_ms": 421383.48,
      "avg_output_tokens_per_task": 9299.894285714287,
      "last_run_timestamp": "2026-09-26T08:07:18.968Z"
    },
    {
      "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": 0.7189780218978102,
      "accuracy_std_dev": 0.03683748767175827,
      "run_count": 11,
      "total_questions": 548,
      "avg_cost_per_task_usd": 0.10041807299270071,
      "avg_duration_per_task_ms": 119245.32116788322,
      "avg_output_tokens_per_task": 9093.193430656935,
      "last_run_timestamp": "2026-09-26T08:22:36.629Z"
    },
    {
      "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": 0.7188806165413533,
      "accuracy_std_dev": 0.09675689930043425,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.16023803007518797,
      "avg_duration_per_task_ms": 435488.2756892231,
      "avg_output_tokens_per_task": 22759.51127819549,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Chutes",
      "accuracy": 0.7188550454545455,
      "accuracy_std_dev": 0.03234119038253501,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0716665202020202,
      "avg_duration_per_task_ms": 390250.71717171714,
      "avg_output_tokens_per_task": 11350.588383838383,
      "last_run_timestamp": "2026-08-29T16:47:08.827Z"
    },
    {
      "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": 0.7186897976878613,
      "accuracy_std_dev": 0.027015736117292515,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.02169187283236994,
      "avg_duration_per_task_ms": 314303.23988439306,
      "avg_output_tokens_per_task": 7575.85549132948,
      "last_run_timestamp": "2026-08-29T16:48:46.265Z"
    },
    {
      "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": 0.7185185555555555,
      "accuracy_std_dev": 0.01771600069059424,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.10858120444444444,
      "avg_duration_per_task_ms": 333763.2088888889,
      "avg_output_tokens_per_task": 15783.317777777778,
      "last_run_timestamp": "2026-08-29T17:17:21.072Z"
    },
    {
      "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": 0.7185184444444445,
      "accuracy_std_dev": 0.03648484962925207,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.06869702888888889,
      "avg_duration_per_task_ms": 618804.8311111111,
      "avg_output_tokens_per_task": 23662.906666666666,
      "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "Novita",
      "accuracy": 0.7183906609195403,
      "accuracy_std_dev": 0.025974738950584678,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.06343264655172413,
      "avg_duration_per_task_ms": 497586.9425287356,
      "avg_output_tokens_per_task": 14530.419540229885,
      "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": "Phala",
      "accuracy": 0.7177420000000001,
      "accuracy_std_dev": 0.05717499919930039,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.2799884032258065,
      "avg_duration_per_task_ms": 377640.3185483871,
      "avg_output_tokens_per_task": 9988.213709677419,
      "last_run_timestamp": "2026-09-26T08:10:43.671Z"
    },
    {
      "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": 0.717499875,
      "accuracy_std_dev": 0.027474714386311185,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.35365771499999993,
      "avg_duration_per_task_ms": 238504.995,
      "avg_output_tokens_per_task": 10803.1675,
      "last_run_timestamp": "2026-09-26T08:10:56.834Z"
    },
    {
      "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": 0.7173334,
      "accuracy_std_dev": 0.025508314986294175,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.327251336,
      "avg_duration_per_task_ms": 427459.224,
      "avg_output_tokens_per_task": 9077.78,
      "last_run_timestamp": "2026-09-26T08:23:18.418Z"
    },
    {
      "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": 0.7173334,
      "accuracy_std_dev": 0.050526018820405805,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.171608576,
      "avg_duration_per_task_ms": 290919.704,
      "avg_output_tokens_per_task": 9956.296,
      "last_run_timestamp": "2026-09-26T08:22:17.301Z"
    },
    {
      "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": 0.717083375,
      "accuracy_std_dev": 0.015629102796845852,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.03730813999999999,
      "avg_duration_per_task_ms": 202282.19375,
      "avg_output_tokens_per_task": 10076.16875,
      "last_run_timestamp": "2026-08-29T16:54:26.210Z"
    },
    {
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Parasail",
      "accuracy": 0.7166667999999999,
      "accuracy_std_dev": 0.03303193740851419,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.11440428,
      "avg_duration_per_task_ms": 101112.032,
      "avg_output_tokens_per_task": 9922.216,
      "last_run_timestamp": "2026-09-26T08:27:14.783Z"
    },
    {
      "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": 0.7166665000000001,
      "accuracy_std_dev": 0.03666649999999999,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.17924758000000002,
      "avg_duration_per_task_ms": 326369.44,
      "avg_output_tokens_per_task": 9562.9,
      "last_run_timestamp": "2026-09-28T18:00:05.928Z"
    },
    {
      "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": 0.7162963333333333,
      "accuracy_std_dev": 0.030325222992676492,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.12511306444444442,
      "avg_duration_per_task_ms": 317088.20666666667,
      "avg_output_tokens_per_task": 9945.955555555556,
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "GMICloud",
      "accuracy": 0.716111,
      "accuracy_std_dev": 0.031056479812324726,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.006504103333333333,
      "avg_duration_per_task_ms": 381723.9266666667,
      "avg_output_tokens_per_task": 12559.69,
      "last_run_timestamp": "2026-09-26T08:05:12.738Z"
    },
    {
      "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": "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": "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Reka",
      "accuracy": 0.715,
      "accuracy_std_dev": 0.04499999999999999,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.16922785,
      "avg_duration_per_task_ms": 299708.78,
      "avg_output_tokens_per_task": 9901.85,
      "last_run_timestamp": "2026-08-29T08:38:56.982Z"
    },
    {
      "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": 0.7144443333333333,
      "accuracy_std_dev": 0.0015716626723173276,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.04586062666666667,
      "avg_duration_per_task_ms": 212784.84,
      "avg_output_tokens_per_task": 30659.346666666668,
      "last_run_timestamp": "2026-09-26T08:22:27.237Z"
    },
    {
      "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": 0.7142854897959183,
      "accuracy_std_dev": 0.039861000000000035,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.11979946938775511,
      "avg_duration_per_task_ms": 677914.6734693878,
      "avg_output_tokens_per_task": 54287.663265306124,
      "last_run_timestamp": "2026-09-01T17:30:15.922Z"
    },
    {
      "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": "Inceptron",
      "accuracy": 0.7140255173041894,
      "accuracy_std_dev": 0.0345616737838867,
      "run_count": 11,
      "total_questions": 549,
      "avg_cost_per_task_usd": 0.06628586703096538,
      "avg_duration_per_task_ms": 272000.2349726776,
      "avg_output_tokens_per_task": 8738.27504553734,
      "last_run_timestamp": "2026-09-26T08:25:04.482Z"
    },
    {
      "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": 0.7133335,
      "accuracy_std_dev": 0.026666499999999982,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.028499629999999998,
      "avg_duration_per_task_ms": 176126.96,
      "avg_output_tokens_per_task": 22001.83,
      "last_run_timestamp": "2026-09-26T08:08:19.643Z"
    },
    {
      "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": 0.7133335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06446003,
      "avg_duration_per_task_ms": 348420.47,
      "avg_output_tokens_per_task": 34152.4,
      "last_run_timestamp": "2026-09-01T12:41:25.939Z"
    },
    {
      "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": 0.7133333333333334,
      "accuracy_std_dev": 0.04268744285956807,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04181986333333333,
      "avg_duration_per_task_ms": 299500.48,
      "avg_output_tokens_per_task": 13842.47,
      "last_run_timestamp": "2026-09-26T08:13:28.082Z"
    },
    {
      "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": 0.7133333333333334,
      "accuracy_std_dev": 0.019625871060640555,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12069053999999999,
      "avg_duration_per_task_ms": 78204.86666666667,
      "avg_output_tokens_per_task": 4256.786666666667,
      "last_run_timestamp": "2026-09-01T12:59:40.225Z"
    },
    {
      "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": 0.7133330000000001,
      "accuracy_std_dev": 0.02828427124746193,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.031477000000000005,
      "avg_duration_per_task_ms": 184080.18666666668,
      "avg_output_tokens_per_task": 16016.74,
      "last_run_timestamp": "2026-09-26T08:01:16.573Z"
    },
    {
      "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": 0.7133330000000001,
      "accuracy_std_dev": 0.040000000000000036,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07565915000000001,
      "avg_duration_per_task_ms": 95308.62,
      "avg_output_tokens_per_task": 3758.94,
      "last_run_timestamp": "2026-09-01T13:00:41.334Z"
    },
    {
      "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": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.043553119999999994,
      "avg_duration_per_task_ms": 152550.44,
      "avg_output_tokens_per_task": 34185.24,
      "last_run_timestamp": "2026-09-26T08:22:24.717Z"
    },
    {
      "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": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.14862698,
      "avg_duration_per_task_ms": 506333.72,
      "avg_output_tokens_per_task": 12832.14,
      "last_run_timestamp": "2026-09-26T16:00:25.688Z"
    },
    {
      "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": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "OpenAI",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.57503108,
      "avg_duration_per_task_ms": 657829.88,
      "avg_output_tokens_per_task": 51958.96,
      "last_run_timestamp": "2026-09-01T13:00:31.045Z"
    },
    {
      "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": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0723164,
      "avg_duration_per_task_ms": 346522.38,
      "avg_output_tokens_per_task": 59152.72,
      "last_run_timestamp": "2026-09-12T16:00:24.399Z"
    },
    {
      "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": 0.7130758729674798,
      "accuracy_std_dev": 0.03492932464076567,
      "run_count": 20,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.09219651422764227,
      "avg_duration_per_task_ms": 75102.03760162601,
      "avg_output_tokens_per_task": 7773.928861788618,
      "last_run_timestamp": "2026-09-26T08:12:44.279Z"
    },
    {
      "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": 0.712,
      "accuracy_std_dev": 0.03249611258904671,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.08037907,
      "avg_duration_per_task_ms": 74455.182,
      "avg_output_tokens_per_task": 8504.276,
      "last_run_timestamp": "2026-09-26T08:11:40.822Z"
    },
    {
      "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": 0.712,
      "accuracy_std_dev": 0.023626914974240687,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.34044856000000007,
      "avg_duration_per_task_ms": 188002.144,
      "avg_output_tokens_per_task": 11667.488,
      "last_run_timestamp": "2026-09-26T08:22:45.792Z"
    },
    {
      "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": 0.7117119324324325,
      "accuracy_std_dev": 0.01247848271581482,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.6902402702702702,
      "avg_duration_per_task_ms": 688360.5945945946,
      "avg_output_tokens_per_task": 56108.03378378379,
      "last_run_timestamp": "2026-09-01T13:02:58.619Z"
    },
    {
      "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": 0.7116669999999999,
      "accuracy_std_dev": 0.008660254037844347,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.013152875000000001,
      "avg_duration_per_task_ms": 1125576.36,
      "avg_output_tokens_per_task": 15885.82,
      "last_run_timestamp": "2026-09-26T08:11:57.022Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.7115385708502023,
      "accuracy_std_dev": 0.03244045710975723,
      "run_count": 10,
      "total_questions": 494,
      "avg_cost_per_task_usd": 0.021128163967611337,
      "avg_duration_per_task_ms": 99825.94129554655,
      "avg_output_tokens_per_task": 6537.080971659919,
      "last_run_timestamp": "2026-09-26T16:00:20.169Z"
    },
    {
      "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": 0.7114285714285714,
      "accuracy_std_dev": 0.0324087457766623,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.034869005714285715,
      "avg_duration_per_task_ms": 406092.0742857143,
      "avg_output_tokens_per_task": 5418.051428571429,
      "last_run_timestamp": "2026-09-30T00:35:04.155Z"
    },
    {
      "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": 0.7111871198630138,
      "accuracy_std_dev": 0.09721728114340919,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.03006977739726028,
      "avg_duration_per_task_ms": 1645574.0684931506,
      "avg_output_tokens_per_task": 44781.23630136986,
      "last_run_timestamp": "2026-09-26T16:00:24.617Z"
    },
    {
      "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": 0.711111,
      "accuracy_std_dev": 0.016629528997138404,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03228578,
      "avg_duration_per_task_ms": 111202.65333333334,
      "avg_output_tokens_per_task": 22255.406666666666,
      "last_run_timestamp": "2026-09-26T08:17:01.712Z"
    },
    {
      "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": 0.711111,
      "accuracy_std_dev": 0.07430303651040564,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.040132593333333334,
      "avg_duration_per_task_ms": 322574.2266666667,
      "avg_output_tokens_per_task": 8636.586666666666,
      "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7106666,
      "accuracy_std_dev": 0.1560397473582933,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.0609269,
      "avg_duration_per_task_ms": 167712.206,
      "avg_output_tokens_per_task": 20248.218,
      "last_run_timestamp": "2026-09-26T16:00:25.686Z"
    },
    {
      "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": 0.7105555,
      "accuracy_std_dev": 0.03132382736166404,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.10523619666666666,
      "avg_duration_per_task_ms": 90620.79166666667,
      "avg_output_tokens_per_task": 10589.016666666666,
      "last_run_timestamp": "2026-09-26T08:22:18.077Z"
    },
    {
      "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": 0.71,
      "accuracy_std_dev": 0.003332999999999975,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.009475860000000001,
      "avg_duration_per_task_ms": 101117.88,
      "avg_output_tokens_per_task": 5437.48,
      "last_run_timestamp": "2026-09-01T17:32:44.473Z"
    },
    {
      "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": 0.7094593108108108,
      "accuracy_std_dev": 0.025596061967072654,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.29917547972972974,
      "avg_duration_per_task_ms": 177661.23648648648,
      "avg_output_tokens_per_task": 4943.472972972973,
      "last_run_timestamp": "2026-09-01T13:00:41.783Z"
    },
    {
      "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": 0.709333,
      "accuracy_std_dev": 0.034409301068170535,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.23672154,
      "avg_duration_per_task_ms": 252134.184,
      "avg_output_tokens_per_task": 20169.304,
      "last_run_timestamp": "2026-09-26T08:14:47.038Z"
    },
    {
      "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": 0.7091838979591837,
      "accuracy_std_dev": 0.022986000000000006,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.021414438775510206,
      "avg_duration_per_task_ms": 174657.33673469388,
