{
  "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": "2026-09-21",
  "retrieved_at": "2026-09-21T01:46:02.485Z",
  "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 newest 10,000 non-zero-score runs: 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": 2068,
  "rows": [
    {
      "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": 0.955357,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 56,
      "avg_cost_per_task_usd": 0.002410089285714286,
      "avg_duration_per_task_ms": 98368.17857142857,
      "avg_output_tokens_per_task": 4744.517857142857,
      "last_run_timestamp": "2026-09-16T18:25:41.949Z"
    },
    {
      "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": "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": "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": 0.9343432500000001,
      "accuracy_std_dev": 0.006628136630117082,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.1361736212121212,
      "avg_duration_per_task_ms": 193730.40025252526,
      "avg_output_tokens_per_task": 8868.685606060606,
      "last_run_timestamp": "2026-09-19T09:16:33.297Z"
    },
    {
      "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": "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Relace",
      "accuracy": 0.9305555,
      "accuracy_std_dev": 0.0012624999999999997,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.08957261616161614,
      "avg_duration_per_task_ms": 260418.87121212122,
      "avg_output_tokens_per_task": 8649.169191919193,
      "last_run_timestamp": "2026-09-19T09:50:43.404Z"
    },
    {
      "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": "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Fireworks",
      "accuracy": 0.9290078158277567,
      "accuracy_std_dev": 0.008556345120594102,
      "run_count": 19,
      "total_questions": 3437,
      "avg_cost_per_task_usd": 0.15611512365434976,
      "avg_duration_per_task_ms": 137193.63078265928,
      "avg_output_tokens_per_task": 8758.286005237125,
      "last_run_timestamp": "2026-09-19T08:58:50.122Z"
    },
    {
      "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": "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": "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Makora",
      "accuracy": 0.9254603241206031,
      "accuracy_std_dev": 0.020055971452135855,
      "run_count": 3,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.08006439949748743,
      "avg_duration_per_task_ms": 226382.3567839196,
      "avg_output_tokens_per_task": 6134.640703517588,
      "last_run_timestamp": "2026-09-05T08:30:49.088Z"
    },
    {
      "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": 0.9240362551020408,
      "accuracy_std_dev": 0.010411205213614818,
      "run_count": 3,
      "total_questions": 588,
      "avg_cost_per_task_usd": 0.11667940476190476,
      "avg_duration_per_task_ms": 366799.1649659864,
      "avg_output_tokens_per_task": 19230.027210884353,
      "last_run_timestamp": "2026-09-01T13:27:43.904Z"
    },
    {
      "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": 0.9225940304841601,
      "accuracy_std_dev": 0.0076093798566288715,
      "run_count": 9,
      "total_questions": 1673,
      "avg_cost_per_task_usd": 0.14876478720860728,
      "avg_duration_per_task_ms": 123940.04423191871,
      "avg_output_tokens_per_task": 9777.336521219366,
      "last_run_timestamp": "2026-09-19T09:02:06.672Z"
    },
    {
      "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": "Alibaba",
      "accuracy": 0.9221377500000001,
      "accuracy_std_dev": 0.007277616312193182,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.1708342297979798,
      "avg_duration_per_task_ms": 352657.0984848485,
      "avg_output_tokens_per_task": 9736.763888888889,
      "last_run_timestamp": "2026-09-19T08:58:45.410Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "NextBit",
      "accuracy": 0.92087525,
      "accuracy_std_dev": 0.005952071335888052,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.137514726010101,
      "avg_duration_per_task_ms": 557212.4103535353,
      "avg_output_tokens_per_task": 37117.12752525252,
      "last_run_timestamp": "2026-09-19T09:21:26.618Z"
    },
    {
      "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": "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": 0.9208333333333333,
      "accuracy_std_dev": 0.005676000000000014,
      "run_count": 2,
      "total_questions": 360,
      "avg_cost_per_task_usd": 0.0029125277777777778,
      "avg_duration_per_task_ms": 461662.2555555556,
      "avg_output_tokens_per_task": 10228.08888888889,
      "last_run_timestamp": "2026-09-19T09:15:41.550Z"
    },
    {
      "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": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07324645454545455,
      "avg_duration_per_task_ms": 236783.91666666666,