      "avg_output_tokens_per_task": 20207.65306122449,
      "last_run_timestamp": "2026-09-26T08:07:31.848Z"
    },
    {
      "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": 0.7091665,
      "accuracy_std_dev": 0.012774768970513724,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.01202336,
      "avg_duration_per_task_ms": 319678.925,
      "avg_output_tokens_per_task": 17962.82,
      "last_run_timestamp": "2026-09-26T08:14:04.244Z"
    },
    {
      "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": 0.708889,
      "accuracy_std_dev": 0.019115994210782406,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.01979013333333333,
      "avg_duration_per_task_ms": 164929.67333333334,
      "avg_output_tokens_per_task": 14302.1,
      "last_run_timestamp": "2026-09-26T08:17:02.501Z"
    },
    {
      "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": 0.7088886666666667,
      "accuracy_std_dev": 0.0339208519579853,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08842117999999999,
      "avg_duration_per_task_ms": 151404.06666666668,
      "avg_output_tokens_per_task": 46789.46333333333,
      "last_run_timestamp": "2026-09-01T17:31:47.762Z"
    },
    {
      "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": 0.7088886666666667,
      "accuracy_std_dev": 0.025724129999325985,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.062267260000000005,
      "avg_duration_per_task_ms": 262899.92,
      "avg_output_tokens_per_task": 8890.053333333333,
      "last_run_timestamp": "2026-09-26T08:07:15.090Z"
    },
    {
      "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": 0.7087542626262627,
      "accuracy_std_dev": 0.01740072369903045,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.09113491919191918,
      "avg_duration_per_task_ms": 1081759.01010101,
      "avg_output_tokens_per_task": 17133.217171717173,
      "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-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.708333375,
      "accuracy_std_dev": 0.028037521983662796,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.04000606,
      "avg_duration_per_task_ms": 322622.4375,
      "avg_output_tokens_per_task": 9103.8975,
      "last_run_timestamp": "2026-08-29T16:33:11.545Z"
    },
    {
      "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": 0.708,
      "accuracy_std_dev": 0.024367428169587355,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.117385436,
      "avg_duration_per_task_ms": 671239.06,
      "avg_output_tokens_per_task": 25704.02,
      "last_run_timestamp": "2026-09-26T08:22:25.076Z"
    },
    {
      "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": 0.7077728430493274,
      "accuracy_std_dev": 0.03674680352169054,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.0593315,
      "avg_duration_per_task_ms": 484083.41704035876,
      "avg_output_tokens_per_task": 14009.72197309417,
      "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": "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": 0.7077207463235294,
      "accuracy_std_dev": 0.07487438084226404,
      "run_count": 11,
      "total_questions": 544,
      "avg_cost_per_task_usd": 0.04842610477941176,
      "avg_duration_per_task_ms": 275262.8474264706,
      "avg_output_tokens_per_task": 27885.965073529413,
      "last_run_timestamp": "2026-09-26T16:00:24.610Z"
    },
    {
      "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": 0.707206945945946,
      "accuracy_std_dev": 0.06900418229514953,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.23642254054054054,
      "avg_duration_per_task_ms": 438110.88513513515,
      "avg_output_tokens_per_task": 23582.5,
      "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": "openrouter/auto-beta",
      "model_slug": "openrouter/auto-beta",
      "model_name": "Auto Router (Beta)",
      "provider_name": "auto-routing",
      "accuracy": 0.706667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.2407015,
      "avg_duration_per_task_ms": 149076.32,
      "avg_output_tokens_per_task": 17938.72,
      "last_run_timestamp": "2026-07-22T00:25:39.534Z"
    },
    {
      "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": "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": 0.706667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02431272,
      "avg_duration_per_task_ms": 304327.62,
      "avg_output_tokens_per_task": 14297.9,
      "last_run_timestamp": "2026-09-12T08:11:25.132Z"
    },
    {
      "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": "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": 0.706667,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00934721,
      "avg_duration_per_task_ms": 103901.71,
      "avg_output_tokens_per_task": 5395.91,
      "last_run_timestamp": "2026-09-01T17:30:15.713Z"
    },
    {
      "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-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": 0.7066669999999999,
      "accuracy_std_dev": 0.019999999999999962,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06360505,
      "avg_duration_per_task_ms": 327814.04,
      "avg_output_tokens_per_task": 33622.71,
      "last_run_timestamp": "2026-09-01T13:01:01.497Z"
    },
    {
      "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": 0.7066666666666667,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.042424746666666666,
      "avg_duration_per_task_ms": 131868.91333333333,
      "avg_output_tokens_per_task": 15613.653333333334,
      "last_run_timestamp": "2026-09-26T08:23:01.726Z"
    },
    {
      "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": 0.7063491754385965,
      "accuracy_std_dev": 0.03950815373122365,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.13163428320802006,
      "avg_duration_per_task_ms": 73963.29824561403,
      "avg_output_tokens_per_task": 10257.41604010025,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Fireworks",
      "accuracy": 0.706237370221328,
      "accuracy_std_dev": 0.032130056570911926,
      "run_count": 10,
      "total_questions": 497,
      "avg_cost_per_task_usd": 0.37140224144869216,
      "avg_duration_per_task_ms": 217046.72837022133,
      "avg_output_tokens_per_task": 8219.961770623742,
      "last_run_timestamp": "2026-09-26T08:16:17.130Z"
    },
    {
      "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": 0.705833375,
      "accuracy_std_dev": 0.13289250276721548,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.03976693,
      "avg_duration_per_task_ms": 263266.415,
      "avg_output_tokens_per_task": 10599.815,
      "last_run_timestamp": "2026-08-29T17:02:01.052Z"
    },
    {
      "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": 0.7055556666666667,
      "accuracy_std_dev": 0.03430073605462263,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.038495376666666664,
      "avg_duration_per_task_ms": 114809.37333333334,
      "avg_output_tokens_per_task": 5174.493333333333,
      "last_run_timestamp": "2026-09-01T17:31:46.866Z"
    },
    {
      "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": 0.7044674329896906,
      "accuracy_std_dev": 0.02710499999999999,
      "run_count": 2,
      "total_questions": 97,
      "avg_cost_per_task_usd": 0.11629732989690722,
      "avg_duration_per_task_ms": 255742.56701030929,
      "avg_output_tokens_per_task": 22402.164948453606,
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.704,
      "accuracy_std_dev": 0.02370180490173691,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.008245141999999999,
      "avg_duration_per_task_ms": 308896.162,
      "avg_output_tokens_per_task": 13728.714,
      "last_run_timestamp": "2026-09-26T08:22:46.390Z"
    },
    {
      "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": 0.7037787044534413,
      "accuracy_std_dev": 0.09864976994620919,
      "run_count": 10,
      "total_questions": 494,
      "avg_cost_per_task_usd": 0.16661966194331984,
      "avg_duration_per_task_ms": 961787.6214574899,
      "avg_output_tokens_per_task": 43098.3036437247,
      "last_run_timestamp": "2026-09-26T08:11:26.492Z"
    },
    {
      "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": 0.7029631111111111,
      "accuracy_std_dev": 0.042758122148362894,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.06687537777777779,
      "avg_duration_per_task_ms": 328702.3044444444,
      "avg_output_tokens_per_task": 9886.144444444444,
      "last_run_timestamp": "2026-08-30T00:01:09.408Z"
    },
    {
      "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": 0.7022223333333333,
      "accuracy_std_dev": 0.011330936570096727,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.4077153666666667,
      "avg_duration_per_task_ms": 153641.3,
      "avg_output_tokens_per_task": 23369.493333333332,
      "last_run_timestamp": "2026-09-01T17:32:03.083Z"
    },
    {
      "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": 0.701666625,
      "accuracy_std_dev": 0.02598062579566502,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.04259596,
      "avg_duration_per_task_ms": 192678.315,
      "avg_output_tokens_per_task": 8418.055,
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Morph",
      "accuracy": 0.7000002,
      "accuracy_std_dev": 0.02828436552868031,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.142434276,
      "avg_duration_per_task_ms": 459844.396,
      "avg_output_tokens_per_task": 11247.26,
      "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": "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": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.9826790200000001,
      "avg_duration_per_task_ms": 401624.28,
      "avg_output_tokens_per_task": 9843.34,
      "last_run_timestamp": "2026-08-26T00:00:04.572Z"
    },
    {
      "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": 0.7,
      "accuracy_std_dev": 0.01440159456449179,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.017548586666666668,
      "avg_duration_per_task_ms": 296799.9066666667,
      "avg_output_tokens_per_task": 10272.013333333334,
      "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": "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": 0.7,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00948338,
      "avg_duration_per_task_ms": 87248.24,
      "avg_output_tokens_per_task": 5625.9,
      "last_run_timestamp": "2026-09-26T08:10:43.550Z"
    },
    {
      "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": "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-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.6983967174348699,
      "accuracy_std_dev": 0.025880910808740845,
      "run_count": 10,
      "total_questions": 499,
      "avg_cost_per_task_usd": 0.03737082565130261,
      "avg_duration_per_task_ms": 243092.8116232465,
      "avg_output_tokens_per_task": 9769.018036072144,
      "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": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "Azure",
      "accuracy": 0.69833325,
      "accuracy_std_dev": 0.009859949857250788,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.14240881,
      "avg_duration_per_task_ms": 514302.935,
      "avg_output_tokens_per_task": 58927.42,
      "last_run_timestamp": "2026-09-01T17:31:48.421Z"
    },
    {
      "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": 0.698333,
      "accuracy_std_dev": 0.02291287847477922,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.05212012666666667,
      "avg_duration_per_task_ms": 385762.69666666666,
      "avg_output_tokens_per_task": 9391.48,
      "last_run_timestamp": "2026-08-22T16:05:40.004Z"
    },
    {
      "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": 0.698312293670886,
      "accuracy_std_dev": 0.0558284276286228,
      "run_count": 8,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.057643389873417716,
      "avg_duration_per_task_ms": 901330.3670886076,
      "avg_output_tokens_per_task": 61107.00506329114,
      "last_run_timestamp": "2026-09-26T16:00:25.696Z"
    },
    {
      "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": 0.019759375021746,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.020142030201342278,
      "avg_duration_per_task_ms": 437469.0536912752,
      "avg_output_tokens_per_task": 10039.570469798658,
      "last_run_timestamp": "2026-08-29T16:32:40.490Z"
    },
    {
      "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": 0.6966669999999999,
      "accuracy_std_dev": 0.02999999999999997,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12995234,
      "avg_duration_per_task_ms": 694377.26,
      "avg_output_tokens_per_task": 58360.09,
      "last_run_timestamp": "2026-09-01T17:23:14.439Z"
    },
    {
      "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": 0.69666675,
      "accuracy_std_dev": 0.019720195515955205,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.27175437,
      "avg_duration_per_task_ms": 364436.225,
      "avg_output_tokens_per_task": 35416.46,
      "last_run_timestamp": "2026-09-01T13:04:22.481Z"
    },
    {
      "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": 0.6966665000000001,
      "accuracy_std_dev": 0.023333500000000007,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03611107,
      "avg_duration_per_task_ms": 366038.43,
      "avg_output_tokens_per_task": 14231.82,
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "auto-routing",
      "accuracy": 0.6965619147982063,
      "accuracy_std_dev": 0.03306560984534301,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.025591688340807175,
      "avg_duration_per_task_ms": 227729.04260089685,
      "avg_output_tokens_per_task": 9657.32735426009,
      "last_run_timestamp": "2026-08-29T16:13:12.597Z"
    },
    {
      "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": 0.6962962222222222,
      "accuracy_std_dev": 0.07160319018774045,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05209685777777778,
      "avg_duration_per_task_ms": 312714.1022222222,
      "avg_output_tokens_per_task": 8152.073333333334,
      "last_run_timestamp": "2026-08-29T16:33:36.567Z"
    },
    {
      "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": "Together",
      "accuracy": 0.695652,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 46,
      "avg_cost_per_task_usd": 0.32337578260869565,
      "avg_duration_per_task_ms": 302938.26086956525,
      "avg_output_tokens_per_task": 52094.65217391304,
      "last_run_timestamp": "2026-07-04T08:06:25.305Z"
    },
    {
      "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": 0.6945686116071429,
      "accuracy_std_dev": 0.2043857173195383,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.06466310491071428,
      "avg_duration_per_task_ms": 392031.1339285714,
      "avg_output_tokens_per_task": 8293.89732142857,
      "last_run_timestamp": "2026-08-29T17:36:31.188Z"
    },
    {
      "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": 0.694166625,
      "accuracy_std_dev": 0.06141286114475023,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.3036356975,
      "avg_duration_per_task_ms": 394520.48,
      "avg_output_tokens_per_task": 9370.85,
      "last_run_timestamp": "2026-09-26T08:08:19.644Z"
    },
    {
      "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": 0.6936023333333333,
      "accuracy_std_dev": 0.000481077111802244,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.08194677777777779,
      "avg_duration_per_task_ms": 693284.398989899,
      "avg_output_tokens_per_task": 14499.550505050505,
      "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": "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": 0.6933333333333334,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.020424833333333333,
      "avg_duration_per_task_ms": 454611.69333333336,
      "avg_output_tokens_per_task": 10363.686666666666,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Mistral",
      "accuracy": 0.6933331428571429,
      "accuracy_std_dev": 0.02519760633715926,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.11111235000000001,
      "avg_duration_per_task_ms": 77728.44714285714,
      "avg_output_tokens_per_task": 9319.027142857143,
      "last_run_timestamp": "2026-09-26T08:24:41.403Z"
    },
    {
      "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": "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": "DeepInfra",
      "accuracy": 0.6921965086705202,
      "accuracy_std_dev": 0.017490253563464578,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.02071824566473988,
      "avg_duration_per_task_ms": 811870.6213872832,
      "avg_output_tokens_per_task": 19057.85838150289,
      "last_run_timestamp": "2026-08-29T16:48:06.159Z"
    },
    {
      "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": 0.69166675,
      "accuracy_std_dev": 0.015184321229067198,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.043835795,
      "avg_duration_per_task_ms": 455940.16,
      "avg_output_tokens_per_task": 10146.67,
      "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-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Venice",
      "accuracy": 0.6909547336683417,
      "accuracy_std_dev": 0.03444369764353668,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.045872309045226135,
      "avg_duration_per_task_ms": 459922.040201005,
      "avg_output_tokens_per_task": 9154.939698492462,
      "last_run_timestamp": "2026-08-29T16:33:38.223Z"
    },
    {
      "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": 0.6900003333333334,
      "accuracy_std_dev": 0.012472191289246483,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.04195019333333334,
      "avg_duration_per_task_ms": 246231.66666666666,
      "avg_output_tokens_per_task": 31170.286666666667,
      "last_run_timestamp": "2026-09-24T09:00:05.215Z"
    },
    {
      "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": 0.69,
      "accuracy_std_dev": 0.032088096328701056,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.18925239666666666,
      "avg_duration_per_task_ms": 1912773.68,
      "avg_output_tokens_per_task": 13075.366666666667,
      "last_run_timestamp": "2026-08-29T16:50:22.450Z"
    },
    {
      "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": 0.68999975,
      "accuracy_std_dev": 0.03605553586742955,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.113432665,
      "avg_duration_per_task_ms": 675363.205,
      "avg_output_tokens_per_task": 27403.155,
      "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": "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": 0.6873795664739885,
      "accuracy_std_dev": 0.030644933065213705,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.03333745375722543,
      "avg_duration_per_task_ms": 810141.4335260115,
      "avg_output_tokens_per_task": 20074.78901734104,
      "last_run_timestamp": "2026-08-29T17:42:21.163Z"
    },
    {
      "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": 0.686667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.1282767,
      "avg_duration_per_task_ms": 113186.12,
      "avg_output_tokens_per_task": 4988.34,
      "last_run_timestamp": "2026-09-23T15:00:14.019Z"
    },
    {
      "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": 0.6866667,
      "accuracy_std_dev": 0.04862554427057861,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.029370520000000004,
      "avg_duration_per_task_ms": 173382.156,
      "avg_output_tokens_per_task": 10540.91,
      "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-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.6866666666666666,
      "accuracy_std_dev": 0.005443174461857887,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.009258993333333333,
      "avg_duration_per_task_ms": 105180.14666666667,
      "avg_output_tokens_per_task": 8673.2,
      "last_run_timestamp": "2026-09-23T06:00:08.557Z"
    },
    {
      "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": 0.6866663333333333,
      "accuracy_std_dev": 0.024944382578492918,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.013524920000000001,
      "avg_duration_per_task_ms": 73418.06,
      "avg_output_tokens_per_task": 5390.6866666666665,
      "last_run_timestamp": "2026-09-01T17:31:44.289Z"
    },
    {
      "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": 0.6860002,
      "accuracy_std_dev": 0.0747409975258559,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.11085599800000001,
      "avg_duration_per_task_ms": 65942.592,
      "avg_output_tokens_per_task": 8834.634,
      "last_run_timestamp": "2026-09-26T08:22:29.056Z"
    },
    {
      "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": 0.6858332500000001,
      "accuracy_std_dev": 0.009538074291359878,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.11306765,
      "avg_duration_per_task_ms": 138784.995,
      "avg_output_tokens_per_task": 8808.785,
      "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-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6857144285714285,
      "accuracy_std_dev": 0.042077516201333307,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.030182771428571425,
      "avg_duration_per_task_ms": 323904.1228571429,
      "avg_output_tokens_per_task": 10162.908571428572,
      "last_run_timestamp": "2026-08-29T16:50:26.708Z"
    },
    {
      "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": 0.6857143795918367,
      "accuracy_std_dev": 0.056091523474496544,
      "run_count": 10,
      "total_questions": 490,
      "avg_cost_per_task_usd": 0.030057089795918368,
      "avg_duration_per_task_ms": 947954.6224489796,
      "avg_output_tokens_per_task": 40436.48163265306,
      "last_run_timestamp": "2026-09-26T16:00:18.485Z"
    },
    {
      "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": 0.6851852777777777,
      "accuracy_std_dev": 0.03997262094623929,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.035166996666666665,
      "avg_duration_per_task_ms": 166784.71777777778,
      "avg_output_tokens_per_task": 10260.99888888889,
      "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": 0.6850458295739348,
      "accuracy_std_dev": 0.02015963904958062,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.006078426065162907,
      "avg_duration_per_task_ms": 231835.2882205514,
      "avg_output_tokens_per_task": 9389.711779448622,
      "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": "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": 0.685,
      "accuracy_std_dev": 0.025110843165851677,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0074074375,
      "avg_duration_per_task_ms": 254406.3375,
      "avg_output_tokens_per_task": 11400.4,
      "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-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "auto-routing",
      "accuracy": 0.683673530612245,
      "accuracy_std_dev": 0.03777750000000002,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.6821915918367347,
      "avg_duration_per_task_ms": 803506.224489796,
      "avg_output_tokens_per_task": 61161.97959183674,
      "last_run_timestamp": "2026-09-01T12:56:59.816Z"
    },
    {
      "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": 0.683333,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.05961793,
      "avg_duration_per_task_ms": 370273.51,
      "avg_output_tokens_per_task": 10007.47,
      "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": "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.6819729591836734,
      "accuracy_std_dev": 0.005416499999999991,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.019869846938775512,
      "avg_duration_per_task_ms": 223014.25510204083,
      "avg_output_tokens_per_task": 10886.510204081633,
      "last_run_timestamp": "2026-07-11T15:34:15.373Z"
    },
    {
      "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": "Amazon Bedrock",
      "accuracy": 0.68,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.65240112,
      "avg_duration_per_task_ms": 90576.24,
      "avg_output_tokens_per_task": 3843.06,
      "last_run_timestamp": "2026-09-26T09:24:36.382Z"
    },
    {
      "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": 0.68,
      "accuracy_std_dev": 0,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.088178175,
      "avg_duration_per_task_ms": 74028.285,
      "avg_output_tokens_per_task": 3643.675,
      "last_run_timestamp": "2026-09-19T08:08:39.705Z"
    },
    {
      "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": 0.68,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06724904000000001,
      "avg_duration_per_task_ms": 102008.58,
      "avg_output_tokens_per_task": 2510.66,
      "last_run_timestamp": "2026-08-22T08:23:03.574Z"
    },
    {
      "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": 0.6788891666666665,
      "accuracy_std_dev": 0.08738317986587706,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.057002019999999994,
      "avg_duration_per_task_ms": 515207.3233333333,
      "avg_output_tokens_per_task": 18064.933333333334,
      "last_run_timestamp": "2026-08-29T17:31:24.151Z"
    },
    {
      "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": 0.6785186666666666,
      "accuracy_std_dev": 0.019570227381861923,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.00874342888888889,
      "avg_duration_per_task_ms": 244185.2288888889,
      "avg_output_tokens_per_task": 10575.206666666667,
      "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-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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Cloudflare",
      "accuracy": 0.6773809285714286,
      "accuracy_std_dev": 0.1326003459074697,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.12869523000000002,
      "avg_duration_per_task_ms": 133454.41285714286,
      "avg_output_tokens_per_task": 8415.124285714286,
      "last_run_timestamp": "2026-09-26T08:16:17.531Z"
    },
    {
      "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": 0.6770371111111111,
      "accuracy_std_dev": 0.2038422884816399,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.04420765333333334,
      "avg_duration_per_task_ms": 387239.0911111111,
      "avg_output_tokens_per_task": 8342.586666666666,
      "last_run_timestamp": "2026-08-30T00:01:10.585Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.6755553333333334,
      "accuracy_std_dev": 0.00314285394011385,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09428396666666666,
      "avg_duration_per_task_ms": 1235635.52,
      "avg_output_tokens_per_task": 16258.493333333334,
      "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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.6744185058139535,
      "accuracy_std_dev": 0.09615398664493471,
      "run_count": 7,
      "total_questions": 344,
      "avg_cost_per_task_usd": 0.0726462965116279,
      "avg_duration_per_task_ms": 198176.1191860465,
      "avg_output_tokens_per_task": 25038.38081395349,
      "last_run_timestamp": "2026-08-29T16:10:33.075Z"
    },
    {
      "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": 0.6741667499999999,
      "accuracy_std_dev": 0.04671857083845245,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.09568332249999999,
      "avg_duration_per_task_ms": 277611.575,
      "avg_output_tokens_per_task": 10319.5475,
      "last_run_timestamp": "2026-08-29T16:43:32.997Z"
    },
    {
      "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": 0.6740741111111112,
      "accuracy_std_dev": 0.19675003227188467,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.2889698533333333,
      "avg_duration_per_task_ms": 475612.9288888889,
      "avg_output_tokens_per_task": 22340.875555555554,
      "last_run_timestamp": "2026-08-29T16:58:32.934Z"
    },
    {
      "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": 0.6733334999999999,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.14733292,
      "avg_duration_per_task_ms": 130030.77,
      "avg_output_tokens_per_task": 5038.48,
      "last_run_timestamp": "2026-09-23T15:13:09.897Z"
    },
    {
      "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-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.673333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.039595319999999996,
      "avg_duration_per_task_ms": 304692.26,
      "avg_output_tokens_per_task": 31675.6,
      "last_run_timestamp": "2026-09-24T09:00:04.937Z"
    },
    {
      "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": 0.6717170404040403,
      "accuracy_std_dev": 0.038697517759541104,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02239533585858586,
      "avg_duration_per_task_ms": 432351.29545454547,
      "avg_output_tokens_per_task": 20630.38888888889,
      "last_run_timestamp": "2026-08-29T16:12:10.466Z"
    },
    {
      "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": 0.6707315609756098,
      "accuracy_std_dev": 0.021850282996794323,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.016168723577235772,
      "avg_duration_per_task_ms": 272417.73170731706,
      "avg_output_tokens_per_task": 10469.817073170732,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Friendli",
      "accuracy": 0.6704762857142856,
      "accuracy_std_dev": 0.09919401452235525,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08830733714285714,
      "avg_duration_per_task_ms": 86182.73714285715,
      "avg_output_tokens_per_task": 8229.205714285714,
      "last_run_timestamp": "2026-08-29T16:33:39.233Z"
    },
    {
      "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": 0.67,
      "accuracy_std_dev": 0.003332999999999975,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03314721,
      "avg_duration_per_task_ms": 151834.99,
      "avg_output_tokens_per_task": 17524.36,
      "last_run_timestamp": "2026-09-01T17:29:39.925Z"
    },
    {
      "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": 0.67,
      "accuracy_std_dev": 0.031446692107437925,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07852493499999999,
      "avg_duration_per_task_ms": 68196.455,
      "avg_output_tokens_per_task": 3640.76,
      "last_run_timestamp": "2026-09-19T08:12:48.272Z"
    },
    {
      "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": 0.67,
      "accuracy_std_dev": 0.02999999999999997,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.014700359999999999,
      "avg_duration_per_task_ms": 1773933.08,
      "avg_output_tokens_per_task": 23989.78,
      "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": "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": 0.6694631409395974,
      "accuracy_std_dev": 0.06606770680752856,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.024180788590604026,
      "avg_duration_per_task_ms": 455932.432885906,
      "avg_output_tokens_per_task": 17012.89932885906,
      "last_run_timestamp": "2026-08-29T16:44:07.735Z"
    },
    {
      "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": 0.6681159130434783,
      "accuracy_std_dev": 0.032598105721055856,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.12360180173913042,
      "avg_duration_per_task_ms": 117551.10695652173,
      "avg_output_tokens_per_task": 12432.393913043477,
      "last_run_timestamp": "2026-09-26T08:11:26.494Z"
    },
    {
      "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": 0.668,
      "accuracy_std_dev": 0.06598329755021341,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.06998024000000001,
      "avg_duration_per_task_ms": 716919.092,
      "avg_output_tokens_per_task": 18456.574,
      "last_run_timestamp": "2026-09-26T08:15:33.177Z"
    },
    {
      "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": 0.667619,
      "accuracy_std_dev": 0.045694607868825315,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.019704325714285712,
      "avg_duration_per_task_ms": 356821.1142857143,
      "avg_output_tokens_per_task": 19482.165714285715,
      "last_run_timestamp": "2026-08-29T17:51:56.583Z"
    },
    {
      "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": 0.6674998750000001,
      "accuracy_std_dev": 0.034710984293295034,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.11729118250000001,
      "avg_duration_per_task_ms": 121864.08,
      "avg_output_tokens_per_task": 12614.4,
      "last_run_timestamp": "2026-09-26T08:10:43.582Z"
    },
    {
      "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": 0.6666665,
      "accuracy_std_dev": 0.026666499999999982,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.11778575999999999,
      "avg_duration_per_task_ms": 148387.42,
      "avg_output_tokens_per_task": 58241.15,
      "last_run_timestamp": "2026-09-01T17:32:23.926Z"
    },
    {
      "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": 0.6666663333333335,
      "accuracy_std_dev": 0.016996731711975913,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.020487560000000002,
      "avg_duration_per_task_ms": 154828.78666666665,
      "avg_output_tokens_per_task": 15481.913333333334,
      "last_run_timestamp": "2026-08-29T16:48:14.218Z"
    },
    {
      "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "CoreWeave",
      "accuracy": 0.6633333333333333,
      "accuracy_std_dev": 0.016555205069369833,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.01703684,
      "avg_duration_per_task_ms": 144424.58666666667,
      "avg_output_tokens_per_task": 21218.593333333334,
      "last_run_timestamp": "2026-09-26T08:04:53.526Z"
    },
    {
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Phala",
      "accuracy": 0.6614868263254112,
      "accuracy_std_dev": 0.20509144592998726,
      "run_count": 11,
      "total_questions": 547,
      "avg_cost_per_task_usd": 0.09529280438756856,
      "avg_duration_per_task_ms": 265157.61608775135,
      "avg_output_tokens_per_task": 8754.009140767825,
      "last_run_timestamp": "2026-09-26T08:24:45.132Z"
    },
    {
      "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": 0.6608527965116279,
      "accuracy_std_dev": 0.03853886998862316,
      "run_count": 7,
      "total_questions": 344,
      "avg_cost_per_task_usd": 0.024110087209302324,
      "avg_duration_per_task_ms": 240569.14244186046,
      "avg_output_tokens_per_task": 17453.627906976744,
      "last_run_timestamp": "2026-08-29T17:44:03.617Z"
    },
    {
      "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": 0.660833375,
      "accuracy_std_dev": 0.08579283793962278,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0750140775,
      "avg_duration_per_task_ms": 256748.24,
      "avg_output_tokens_per_task": 24550.89,
      "last_run_timestamp": "2026-08-29T17:00:54.109Z"
    },
    {
      "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": 0.66,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00507128,
      "avg_duration_per_task_ms": 72920.3,
      "avg_output_tokens_per_task": 17941.8,
      "last_run_timestamp": "2026-09-28T18:00:05.887Z"
    },
    {
      "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": 0.66,
      "accuracy_std_dev": 0.01911622670810208,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.02949636222222222,
      "avg_duration_per_task_ms": 349016.0333333333,
      "avg_output_tokens_per_task": 10678.31111111111,
      "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": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.66,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00819156,
      "avg_duration_per_task_ms": 91056.4,
      "avg_output_tokens_per_task": 8317.36,
      "last_run_timestamp": "2026-09-23T06:00:08.542Z"
    },
    {
      "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": 0.658750125,
      "accuracy_std_dev": 0.021339771978851493,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.113134535,
      "avg_duration_per_task_ms": 123266.8825,
      "avg_output_tokens_per_task": 12095.085,
      "last_run_timestamp": "2026-09-26T08:08:19.644Z"
    },
    {
      "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": 0.65833325,
      "accuracy_std_dev": 0.020480390351199366,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.022493505000000004,
      "avg_duration_per_task_ms": 261094.5125,
      "avg_output_tokens_per_task": 10075.545,
      "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": "Friendli",
      "accuracy": 0.6570371111111111,
      "accuracy_std_dev": 0.027596789142718296,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.029809600000000006,
      "avg_duration_per_task_ms": 233052.69555555555,
      "avg_output_tokens_per_task": 11101.757777777777,
      "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-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.6570371111111111,
      "accuracy_std_dev": 0.023330540692141164,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.022303715555555555,
      "avg_duration_per_task_ms": 285586.0155555556,
      "avg_output_tokens_per_task": 10395.377777777778,
      "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": 0.6566665,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.008976079999999999,
      "avg_duration_per_task_ms": 368613.48,
      "avg_output_tokens_per_task": 25363.16,
      "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Phala",
      "accuracy": 0.6565315,
      "accuracy_std_dev": 0.011837568509716087,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.014823912162162164,
      "avg_duration_per_task_ms": 271525.86486486485,
      "avg_output_tokens_per_task": 19985.412162162163,
      "last_run_timestamp": "2026-09-19T08:01:36.456Z"
    },
    {
      "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": 0.6555556666666666,
      "accuracy_std_dev": 0.03189478845969805,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.023210153333333334,
      "avg_duration_per_task_ms": 207735.58,
      "avg_output_tokens_per_task": 33693.153333333335,
      "last_run_timestamp": "2026-09-26T08:25:43.443Z"
    },
    {
      "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": 0.655102163265306,
      "accuracy_std_dev": 0.015608943341559031,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.012813330612244898,
      "avg_duration_per_task_ms": 278688.9183673469,
      "avg_output_tokens_per_task": 16201.359183673469,
      "last_run_timestamp": "2026-08-29T16:33:39.228Z"
    },
    {
      "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": 0.6549999166666667,
      "accuracy_std_dev": 0.02926570645242861,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.12179273749999998,
      "avg_duration_per_task_ms": 120187.11666666667,
      "avg_output_tokens_per_task": 12516.864166666666,
      "last_run_timestamp": "2026-09-26T08:13:57.277Z"
    },
    {
      "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": 0.6533335,
      "accuracy_std_dev": 0.053333500000000034,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.22863909999999998,
      "avg_duration_per_task_ms": 717866.38,
      "avg_output_tokens_per_task": 11422.01,
      "last_run_timestamp": "2026-09-26T08:13:25.231Z"
    },
    {
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Novita",
      "accuracy": 0.6525,
      "accuracy_std_dev": 0.054102510725011654,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.021760159999999997,
      "avg_duration_per_task_ms": 206554.95,
      "avg_output_tokens_per_task": 8661.865,
      "last_run_timestamp": "2026-08-29T16:37:21.285Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.6514084084507042,
      "accuracy_std_dev": 0.140951893455415,
      "run_count": 3,
      "total_questions": 142,
      "avg_cost_per_task_usd": 0.04356793661971831,
      "avg_duration_per_task_ms": 110031.35915492958,
      "avg_output_tokens_per_task": 6698.830985915493,
      "last_run_timestamp": "2026-08-01T08:06:01.594Z"
    },
    {
      "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": 0.651111,
      "accuracy_std_dev": 0.09803010371649448,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.010906386666666667,
      "avg_duration_per_task_ms": 303175.14,
      "avg_output_tokens_per_task": 4777.48,
      "last_run_timestamp": "2026-09-26T08:13:57.291Z"
    },
    {
      "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": 0.6493332,
      "accuracy_std_dev": 0.024801501748079685,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.024958387999999998,
      "avg_duration_per_task_ms": 169231.86,
      "avg_output_tokens_per_task": 11298.596,
      "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": "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": "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": 0.6476509932885907,
      "accuracy_std_dev": 0.013206828848070314,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.018146744966442953,
      "avg_duration_per_task_ms": 194939.0268456376,
      "avg_output_tokens_per_task": 22002.395973154362,
      "last_run_timestamp": "2026-09-26T08:13:57.279Z"
    },
    {
      "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": 0.6474073333333333,
      "accuracy_std_dev": 0.028011374332581395,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021467380000000005,
      "avg_duration_per_task_ms": 230369.31333333332,
      "avg_output_tokens_per_task": 10678.271111111111,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Venice",
      "accuracy": 0.6466667777777777,
      "accuracy_std_dev": 0.17752752803424773,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.06551144888888888,
      "avg_duration_per_task_ms": 209379.12444444443,
      "avg_output_tokens_per_task": 9860.064444444444,
      "last_run_timestamp": "2026-08-29T17:03:33.193Z"
    },
    {
      "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": 0.6461902857142857,
      "accuracy_std_dev": 0.024196977005487303,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.022573365714285715,
      "avg_duration_per_task_ms": 279162.21142857143,
      "avg_output_tokens_per_task": 11143.031428571428,
      "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": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Azure",
      "accuracy": 0.645833375,
      "accuracy_std_dev": 0.02526261098905604,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.11683335,
      "avg_duration_per_task_ms": 147559.335,
      "avg_output_tokens_per_task": 12607.1725,
      "last_run_timestamp": "2026-09-26T08:22:25.227Z"
    },
    {
      "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": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.6433335,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.034682500000000005,
      "avg_duration_per_task_ms": 159195.12,
      "avg_output_tokens_per_task": 18151.16,
      "last_run_timestamp": "2026-09-01T17:31:43.288Z"
    },
    {
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Novita",
      "accuracy": 0.6429628888888889,
      "accuracy_std_dev": 0.020872536285455455,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.024601560000000005,
      "avg_duration_per_task_ms": 160050.17111111112,
      "avg_output_tokens_per_task": 10753.557777777778,
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Alibaba",
      "accuracy": 0.6418009435483871,
      "accuracy_std_dev": 0.15603995157958747,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.04030671370967741,
      "avg_duration_per_task_ms": 276664.12096774194,
      "avg_output_tokens_per_task": 11942.516129032258,
      "last_run_timestamp": "2026-08-29T16:15:12.662Z"
    },
    {
      "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": 0.64166675,
      "accuracy_std_dev": 0.00985994985725082,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.0329834,
      "avg_duration_per_task_ms": 238262.615,
      "avg_output_tokens_per_task": 16910.1,
      "last_run_timestamp": "2026-09-01T17:31:47.440Z"
    },
    {
      "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": 0.6415410502512563,
      "accuracy_std_dev": 0.10999376007351223,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.09234448994974874,
      "avg_duration_per_task_ms": 238857.3718592965,
      "avg_output_tokens_per_task": 26390.881909547737,
      "last_run_timestamp": "2026-08-29T16:12:10.428Z"
    },
    {
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "WandB",
      "accuracy": 0.64,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.08159856,
      "avg_duration_per_task_ms": 189065.52,
      "avg_output_tokens_per_task": 12650.4,
      "last_run_timestamp": "2026-07-04T16:00:23.966Z"
    },
    {
      "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": "Phala",
      "accuracy": 0.64,
      "accuracy_std_dev": 0.026666999999999996,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.036191219999999996,
      "avg_duration_per_task_ms": 572703.13,
      "avg_output_tokens_per_task": 26589.6,
      "last_run_timestamp": "2026-07-11T15:34:14.992Z"
    },
    {
      "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": 0.64,
      "accuracy_std_dev": 0.025647213408087847,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.218415728,
      "avg_duration_per_task_ms": 1219489.072,
      "avg_output_tokens_per_task": 31443.888,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Venice",
      "accuracy": 0.6397003505617977,
      "accuracy_std_dev": 0.15332848831846185,
      "run_count": 9,
      "total_questions": 445,
      "avg_cost_per_task_usd": 0.01602385617977528,
      "avg_duration_per_task_ms": 268837.0134831461,
      "avg_output_tokens_per_task": 8576.319101123596,
      "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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Novita",
      "accuracy": 0.6386664000000001,
      "accuracy_std_dev": 0.04139773692172073,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.02512048,
      "avg_duration_per_task_ms": 174653.388,
      "avg_output_tokens_per_task": 11467.556,
      "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": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6360152471264368,
      "accuracy_std_dev": 0.13554242758288354,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.04763444827586207,
      "avg_duration_per_task_ms": 2222096.040229885,
      "avg_output_tokens_per_task": 71368.38505747127,
      "last_run_timestamp": "2026-09-26T08:08:19.641Z"
    },
    {
      "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": 0.6356786532663317,
      "accuracy_std_dev": 0.023893543066067047,
      "run_count": 4,
      "total_questions": 199,
      "avg_cost_per_task_usd": 0.018825286432160804,
      "avg_duration_per_task_ms": 163374.6633165829,
      "avg_output_tokens_per_task": 9394.713567839195,
      "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/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "NextBit",
      "accuracy": 0.6333333333333333,
      "accuracy_std_dev": 0.014401903168987377,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.07877431333333335,
      "avg_duration_per_task_ms": 1233888.1733333333,
      "avg_output_tokens_per_task": 25898.94,
      "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": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "Azure",
      "accuracy": 0.6333332500000001,
      "accuracy_std_dev": 0.03299835012826401,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.102073965,
      "avg_duration_per_task_ms": 160807.645,
      "avg_output_tokens_per_task": 12456.285,
      "last_run_timestamp": "2026-09-01T17:31:46.635Z"
    },
    {
      "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": 0.6331093959731544,
      "accuracy_std_dev": 0.01453205911080741,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.012315053691275168,
      "avg_duration_per_task_ms": 1064929.322147651,
      "avg_output_tokens_per_task": 21067.724832214764,
      "last_run_timestamp": "2026-08-29T17:18:10.895Z"
    },
    {
      "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": 0.6319998,
      "accuracy_std_dev": 0.0363685871784979,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.09467958400000001,
      "avg_duration_per_task_ms": 55836.916,
      "avg_output_tokens_per_task": 3550.436,
      "last_run_timestamp": "2026-08-29T16:15:31.797Z"
    },
    {
      "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": 0.6316667499999999,
      "accuracy_std_dev": 0.030867157124320685,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.021079455,
      "avg_duration_per_task_ms": 289742.9075,
      "avg_output_tokens_per_task": 22034.425,
      "last_run_timestamp": "2026-08-29T16:25:26.695Z"
    },
    {
      "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": 0.6308725570469799,
      "accuracy_std_dev": 0.09289222799567248,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.06755993288590603,
      "avg_duration_per_task_ms": 587099.4328859061,
      "avg_output_tokens_per_task": 38157.05369127517,
      "last_run_timestamp": "2026-08-29T16:07:02.358Z"
    },
    {
      "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": 0.6303702222222223,
      "accuracy_std_dev": 0.02063465195989924,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.029002144444444447,
      "avg_duration_per_task_ms": 221656.39333333334,
      "avg_output_tokens_per_task": 10431.462222222222,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "OpenInference",
      "accuracy": 0.63,
      "accuracy_std_dev": 0.043332999999999955,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.08937484999999999,
      "avg_duration_per_task_ms": 3066082.31,
      "avg_output_tokens_per_task": 80584.37,
      "last_run_timestamp": "2026-09-29T01:28:53.566Z"
    },
    {
      "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": "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": 0.63,
      "accuracy_std_dev": 0.050000000000000044,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.09203941,
      "avg_duration_per_task_ms": 133131.18,
      "avg_output_tokens_per_task": 12511.49,
      "last_run_timestamp": "2026-09-01T17:23:15.998Z"
    },
    {
      "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": 0.6295238571428571,
      "accuracy_std_dev": 0.055874856932839716,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.0383867,
      "avg_duration_per_task_ms": 389225.1657142857,
      "avg_output_tokens_per_task": 20771.2,
      "last_run_timestamp": "2026-08-29T16:50:48.984Z"
    },
    {
      "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": 0.62946425,
      "accuracy_std_dev": 0.027834597690954414,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.03242941071428571,
      "avg_duration_per_task_ms": 84991.91294642857,
      "avg_output_tokens_per_task": 21655.93526785714,
      "last_run_timestamp": "2026-08-29T17:17:57.467Z"
    },
    {
      "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": 0.625850418367347,
      "accuracy_std_dev": 0.1514466649820573,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.0701549693877551,
      "avg_duration_per_task_ms": 396505.55102040817,
      "avg_output_tokens_per_task": 11691.816326530612,
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Google",
      "accuracy": 0.622650703264095,
      "accuracy_std_dev": 0.0769236199560503,
      "run_count": 7,
      "total_questions": 337,
      "avg_cost_per_task_usd": 0.14786436795252222,
      "avg_duration_per_task_ms": 364083.8011869436,
      "avg_output_tokens_per_task": 11314.480712166172,
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Venice",
      "accuracy": 0.6224298747663553,
      "accuracy_std_dev": 0.06407885487325574,
      "run_count": 11,
      "total_questions": 535,
      "avg_cost_per_task_usd": 0.022028269158878504,
      "avg_duration_per_task_ms": 247156.75327102805,
      "avg_output_tokens_per_task": 24158.08785046729,
      "last_run_timestamp": "2026-09-26T16:00:24.609Z"
    },
    {
      "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": 0.6222782361809045,
      "accuracy_std_dev": 0.013777638925424605,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.10842231658291457,
      "avg_duration_per_task_ms": 712075.7864321608,
      "avg_output_tokens_per_task": 34125.979899497484,
      "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": "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": 0.6211111666666667,
      "accuracy_std_dev": 0.013005359772246037,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.10838059000000001,
      "avg_duration_per_task_ms": 725500.3766666667,
      "avg_output_tokens_per_task": 33770.87,
      "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": "google/gemini-2.5-pro",
      "model_slug": "google/gemini-2.5-pro",
      "model_name": "Google: Gemini 2.5 Pro",
      "provider_name": "Google",
      "accuracy": 0.620246288590604,
      "accuracy_std_dev": 0.024043751131473293,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.19689856375838927,
      "avg_duration_per_task_ms": 175265.28187919463,
      "avg_output_tokens_per_task": 13638.724832214764,
      "last_run_timestamp": "2026-09-01T13:00:41.540Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Venice",
      "accuracy": 0.6176750230547552,
      "accuracy_std_dev": 0.22454644005875757,
      "run_count": 7,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.04794541498559077,
      "avg_duration_per_task_ms": 857866.4149855907,
      "avg_output_tokens_per_task": 11712.968299711816,
      "last_run_timestamp": "2026-08-29T17:09:01.494Z"
    },
    {
      "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": 0.6173332,
      "accuracy_std_dev": 0.03309246574312648,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.09146958399999999,
      "avg_duration_per_task_ms": 147054.108,
      "avg_output_tokens_per_task": 3898.968,
      "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": "DeepInfra",
      "accuracy": 0.6163948885869566,
      "accuracy_std_dev": 0.028746044099319138,
      "run_count": 15,
      "total_questions": 736,
      "avg_cost_per_task_usd": 0.019574407608695648,
      "avg_duration_per_task_ms": 331626.9605978261,
      "avg_output_tokens_per_task": 20514.514945652172,
      "last_run_timestamp": "2026-08-29T16:49:23.763Z"
    },
    {
      "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": 0.6159998000000001,
      "accuracy_std_dev": 0.03440923906976152,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.045486667999999994,
      "avg_duration_per_task_ms": 195084.972,
      "avg_output_tokens_per_task": 3358.316,
      "last_run_timestamp": "2026-09-19T08:08:42.956Z"
    },
    {
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.61401515,
      "accuracy_std_dev": 0.08680383555799967,
      "run_count": 9,
      "total_questions": 440,
      "avg_cost_per_task_usd": 0.021259502272727274,
      "avg_duration_per_task_ms": 807630.4068181819,
      "avg_output_tokens_per_task": 27761.55,
      "last_run_timestamp": "2026-09-26T16:00:20.168Z"
    },
    {
      "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": 0.6133335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.22014136,
      "avg_duration_per_task_ms": 192160.83,
      "avg_output_tokens_per_task": 15089.79,
      "last_run_timestamp": "2026-09-01T12:57:05.201Z"
    },
    {
      "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": 0.60958325,
      "accuracy_std_dev": 0.03034222282871678,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.06603929000000001,
      "avg_duration_per_task_ms": 26897.2,
      "avg_output_tokens_per_task": 20840.3025,
      "last_run_timestamp": "2026-08-29T16:55:05.300Z"
    },
    {
      "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": 0.608333,
      "accuracy_std_dev": 0.0050000000000000044,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03168119,
      "avg_duration_per_task_ms": 488065.11,
      "avg_output_tokens_per_task": 10898.63,
      "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/qwen3.5-35b-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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "GMICloud",
      "accuracy": 0.6066666666666667,
      "accuracy_std_dev": 0.15241273398753646,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.017056396666666668,
      "avg_duration_per_task_ms": 233868.72666666665,
      "avg_output_tokens_per_task": 8308.71,
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Parasail",
      "accuracy": 0.6045386517857143,
      "accuracy_std_dev": 0.02551152635115024,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.019703399553571425,
      "avg_duration_per_task_ms": 94813.06026785714,
      "avg_output_tokens_per_task": 4259.450892857143,
      "last_run_timestamp": "2026-08-29T16:08:08.875Z"
    },
    {
      "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": 0.604,
      "accuracy_std_dev": 0.19783738552457672,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.131018624,
      "avg_duration_per_task_ms": 356662.58,
      "avg_output_tokens_per_task": 7644.116,
      "last_run_timestamp": "2026-09-26T08:24:48.090Z"
    },
    {
      "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": 0.6033335000000001,
      "accuracy_std_dev": 0.028867224784346696,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.04717542499999999,
      "avg_duration_per_task_ms": 204909.125,
      "avg_output_tokens_per_task": 3753.575,
      "last_run_timestamp": "2026-09-19T08:01:34.089Z"
    },
    {
      "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": 0.6023054178674352,
      "accuracy_std_dev": 0.20058517615217727,
      "run_count": 7,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.010320801152737752,
      "avg_duration_per_task_ms": 214339.75504322766,
      "avg_output_tokens_per_task": 8030.982708933718,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "SambaNova",
      "accuracy": 0.6021870316742081,
      "accuracy_std_dev": 0.025520253734854576,
      "run_count": 9,
      "total_questions": 442,
      "avg_cost_per_task_usd": 0.04226960180995475,
      "avg_duration_per_task_ms": 82823.46606334842,
      "avg_output_tokens_per_task": 26213.19230769231,
      "last_run_timestamp": "2026-08-29T17:17:47.080Z"
    },
    {
      "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": 0.6004564109589041,
      "accuracy_std_dev": 0.09338877055873593,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.485756647260274,
      "avg_duration_per_task_ms": 2654805.914383562,
      "avg_output_tokens_per_task": 83086.6404109589,
      "last_run_timestamp": "2026-09-26T08:17:02.590Z"
    },
    {
      "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": 0.5993149931506849,
      "accuracy_std_dev": 0.055256651291194496,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.0717366712328767,
      "avg_duration_per_task_ms": 725826.4726027397,
      "avg_output_tokens_per_task": 48647.558219178085,
      "last_run_timestamp": "2026-08-29T17:09:45.162Z"
    },
    {
      "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": 0.5935375204081632,
      "accuracy_std_dev": 0.22137617923387215,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.04782920408163266,
      "avg_duration_per_task_ms": 235539.23979591837,
      "avg_output_tokens_per_task": 7215.7295918367345,
      "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": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.5933335000000001,
      "accuracy_std_dev": 0.05333349999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.08864457999999999,
      "avg_duration_per_task_ms": 125444.95,
      "avg_output_tokens_per_task": 11722.21,
      "last_run_timestamp": "2026-09-01T17:31:29.315Z"
    },
    {
      "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": "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": 0.5909962011494253,
      "accuracy_std_dev": 0.028655941146289218,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.16123717816091956,
      "avg_duration_per_task_ms": 636124.5057471264,
      "avg_output_tokens_per_task": 30137.091954022988,
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "auto-routing",
      "accuracy": 0.5887097016129033,
      "accuracy_std_dev": 0.03627146285111756,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.0998679556451613,
      "avg_duration_per_task_ms": 111335.03629032258,
      "avg_output_tokens_per_task": 3689.3790322580644,
      "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": "Google",
      "accuracy": 0.5866666666666667,
      "accuracy_std_dev": 0.0066666666750000055,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.029609283333333326,
      "avg_duration_per_task_ms": 75747.02,
      "avg_output_tokens_per_task": 8199.643333333333,
      "last_run_timestamp": "2026-09-01T17:32:36.461Z"
    },
    {
      "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": 0.5857141428571428,
      "accuracy_std_dev": 0.041622460359860544,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.011210711428571428,
      "avg_duration_per_task_ms": 363659.66,
      "avg_output_tokens_per_task": 19214.505714285715,
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "StreamLake",
      "accuracy": 0.5855555555555556,
      "accuracy_std_dev": 0.04898963357262684,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.04505041111111111,
      "avg_duration_per_task_ms": 84460.94888888889,
      "avg_output_tokens_per_task": 4504.664444444445,
      "last_run_timestamp": "2026-08-29T16:44:17.204Z"
    },
    {
      "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": 0.5811856778350516,
      "accuracy_std_dev": 0.03980730883718937,
      "run_count": 8,
      "total_questions": 388,
      "avg_cost_per_task_usd": 0.07594913402061856,
      "avg_duration_per_task_ms": 92087.82731958762,
      "avg_output_tokens_per_task": 5709.461340206186,
      "last_run_timestamp": "2026-08-29T16:11:10.141Z"
    },
    {
      "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": 0.5803722081218273,
      "accuracy_std_dev": 0.06797104807149512,
      "run_count": 8,
      "total_questions": 394,
      "avg_cost_per_task_usd": 0.05299261167512691,
      "avg_duration_per_task_ms": 433958.16751269036,
      "avg_output_tokens_per_task": 10198.129441624365,
      "last_run_timestamp": "2026-08-30T00:01:02.157Z"
    },
    {
      "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": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.5766665,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.028599199999999998,
      "avg_duration_per_task_ms": 78841.33,
      "avg_output_tokens_per_task": 8172.09,
      "last_run_timestamp": "2026-09-01T17:31:45.195Z"
    },
    {
      "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": 0.573333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.012256119999999999,
      "avg_duration_per_task_ms": 120534.1,
      "avg_output_tokens_per_task": 16750,
      "last_run_timestamp": "2026-08-07T00:00:02.467Z"
    },
    {
      "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": 0.5726744011627907,
      "accuracy_std_dev": 0.020765505569357568,
      "run_count": 7,
      "total_questions": 344,
      "avg_cost_per_task_usd": 0.13094280523255813,
      "avg_duration_per_task_ms": 1202301.4127906978,
      "avg_output_tokens_per_task": 25694.97965116279,
      "last_run_timestamp": "2026-08-29T16:31:42.155Z"
    },
    {
      "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": 0.569727612244898,
      "accuracy_std_dev": 0.026944500000000038,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.0024336428571428573,
      "avg_duration_per_task_ms": 317617.1530612245,
      "avg_output_tokens_per_task": 5820.755102040816,
      "last_run_timestamp": "2026-08-29T16:57:41.635Z"
    },
    {
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Mancer 2",
      "accuracy": 0.5639731717171717,
      "accuracy_std_dev": 0.09459911998137191,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.04424948484848485,
      "avg_duration_per_task_ms": 3137109.6868686867,
      "avg_output_tokens_per_task": 95124.10606060606,
      "last_run_timestamp": "2026-08-29T17:24:25.648Z"
    },
    {
      "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": 0.5604491347150259,
      "accuracy_std_dev": 0.10678459157041047,
      "run_count": 8,
      "total_questions": 386,
      "avg_cost_per_task_usd": 0.07120025388601035,
      "avg_duration_per_task_ms": 510444.22279792745,
      "avg_output_tokens_per_task": 29300.067357512955,
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "auto-routing",
      "accuracy": 0.5562962222222222,
      "accuracy_std_dev": 0.09823697755233152,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.04116750222222223,
      "avg_duration_per_task_ms": 96293.60222222222,
      "avg_output_tokens_per_task": 4303.308888888889,
      "last_run_timestamp": "2026-08-29T16:43:55.088Z"
    },
    {
      "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": 0.5533335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03619409,
      "avg_duration_per_task_ms": 60452.63,
      "avg_output_tokens_per_task": 6631.82,
      "last_run_timestamp": "2026-09-01T17:31:47.754Z"
    },
    {
      "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": 0.55,
      "accuracy_std_dev": 0.036667000000000005,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.23514412,
      "avg_duration_per_task_ms": 166142.51,
      "avg_output_tokens_per_task": 13675.21,
      "last_run_timestamp": "2026-09-01T13:01:57.044Z"
    },
    {
      "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": "BaseTen",
      "accuracy": 0.5382882432432433,
      "accuracy_std_dev": 0.2423313405275696,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.02003441891891892,
      "avg_duration_per_task_ms": 116748.24662162163,
      "avg_output_tokens_per_task": 20889.20608108108,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Nebius",
      "accuracy": 0.5355556666666667,
      "accuracy_std_dev": 0.02199873172606905,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09782582666666667,
      "avg_duration_per_task_ms": 375218.6066666667,
      "avg_output_tokens_per_task": 16827.4,
      "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-4.1-2025-04-14",
      "model_slug": "openai/gpt-4.1",
      "model_name": "OpenAI: GPT-4.1",
      "provider_name": "Azure",
      "accuracy": 0.5342466164383561,
      "accuracy_std_dev": 0.014081561995113387,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.1389923630136986,
      "avg_duration_per_task_ms": 68151.21917808219,
      "avg_output_tokens_per_task": 2403.2671232876714,
      "last_run_timestamp": "2026-09-01T13:01:05.433Z"
    },
    {
      "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": 0.529239730994152,
      "accuracy_std_dev": 0.016446009428504185,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.02731367543859649,
      "avg_duration_per_task_ms": 1767438.9385964912,
      "avg_output_tokens_per_task": 142501.04678362573,
      "last_run_timestamp": "2026-08-29T16:33:29.988Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Novita",
      "accuracy": 0.5234740915492958,
      "accuracy_std_dev": 0.24433860528145954,
      "run_count": 3,
      "total_questions": 142,
      "avg_cost_per_task_usd": 0.12058070422535211,
      "avg_duration_per_task_ms": 209796.21126760563,
      "avg_output_tokens_per_task": 11881.387323943662,
      "last_run_timestamp": "2026-07-18T16:00:09.881Z"
    },
    {
      "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": 0.5228572857142858,
      "accuracy_std_dev": 0.021925379532237367,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.028223582857142857,
      "avg_duration_per_task_ms": 445218.9428571428,
      "avg_output_tokens_per_task": 57366.497142857144,
      "last_run_timestamp": "2026-09-26T16:00:18.487Z"
    },
    {
      "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": 0.5183044971098266,
      "accuracy_std_dev": 0.03913581029998911,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.1411618005780347,
      "avg_duration_per_task_ms": 1537742.4248554914,
      "avg_output_tokens_per_task": 30943.121387283238,
      "last_run_timestamp": "2026-08-29T16:25:50.849Z"
    },
    {
      "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": 0.5151512828282828,
      "accuracy_std_dev": 0.04224499999999998,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.026153939393939396,
      "avg_duration_per_task_ms": 213641.58585858587,
      "avg_output_tokens_per_task": 54161.38383838384,
      "last_run_timestamp": "2026-09-01T17:31:30.190Z"
    },
    {
      "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": 0.5119049598214286,
      "accuracy_std_dev": 0.02807387800381059,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.06541727232142856,
      "avg_duration_per_task_ms": 304953.5691964286,
      "avg_output_tokens_per_task": 4005.4308035714284,
      "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": "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-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DeepInfra",
      "accuracy": 0.5077973684210526,
      "accuracy_std_dev": 0.01607029428505164,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.021449652046783625,
      "avg_duration_per_task_ms": 1262349.283625731,
      "avg_output_tokens_per_task": 119951.4649122807,
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.5072072378378378,
      "accuracy_std_dev": 0.016622828858597505,
      "run_count": 15,
      "total_questions": 740,
      "avg_cost_per_task_usd": 0.04104409594594594,
      "avg_duration_per_task_ms": 771120.0337837838,
      "avg_output_tokens_per_task": 212846.60270270272,
      "last_run_timestamp": "2026-08-29T16:44:16.523Z"
    },
    {
      "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": 0.5044443333333334,
      "accuracy_std_dev": 0.01571379829604824,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.022218646666666668,
      "avg_duration_per_task_ms": 1815315.58,
      "avg_output_tokens_per_task": 146379.76666666666,
      "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": "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.5028571428571429,
      "accuracy_std_dev": 0.010605205587400014,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.01953936,
      "avg_duration_per_task_ms": 356098.7342857143,
      "avg_output_tokens_per_task": 10947.26,
      "last_run_timestamp": "2026-08-29T16:44:22.818Z"
    },
    {
      "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": 0.49666650000000007,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.13482986,
      "avg_duration_per_task_ms": 58783.54,
      "avg_output_tokens_per_task": 2896.5,
      "last_run_timestamp": "2026-09-01T13:00:51.538Z"
    },
    {
      "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "SiliconFlow",
      "accuracy": 0.49523814285714285,
      "accuracy_std_dev": 0.023289352137395136,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.06688620571428572,
      "avg_duration_per_task_ms": 308997.38285714283,
      "avg_output_tokens_per_task": 4074.4114285714286,
      "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": "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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.49319713265306125,
      "accuracy_std_dev": 0.03082350135347216,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.05686764285714286,
      "avg_duration_per_task_ms": 2675498.8775510206,
      "avg_output_tokens_per_task": 42954.67857142857,
      "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": "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": 0.4926666,
      "accuracy_std_dev": 0.017688487155209202,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.033023068,
      "avg_duration_per_task_ms": 659102.76,
      "avg_output_tokens_per_task": 61671.012,
      "last_run_timestamp": "2026-09-26T16:00:20.149Z"
    },
    {
      "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": 0.4915826313131313,
      "accuracy_std_dev": 0.024392630759296454,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.023856613636363638,
      "avg_duration_per_task_ms": 839892.5429292929,
      "avg_output_tokens_per_task": 22724.651515151516,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Together",
      "accuracy": 0.49028390582959647,
      "accuracy_std_dev": 0.024789823060240374,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.02451580269058296,
      "avg_duration_per_task_ms": 275409.5044843049,
      "avg_output_tokens_per_task": 18530.174887892375,
      "last_run_timestamp": "2026-08-29T16:33:30.990Z"
    },
    {
      "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": 0.48857142857142855,
      "accuracy_std_dev": 0.018418205526498126,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.034540077142857145,
      "avg_duration_per_task_ms": 44715.205714285716,
      "avg_output_tokens_per_task": 3650.385714285714,
      "last_run_timestamp": "2026-08-29T16:42:36.232Z"
    },
    {
      "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": "Chutes",
      "accuracy": 0.48666699999999996,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.058167340000000005,
      "avg_duration_per_task_ms": 169001.24,
      "avg_output_tokens_per_task": 5056.06,
      "last_run_timestamp": "2026-07-04T16:00:23.500Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.48666650000000006,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12261337999999998,
      "avg_duration_per_task_ms": 50813.54,
      "avg_output_tokens_per_task": 2655.2,
      "last_run_timestamp": "2026-09-01T12:53:49.436Z"
    },
    {
      "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": 0.48333325000000005,
      "accuracy_std_dev": 0.025166081671319035,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.021045685,
      "avg_duration_per_task_ms": 161568.115,
      "avg_output_tokens_per_task": 3811.995,
      "last_run_timestamp": "2026-08-29T16:46:44.671Z"
    },
    {
      "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": 0.4829628888888889,
      "accuracy_std_dev": 0.029666526102455164,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.02247696444444445,
      "avg_duration_per_task_ms": 720948.6155555556,
      "avg_output_tokens_per_task": 22403.7,
      "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": "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": 0.48243642152466365,
      "accuracy_std_dev": 0.023113682722310937,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.028583977578475337,
      "avg_duration_per_task_ms": 385123.331838565,
      "avg_output_tokens_per_task": 66962.30269058296,
      "last_run_timestamp": "2026-09-26T16:00:20.148Z"
    },
    {
      "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": 0.4822222222222222,
      "accuracy_std_dev": 0.04036851236016555,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05617918666666666,
      "avg_duration_per_task_ms": 128221.7311111111,
      "avg_output_tokens_per_task": 3794.2533333333336,
      "last_run_timestamp": "2026-08-29T16:33:38.229Z"
    },
    {
      "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": 0.4805680986547085,
      "accuracy_std_dev": 0.032856388856151955,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.023212632286995515,
      "avg_duration_per_task_ms": 56208.64349775785,
      "avg_output_tokens_per_task": 3761.609865470852,
      "last_run_timestamp": "2026-08-29T16:12:10.468Z"
    },
    {
      "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-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Parasail",
      "accuracy": 0.47716860273972606,
      "accuracy_std_dev": 0.06078806821700747,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.026057061643835613,
      "avg_duration_per_task_ms": 2288262.0684931506,
      "avg_output_tokens_per_task": 135316.74657534246,
      "last_run_timestamp": "2026-08-29T17:10:29.433Z"
    },
    {
      "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": 0.47701135057471267,
      "accuracy_std_dev": 0.009988678401274195,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.1616835948275862,
      "avg_duration_per_task_ms": 2419122.9655172415,
      "avg_output_tokens_per_task": 43666.33620689655,
      "last_run_timestamp": "2026-08-29T16:20:58.736Z"
    },
    {
      "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": 0.47619069387755103,
      "accuracy_std_dev": 0.04472200000000001,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.042756918367346936,
      "avg_duration_per_task_ms": 697970.0612244898,
      "avg_output_tokens_per_task": 99743.38775510204,
      "last_run_timestamp": "2026-09-01T17:27:02.043Z"
    },
    {
      "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": 0.4737037777777778,
      "accuracy_std_dev": 0.04287349816683645,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.0567969311111111,
      "avg_duration_per_task_ms": 70294.79555555555,
      "avg_output_tokens_per_task": 3592.2155555555555,
      "last_run_timestamp": "2026-08-29T17:01:46.593Z"
    },
    {
      "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": 0.02492969,
      "avg_duration_per_task_ms": 126362.985,
      "avg_output_tokens_per_task": 3798.445,
      "last_run_timestamp": "2026-08-29T16:22:50.200Z"
    },
    {
      "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": 0.473333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.023663,
      "avg_duration_per_task_ms": 175642.6,
      "avg_output_tokens_per_task": 45106.7,
      "last_run_timestamp": "2026-08-01T17:39:14.289Z"
    },
    {
      "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": "Phala",
      "accuracy": 0.473333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.023599600000000002,
      "avg_duration_per_task_ms": 409207.06,
      "avg_output_tokens_per_task": 7009.96,
      "last_run_timestamp": "2026-07-04T16:00:25.388Z"
    },
    {
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Together",
      "accuracy": 0.473333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05687784,
      "avg_duration_per_task_ms": 126514.22,
      "avg_output_tokens_per_task": 3833.8,
      "last_run_timestamp": "2026-07-04T16:00:25.259Z"
    },
    {
      "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": 0.469999875,
      "accuracy_std_dev": 0.031798003956370814,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.113410725,
      "avg_duration_per_task_ms": 103888.355,
      "avg_output_tokens_per_task": 2755.44,
      "last_run_timestamp": "2026-08-29T16:33:30.427Z"
    },
    {
      "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": 0.019040825714285718,
      "avg_duration_per_task_ms": 89246.73142857142,
      "avg_output_tokens_per_task": 3672.6485714285714,
      "last_run_timestamp": "2026-08-29T16:43:11.444Z"
    },
    {
      "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": 0.46666633333333335,
      "accuracy_std_dev": 0.009428090415820642,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03296042,
      "avg_duration_per_task_ms": 474531.4533333333,
      "avg_output_tokens_per_task": 36019.21333333333,
      "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Novita",
      "accuracy": 0.46583349999999996,
      "accuracy_std_dev": 0.051363408397905215,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.127061305,
      "avg_duration_per_task_ms": 112391.425,
      "avg_output_tokens_per_task": 3072.22,
      "last_run_timestamp": "2026-08-29T16:25:26.800Z"
    },
    {
      "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": 0.4652775,
      "accuracy_std_dev": 0.006944499999999992,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.027522666666666668,
      "avg_duration_per_task_ms": 29612.604166666668,
      "avg_output_tokens_per_task": 1399.6145833333333,
      "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": "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": "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": 0.464166625,
      "accuracy_std_dev": 0.03843292588047877,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.05392122499999999,
      "avg_duration_per_task_ms": 149701.2725,
      "avg_output_tokens_per_task": 3823.17,
      "last_run_timestamp": "2026-08-29T17:03:47.822Z"
    },
    {
      "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": 0.46333349999999995,
      "accuracy_std_dev": 0.0166665,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.20313888,
      "avg_duration_per_task_ms": 52334.34,
      "avg_output_tokens_per_task": 2078.1,
      "last_run_timestamp": "2026-09-01T12:41:31.668Z"
    },
    {
      "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": "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": 0.46279764285714287,
      "accuracy_std_dev": 0.03220534007479711,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.022562319196428572,
      "avg_duration_per_task_ms": 190962.13169642858,
      "avg_output_tokens_per_task": 3688.5491071428573,
      "last_run_timestamp": "2026-08-29T16:22:50.200Z"
    },
    {
      "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": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": null,
      "avg_duration_per_task_ms": 156315.84,
      "avg_output_tokens_per_task": 4524,
      "last_run_timestamp": "2026-09-01T12:59:40.525Z"
    },
    {
      "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/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "DeepInfra",
      "accuracy": 0.46,
      "accuracy_std_dev": 0.013333000000000012,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02811872,
      "avg_duration_per_task_ms": 579611.05,
      "avg_output_tokens_per_task": 20031.27,
      "last_run_timestamp": "2026-09-01T13:02:37.014Z"
    },
    {
      "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": 0.45920115624999996,
      "accuracy_std_dev": 0.020597873615982785,
      "run_count": 4,
      "total_questions": 192,
      "avg_cost_per_task_usd": 0.08208188020833333,
      "avg_duration_per_task_ms": 47993.442708333336,
      "avg_output_tokens_per_task": 3348.3072916666665,
      "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": "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Nebius",
      "accuracy": 0.4576988578680203,
      "accuracy_std_dev": 0.05672366902457362,
      "run_count": 8,
      "total_questions": 394,
      "avg_cost_per_task_usd": 0.16923886294416243,
      "avg_duration_per_task_ms": 401189.9695431472,
      "avg_output_tokens_per_task": 63938.149746192896,
      "last_run_timestamp": "2026-08-29T17:20:10.689Z"
    },
    {
      "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": 0.45542626162790695,
      "accuracy_std_dev": 0.018912652487751393,
      "run_count": 7,
      "total_questions": 344,
      "avg_cost_per_task_usd": 0.046210372093023255,
      "avg_duration_per_task_ms": 174189.27034883722,
      "avg_output_tokens_per_task": 1898.7906976744187,
      "last_run_timestamp": "2026-08-29T16:23:01.234Z"
    },
    {
      "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": 0.45333349999999994,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.2293187,
      "avg_duration_per_task_ms": 78949.14,
      "avg_output_tokens_per_task": 2726.37,
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.4533333333333333,
      "accuracy_std_dev": 0.02553153460592084,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.027975164444444443,
      "avg_duration_per_task_ms": 208757.2288888889,
      "avg_output_tokens_per_task": 3694.0333333333333,
      "last_run_timestamp": "2026-08-29T16:33:39.618Z"
    },
    {
      "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": 0.44891123618090456,
      "accuracy_std_dev": 0.023521463575517047,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.020988829145728645,
      "avg_duration_per_task_ms": 492653.43467336684,
      "avg_output_tokens_per_task": 21696.05025125628,
      "last_run_timestamp": "2026-08-22T16:08:38.678Z"
    },
    {
      "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": "Google",
      "accuracy": 0.4488886666666667,
      "accuracy_std_dev": 0.05137597555494419,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0494734,
      "avg_duration_per_task_ms": 36454.72,
      "avg_output_tokens_per_task": 2419.86,
      "last_run_timestamp": "2026-08-29T16:58:32.929Z"
    },
    {
      "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": 0.44666675,
      "accuracy_std_dev": 0.031972270973259,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.3239053575,
      "avg_duration_per_task_ms": 100810.8825,
      "avg_output_tokens_per_task": 2994.355,
      "last_run_timestamp": "2026-08-29T16:12:10.427Z"
    },
    {
      "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": 0.4466665,
      "accuracy_std_dev": 0.006666499999999992,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.020371870000000004,
      "avg_duration_per_task_ms": 172937.47,
      "avg_output_tokens_per_task": 42830.02,
      "last_run_timestamp": "2026-08-01T17:38:24.753Z"
    },
    {
      "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": 0.4466665,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.20957163,
      "avg_duration_per_task_ms": 47193.8,
      "avg_output_tokens_per_task": 2215.92,
      "last_run_timestamp": "2026-09-01T12:58:37.172Z"
    },
    {
      "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": 0.4463601206896552,
      "accuracy_std_dev": 0.026366188567007943,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.01648494540229885,
      "avg_duration_per_task_ms": 206287.91091954024,
      "avg_output_tokens_per_task": 2854.1609195402298,
      "last_run_timestamp": "2026-08-29T16:11:10.102Z"
    },
    {
      "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": 0.44500016666666664,
      "accuracy_std_dev": 0.029674109553709534,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.028000296666666664,
      "avg_duration_per_task_ms": 79011.23,
      "avg_output_tokens_per_task": 2803.3,
      "last_run_timestamp": "2026-08-29T17:27:31.024Z"
    },
    {
      "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": 0.445,
      "accuracy_std_dev": 0.001667000000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06940407,
      "avg_duration_per_task_ms": 279300.93,
      "avg_output_tokens_per_task": 3446.49,
      "last_run_timestamp": "2026-09-01T12:58:16.471Z"
    },
    {
      "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": 0.4444444444444444,
      "accuracy_std_dev": 0.02196410423850697,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0480205202020202,
      "avg_duration_per_task_ms": 850576.6565656565,
      "avg_output_tokens_per_task": 104088.0202020202,
      "last_run_timestamp": "2026-09-01T17:31:48.056Z"
    },
    {
      "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": 0.4444443333333334,
      "accuracy_std_dev": 0.01617809828818648,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09676157666666667,
      "avg_duration_per_task_ms": 911223.4233333333,
      "avg_output_tokens_per_task": 20700.716666666667,
      "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-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": 0.44,
      "accuracy_std_dev": 0.013332999999999984,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02857548,
      "avg_duration_per_task_ms": 74491.22,
      "avg_output_tokens_per_task": 2449.6,
      "last_run_timestamp": "2026-09-01T17:30:53.043Z"
    },
    {
      "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/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.44,
      "accuracy_std_dev": 0.02788886675718467,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.03487082,
      "avg_duration_per_task_ms": 281485.155,
      "avg_output_tokens_per_task": 21754.5,
      "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": "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": 0.43965518965517236,
      "accuracy_std_dev": 0.020525836824522862,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.045893882183908045,
      "avg_duration_per_task_ms": 128881.75862068965,
      "avg_output_tokens_per_task": 2035.2413793103449,
      "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-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "CoreWeave",
      "accuracy": 0.43902436585365856,
      "accuracy_std_dev": 0.046630122309082575,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.15270590650406504,
      "avg_duration_per_task_ms": 175421.88617886178,
      "avg_output_tokens_per_task": 6311.357723577235,
      "last_run_timestamp": "2026-08-30T00:01:09.486Z"
    },
    {
      "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": 0.4387170412371134,
      "accuracy_std_dev": 0.022793011768351766,
      "run_count": 6,
      "total_questions": 291,
      "avg_cost_per_task_usd": 0.031083333333333338,
      "avg_duration_per_task_ms": 487106.5326460481,
      "avg_output_tokens_per_task": 4378.945017182131,
      "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/qwen3.6-35b-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": "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": 0.4379670089686099,
      "accuracy_std_dev": 0.04940897449394402,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.03162573766816144,
      "avg_duration_per_task_ms": 87326.9730941704,
      "avg_output_tokens_per_task": 3019.251121076233,
      "last_run_timestamp": "2026-08-29T17:17:42.808Z"
    },
    {
      "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": 0.4373334,
      "accuracy_std_dev": 0.019136942750606727,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.077255652,
      "avg_duration_per_task_ms": 788193.684,
      "avg_output_tokens_per_task": 19596.964,
      "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": "openai/gpt-4o",
      "model_slug": "openai/gpt-4o",
      "model_name": "OpenAI: GPT-4o",
      "provider_name": "OpenAI",
      "accuracy": 0.43666675,
      "accuracy_std_dev": 0.011055139582452123,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.21359883,
      "avg_duration_per_task_ms": 47801.905,
      "avg_output_tokens_per_task": 2174.32,
      "last_run_timestamp": "2026-09-01T13:02:23.243Z"
    },
    {
      "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": "Parasail",
      "accuracy": 0.4353740544217687,
      "accuracy_std_dev": 0.0333418611528944,
      "run_count": 6,
      "total_questions": 294,
      "avg_cost_per_task_usd": 0.03948419727891156,
      "avg_duration_per_task_ms": 53655.58163265306,
      "avg_output_tokens_per_task": 2073.0680272108843,
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "DeepInfra",
      "accuracy": 0.43348282959641254,
      "accuracy_std_dev": 0.021733670805150856,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.024745609865470853,
      "avg_duration_per_task_ms": 267535.8094170404,
      "avg_output_tokens_per_task": 3614.85201793722,
      "last_run_timestamp": "2026-08-29T16:17:35.870Z"
    },
    {
      "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": 0.4333336666666667,
      "accuracy_std_dev": 0.0339934634239519,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03350491333333334,
      "avg_duration_per_task_ms": 439900.87333333335,
      "avg_output_tokens_per_task": 22781.906666666666,
      "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/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.43333333333333335,
      "accuracy_std_dev": 0.014401440259764139,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03429177333333333,
      "avg_duration_per_task_ms": 515166.3933333333,
      "avg_output_tokens_per_task": 24569.606666666667,
      "last_run_timestamp": "2026-08-01T17:38:35.770Z"
    },
    {
      "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": "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": 0.43,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02958018,
      "avg_duration_per_task_ms": 73337.54,
      "avg_output_tokens_per_task": 2441.2,
      "last_run_timestamp": "2026-09-01T17:31:46.578Z"
    },
    {
      "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": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.23500686999999998,
      "avg_duration_per_task_ms": 47776.24,
      "avg_output_tokens_per_task": 2493.12,
      "last_run_timestamp": "2026-09-01T13:00:52.000Z"
    },
    {
      "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": 0.42481477777777776,
      "accuracy_std_dev": 0.01055379582875573,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.024998100000000002,
      "avg_duration_per_task_ms": 230644.6577777778,
      "avg_output_tokens_per_task": 3644.9155555555553,
      "last_run_timestamp": "2026-08-29T16:50:29.867Z"
    },
    {
      "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": 0.423333,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02960835,
      "avg_duration_per_task_ms": 524604.61,
      "avg_output_tokens_per_task": 20056.49,
      "last_run_timestamp": "2026-09-01T12:41:25.851Z"
    },
    {
      "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": 0.4222224444444444,
      "accuracy_std_dev": 0.03624331355991708,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05411629777777777,
      "avg_duration_per_task_ms": 135677.24,
      "avg_output_tokens_per_task": 3556.682222222222,
      "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": 0.4193990491803279,
      "accuracy_std_dev": 0.02746729167500866,
      "run_count": 10,
      "total_questions": 488,
      "avg_cost_per_task_usd": 0.05026936680327868,
      "avg_duration_per_task_ms": 211126.46106557376,
      "avg_output_tokens_per_task": 4020.3565573770493,
      "last_run_timestamp": "2026-09-26T08:09:25.362Z"
    },
    {
      "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": 0.41851855555555556,
      "accuracy_std_dev": 0.03580157095594522,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.0414691511111111,
      "avg_duration_per_task_ms": 80026.80444444444,
      "avg_output_tokens_per_task": 3600.311111111111,
      "last_run_timestamp": "2026-08-29T16:10:58.039Z"
    },
    {
      "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": 0.4183674081632653,
      "accuracy_std_dev": 0.008472500000000022,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.048346357142857135,
      "avg_duration_per_task_ms": 818319.5204081633,
      "avg_output_tokens_per_task": 111719.33673469388,
      "last_run_timestamp": "2026-09-01T17:32:02.570Z"
    },
    {
      "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": 0.41666675,
      "accuracy_std_dev": 0.013743584364622648,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.027539789999999998,
      "avg_duration_per_task_ms": 73803.845,
      "avg_output_tokens_per_task": 2931.665,
      "last_run_timestamp": "2026-09-01T17:32:02.600Z"
    },
    {
      "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/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "StreamLake",
      "accuracy": 0.41306208651399495,
      "accuracy_std_dev": 0.03034144422624565,
      "run_count": 8,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.01169636641221374,
      "avg_duration_per_task_ms": 132148.3969465649,
      "avg_output_tokens_per_task": 3534.709923664122,
      "last_run_timestamp": "2026-08-29T16:17:15.091Z"
    },
    {
      "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": 0.41228081754385965,
      "accuracy_std_dev": 0.14934019625078246,
      "run_count": 6,
      "total_questions": 285,
      "avg_cost_per_task_usd": 0.03923479298245614,
      "avg_duration_per_task_ms": 212298.04912280702,
      "avg_output_tokens_per_task": 88632.91228070176,
      "last_run_timestamp": "2026-08-29T17:06:38.202Z"
    },
    {
      "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": "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": 0.40851855555555555,
      "accuracy_std_dev": 0.0222838565040111,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.06348359111111111,
      "avg_duration_per_task_ms": 98874.61111111111,
      "avg_output_tokens_per_task": 2927.2933333333335,
      "last_run_timestamp": "2026-08-29T16:15:12.659Z"
    },
    {
      "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": 0.40778087608069163,
      "accuracy_std_dev": 0.020533700034383498,
      "run_count": 7,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.030645242074927952,
      "avg_duration_per_task_ms": 639129.7435158502,
      "avg_output_tokens_per_task": 7195.42939481268,
      "last_run_timestamp": "2026-08-30T00:01:10.605Z"
    },
    {
      "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": 0.40666674999999997,
      "accuracy_std_dev": 0.024266872711321913,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.02623856,
      "avg_duration_per_task_ms": 288008.1325,
      "avg_output_tokens_per_task": 4171.565,
      "last_run_timestamp": "2026-08-29T16:17:08.788Z"
    },
    {
      "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": 0.40666650000000004,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06123743,
      "avg_duration_per_task_ms": 234369.43,
      "avg_output_tokens_per_task": 8023.68,
      "last_run_timestamp": "2026-09-01T12:41:31.874Z"
    },
    {
      "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": 0.4051851111111111,
      "accuracy_std_dev": 0.14120402210076063,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.03654666,
      "avg_duration_per_task_ms": 265671.77777777775,
      "avg_output_tokens_per_task": 3253.0866666666666,
      "last_run_timestamp": "2026-08-29T16:33:29.143Z"
    },
    {
      "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": 0.4051851111111111,
      "accuracy_std_dev": 0.039475118659858095,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05917665777777779,
      "avg_duration_per_task_ms": 299711.4088888889,
      "avg_output_tokens_per_task": 7407.624444444445,
      "last_run_timestamp": "2026-08-29T16:11:10.099Z"
    },
    {
      "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": 0.40379423577235773,
      "accuracy_std_dev": 0.19339763406453553,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.01662922357723577,
      "avg_duration_per_task_ms": 1820667.5203252032,
      "avg_output_tokens_per_task": 111556.55284552845,
      "last_run_timestamp": "2026-08-29T17:12:32.860Z"
    },
    {
      "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-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.3977775555555556,
      "accuracy_std_dev": 0.01859228567092104,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.07825311555555554,
      "avg_duration_per_task_ms": 125269.89111111111,
      "avg_output_tokens_per_task": 3131.8733333333334,
      "last_run_timestamp": "2026-08-29T16:33:39.765Z"
    },
    {
      "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": 0.3971632127659574,
      "accuracy_std_dev": 0.3419385,
      "run_count": 2,
      "total_questions": 94,
      "avg_cost_per_task_usd": 0.030152606382978724,
      "avg_duration_per_task_ms": 146561.8404255319,
      "avg_output_tokens_per_task": 4856.648936170212,
      "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": "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": 0.39666650000000003,
      "accuracy_std_dev": 0.02333349999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.0319114,
      "avg_duration_per_task_ms": 79638.86,
      "avg_output_tokens_per_task": 5369.64,
      "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": "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": 0.39062490277777784,
      "accuracy_std_dev": 0.02073837202707419,
      "run_count": 6,
      "total_questions": 288,
      "avg_cost_per_task_usd": 0.010834395833333335,
      "avg_duration_per_task_ms": 31261.70486111111,
      "avg_output_tokens_per_task": 586.9097222222222,
      "last_run_timestamp": "2026-08-15T14:15:11.576Z"
    },
    {
      "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": 0.3903702222222223,
      "accuracy_std_dev": 0.023541163785910455,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.03730226666666667,
      "avg_duration_per_task_ms": 195629.94666666666,
      "avg_output_tokens_per_task": 3811.4333333333334,
      "last_run_timestamp": "2026-08-29T16:31:33.864Z"
    },
    {
      "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": 0.39,
      "accuracy_std_dev": 0.023094082938060125,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.02572764,
      "avg_duration_per_task_ms": 216676.0775,
      "avg_output_tokens_per_task": 4467.8225,
      "last_run_timestamp": "2026-08-29T16:10:20.675Z"
    },
    {
      "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": 0.38666666666666666,
      "accuracy_std_dev": 0.012957518435256032,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.012331926666666666,
      "avg_duration_per_task_ms": 31907.722222222223,
      "avg_output_tokens_per_task": 547.9311111111111,
      "last_run_timestamp": "2026-08-29T16:10:09.208Z"
    },
    {
      "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": 0.385926,
      "accuracy_std_dev": 0.0079780288153013,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.01442735111111111,
      "avg_duration_per_task_ms": 79203.29111111112,
      "avg_output_tokens_per_task": 598.6177777777777,
      "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": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "auto-routing",
      "accuracy": 0.38333349999999994,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07100836,
      "avg_duration_per_task_ms": 283936.54,
      "avg_output_tokens_per_task": 7506.19,
      "last_run_timestamp": "2026-09-01T12:58:37.380Z"
    },
    {
      "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": 0.37916662500000003,
      "accuracy_std_dev": 0.02933276396343131,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.04733663,
      "avg_duration_per_task_ms": 70203.2925,
      "avg_output_tokens_per_task": 4327.02,
      "last_run_timestamp": "2026-08-29T16:25:50.803Z"
    },
    {
      "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": 0.3771427142857142,
      "accuracy_std_dev": 0.008624165809999659,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03271254285714286,
      "avg_duration_per_task_ms": 138753.44,
      "avg_output_tokens_per_task": 4451.42,
      "last_run_timestamp": "2026-08-29T16:13:32.038Z"
    },
    {
      "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": 0.3744443333333334,
      "accuracy_std_dev": 0.01940465134674902,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06861020666666667,
      "avg_duration_per_task_ms": 17591.08,
      "avg_output_tokens_per_task": 793.6233333333333,
      "last_run_timestamp": "2026-08-29T16:18:42.229Z"
    },
    {
      "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": 0.3712121363636363,
      "accuracy_std_dev": 0.203715753368511,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.01041270707070707,
      "avg_duration_per_task_ms": 88573.35858585859,
      "avg_output_tokens_per_task": 2290.378787878788,
      "last_run_timestamp": "2026-08-29T17:44:16.832Z"
    },
    {
      "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": 0.37016358035714286,
      "accuracy_std_dev": 0.060232280320582955,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.05735951562499999,
      "avg_duration_per_task_ms": 340422.8660714286,
      "avg_output_tokens_per_task": 3670.198660714286,
      "last_run_timestamp": "2026-08-29T16:13:54.381Z"
    },
    {
      "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": 0.3688886666666667,
      "accuracy_std_dev": 0.011331263428624758,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.07234757333333333,
      "avg_duration_per_task_ms": 16758.733333333334,
      "avg_output_tokens_per_task": 864.72,
      "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": "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DekaLLM",
      "accuracy": 0.3639454693877551,
      "accuracy_std_dev": 0.261389,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.004816234693877552,
      "avg_duration_per_task_ms": 133358.81632653062,
      "avg_output_tokens_per_task": 12437.69387755102,
      "last_run_timestamp": "2026-07-11T15:34:14.382Z"
    },
    {
      "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": 0.362963,
      "accuracy_std_dev": 0.09661485400174123,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05752990444444444,
      "avg_duration_per_task_ms": 257852.61777777778,
      "avg_output_tokens_per_task": 10247.41111111111,
      "last_run_timestamp": "2026-08-29T16:11:34.922Z"
    },
    {
      "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": 0.3616235461254612,
      "accuracy_std_dev": 0.028902948389451646,
      "run_count": 11,
      "total_questions": 542,
      "avg_cost_per_task_usd": 0.07057725645756457,
      "avg_duration_per_task_ms": 374277.7712177122,
      "avg_output_tokens_per_task": 7489.30258302583,
      "last_run_timestamp": "2026-09-26T08:14:46.867Z"
    },
    {
      "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": 0.358859,
      "accuracy_std_dev": 0.037673584399498986,
      "run_count": 9,
      "total_questions": 444,
      "avg_cost_per_task_usd": 0.021177574324324325,
      "avg_duration_per_task_ms": 100388.78153153154,
      "avg_output_tokens_per_task": 1092.0855855855855,
      "last_run_timestamp": "2026-08-29T16:14:38.085Z"
    },
    {
      "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": 0.3555554444444445,
      "accuracy_std_dev": 0.3294587300632867,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021335997777777782,
      "avg_duration_per_task_ms": 65350.72222222222,
      "avg_output_tokens_per_task": 11586.551111111112,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "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": 0.35542168674698793,
      "accuracy_std_dev": 0.02435901384703412,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.08032603012048192,
      "avg_duration_per_task_ms": 238054.84738955824,
      "avg_output_tokens_per_task": 7629.79718875502,
      "last_run_timestamp": "2026-09-26T08:01:16.588Z"
    },
    {
      "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": 0.35266659999999994,
      "accuracy_std_dev": 0.014047424441512414,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.039413411999999995,
      "avg_duration_per_task_ms": 97771.824,
      "avg_output_tokens_per_task": 4606.912,
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Chutes",
      "accuracy": 0.34641463372093023,
      "accuracy_std_dev": 0.07304902258240625,
      "run_count": 7,
      "total_questions": 344,
      "avg_cost_per_task_usd": 0.00352575,
      "avg_duration_per_task_ms": 248041.55523255814,
      "avg_output_tokens_per_task": 3801.1162790697676,
      "last_run_timestamp": "2026-08-29T16:33:36.263Z"
    },
    {
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Parasail",
      "accuracy": 0.3370835,
      "accuracy_std_dev": 0.06602697155140465,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.004872619999999999,
      "avg_duration_per_task_ms": 111751.1125,
      "avg_output_tokens_per_task": 3307.625,
      "last_run_timestamp": "2026-08-29T17:02:26.939Z"
    },
    {
      "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": 0.3322147852348994,
      "accuracy_std_dev": 0.07110040643109339,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.03765286577181208,
      "avg_duration_per_task_ms": 95171.10067114094,
      "avg_output_tokens_per_task": 6163.89932885906,
      "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": "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": 0.3309091818181818,
      "accuracy_std_dev": 0.022966516131961975,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.08758338363636364,
      "avg_duration_per_task_ms": 154847.89454545453,
      "avg_output_tokens_per_task": 8333.869090909091,
      "last_run_timestamp": "2026-09-26T08:09:25.338Z"
    },
    {
      "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": 0.3219998,
      "accuracy_std_dev": 0.024908747237868067,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.03118176,
      "avg_duration_per_task_ms": 78513.244,
      "avg_output_tokens_per_task": 3895.62,
      "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": "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": 0.32068464732142854,
      "accuracy_std_dev": 0.18170803102970257,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.12890572098214284,
      "avg_duration_per_task_ms": 772274.0200892857,
      "avg_output_tokens_per_task": 5070.225446428572,
      "last_run_timestamp": "2026-09-26T08:05:39.969Z"
    },
    {
      "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": "Groq",
      "accuracy": 0.32,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.07498856,
      "avg_duration_per_task_ms": 19652.8,
      "avg_output_tokens_per_task": 2138.96,
      "last_run_timestamp": "2026-07-04T16:00:20.920Z"
    },
    {
      "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": 0.3143632967479675,
      "accuracy_std_dev": 0.23029092136912385,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.008914227642276422,
      "avg_duration_per_task_ms": 59356.63821138212,
      "avg_output_tokens_per_task": 1974.3902439024391,
      "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": "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": 0.31282059999999995,
      "accuracy_std_dev": 0.030993226130882217,
      "run_count": 8,
      "total_questions": 390,
      "avg_cost_per_task_usd": 0.015985989743589744,
      "avg_duration_per_task_ms": 332807.73846153845,
      "avg_output_tokens_per_task": 4072.7410256410258,
      "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": 0.3081481111111111,
      "accuracy_std_dev": 0.03755203650771211,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.15946950444444447,
      "avg_duration_per_task_ms": 119529.95333333334,
      "avg_output_tokens_per_task": 1809.4622222222222,
      "last_run_timestamp": "2026-08-29T17:05:59.043Z"
    },
    {
      "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": 0.27,
      "accuracy_std_dev": 0.03,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02140288,
      "avg_duration_per_task_ms": 65620.09,
      "avg_output_tokens_per_task": 3949.38,
      "last_run_timestamp": "2026-09-01T17:29:05.378Z"
    },
    {
      "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": "DeepInfra",
      "accuracy": 0.266667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00574078,
      "avg_duration_per_task_ms": 359216.92,
      "avg_output_tokens_per_task": 2727.92,
      "last_run_timestamp": "2026-07-11T08:03:53.497Z"
    },
    {
      "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": 0.266667,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02052575,
      "avg_duration_per_task_ms": 64186.84,
      "avg_output_tokens_per_task": 3663.19,
      "last_run_timestamp": "2026-09-01T17:31:47.037Z"
    },
    {
      "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": 0.26148144444444443,
      "accuracy_std_dev": 0.06338407627246437,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.015817097777777775,
      "avg_duration_per_task_ms": 25942.504444444443,
      "avg_output_tokens_per_task": 1328.6466666666668,
      "last_run_timestamp": "2026-08-29T16:17:15.091Z"
    },
    {
      "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": 0.26041658333333334,
      "accuracy_std_dev": 0.01347799999999999,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.07850963541666667,
      "avg_duration_per_task_ms": 588191.84375,
      "avg_output_tokens_per_task": 5667.5625,
      "last_run_timestamp": "2026-09-01T12:59:55.605Z"
    },
    {
      "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": "Novita",
      "accuracy": 0.25,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.08593323,
      "avg_duration_per_task_ms": 547480.87,
      "avg_output_tokens_per_task": 5216.43,
      "last_run_timestamp": "2026-09-01T12:57:03.395Z"
    },
    {
      "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": 0.23299353061224487,
      "accuracy_std_dev": 0.02687500000000001,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.028852255102040817,
      "avg_duration_per_task_ms": 377353.28571428574,
      "avg_output_tokens_per_task": 5486.489795918367,
      "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": 0.2316665,
      "accuracy_std_dev": 0.018333500000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02960246,
      "avg_duration_per_task_ms": 418677.18,
      "avg_output_tokens_per_task": 5444.6,
      "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": "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": 0.21166649999999998,
      "accuracy_std_dev": 0.046278414695730456,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.0049267899999999995,
      "avg_duration_per_task_ms": 45675.635,
      "avg_output_tokens_per_task": 1348.715,
      "last_run_timestamp": "2026-08-22T16:04:58.329Z"
    },
    {
      "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": 0.20642976649746192,
      "accuracy_std_dev": 0.1484845071697297,
      "run_count": 8,
      "total_questions": 394,
      "avg_cost_per_task_usd": 0.17611084517766495,
      "avg_duration_per_task_ms": 524637.7563451777,
      "avg_output_tokens_per_task": 9406.852791878173,
      "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": 0.19841299319727893,
      "accuracy_std_dev": 0.016383133956060492,
      "run_count": 3,
      "total_questions": 147,
      "avg_cost_per_task_usd": 0.011659061224489795,
      "avg_duration_per_task_ms": 134707.1224489796,
      "avg_output_tokens_per_task": 5676.0204081632655,
      "last_run_timestamp": "2026-09-01T13:00:50.461Z"
    },
    {
      "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": 0.17333325,
      "accuracy_std_dev": 0.11522926663368771,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0127158975,
      "avg_duration_per_task_ms": 62680.3375,
      "avg_output_tokens_per_task": 1280.3475,
      "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": "OpenAI",
      "accuracy": 0.1366665,
      "accuracy_std_dev": 0.003333500000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00694106,
      "avg_duration_per_task_ms": 56167.4,
      "avg_output_tokens_per_task": 2292.83,
      "last_run_timestamp": "2026-09-01T17:31:40.721Z"
    },
    {
      "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": 0.125,
      "accuracy_std_dev": 0.020750046855368784,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.00834746,
      "avg_duration_per_task_ms": 82234.105,
      "avg_output_tokens_per_task": 2517.12,
      "last_run_timestamp": "2026-09-01T17:31:43.416Z"
    },
    {
      "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": 0.1033335,
      "accuracy_std_dev": 0.003333499999999996,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.0070183400000000005,
      "avg_duration_per_task_ms": 61321.36,
      "avg_output_tokens_per_task": 2324.09,
      "last_run_timestamp": "2026-09-01T17:31:46.301Z"
    },
    {
      "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": "SambaNova",
      "accuracy": 0.013333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0036388199999999996,
      "avg_duration_per_task_ms": 4217.8,
      "avg_output_tokens_per_task": 129.92,
      "last_run_timestamp": "2026-07-04T16:00:25.000Z"
    }
  ]
}