      "avg_output_tokens_per_task": 25006.20707070707,
      "last_run_timestamp": "2026-09-20T10:38:54.342Z"
    },
    {
      "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": 0.9194322857142857,
      "accuracy_std_dev": 0.008972904191656873,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.17085740764790766,
      "avg_duration_per_task_ms": 272930.75685425685,
      "avg_output_tokens_per_task": 13023.35064935065,
      "last_run_timestamp": "2026-09-19T08:49:38.168Z"
    },
    {
      "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": 0.9185987869565218,
      "accuracy_std_dev": 0.013465663103476456,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.1389484079710145,
      "avg_duration_per_task_ms": 369171.10942028987,
      "avg_output_tokens_per_task": 10521.670289855072,
      "last_run_timestamp": "2026-09-19T08:59:31.029Z"
    },
    {
      "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": 0.9185184,
      "accuracy_std_dev": 0.0060794207816205845,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1914820111111111,
      "avg_duration_per_task_ms": 356646.5292929293,
      "avg_output_tokens_per_task": 43359.274747474745,
      "last_run_timestamp": "2026-09-20T10:34:51.684Z"
    },
    {
      "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": 0.9176481996466431,
      "accuracy_std_dev": 0.014209394819477809,
      "run_count": 9,
      "total_questions": 1698,
      "avg_cost_per_task_usd": 0.118061925795053,
      "avg_duration_per_task_ms": 124721.6089517079,
      "avg_output_tokens_per_task": 7956.564782096584,
      "last_run_timestamp": "2026-09-19T09:54:41.293Z"
    },
    {
      "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": 0.9163074598792063,
      "accuracy_std_dev": 0.01614503396730882,
      "run_count": 6,
      "total_questions": 1159,
      "avg_cost_per_task_usd": 0.007566448662640208,
      "avg_duration_per_task_ms": 499842.77135461604,
      "avg_output_tokens_per_task": 19527.920621225196,
      "last_run_timestamp": "2026-09-20T10:39:48.638Z"
    },
    {
      "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": 0.9155657692307693,
      "accuracy_std_dev": 0.012784482626358756,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.02835553613053613,
      "avg_duration_per_task_ms": 214027.3453768454,
      "avg_output_tokens_per_task": 23521.37567987568,
      "last_run_timestamp": "2026-09-19T16:08:21.510Z"
    },
    {
      "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": "Baidu",
      "accuracy": 0.91456225,
      "accuracy_std_dev": 0.004014799893830335,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.10835830176767677,
      "avg_duration_per_task_ms": 621940.4457070708,
      "avg_output_tokens_per_task": 40133.22601010101,
      "last_run_timestamp": "2026-09-19T08:53:55.203Z"
    },
    {
      "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": 0.9138044000000001,
      "accuracy_std_dev": 0.003749296606031599,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.28000743232323233,
      "avg_duration_per_task_ms": 356725.703030303,
      "avg_output_tokens_per_task": 38588.78080808081,
      "last_run_timestamp": "2026-09-19T08:49:14.867Z"
    },
    {
      "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": 0.9133028235677082,
      "accuracy_std_dev": 0.006692965934471823,
      "run_count": 8,
      "total_questions": 1536,
      "avg_cost_per_task_usd": 0.1390292682291667,
      "avg_duration_per_task_ms": 291227.6529947917,
      "avg_output_tokens_per_task": 9127.681640625,
      "last_run_timestamp": "2026-09-19T08:55:48.574Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "AtlasCloud",
      "accuracy": 0.9132995,
      "accuracy_std_dev": 0.01430950000000003,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.15749233333333332,
      "avg_duration_per_task_ms": 547358.5101010101,
      "avg_output_tokens_per_task": 39453.482323232325,
      "last_run_timestamp": "2026-09-19T08:48:18.401Z"
    },
    {
      "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": 0.9122337978723405,
      "accuracy_std_dev": 0.011253882096414569,
      "run_count": 5,
      "total_questions": 940,
      "avg_cost_per_task_usd": 0.01295165319148936,
      "avg_duration_per_task_ms": 355502.26914893615,
      "avg_output_tokens_per_task": 16382.894680851065,
      "last_run_timestamp": "2026-09-20T10:40:15.786Z"
    },
    {
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.9099324999999999,
      "accuracy_std_dev": 0.005892500000000023,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.030082277777777777,
      "avg_duration_per_task_ms": 178098.15404040404,
      "avg_output_tokens_per_task": 24895.11111111111,
      "last_run_timestamp": "2026-09-19T08:54:37.084Z"
    },
    {
      "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": 0.909652,
      "accuracy_std_dev": 0.0020998715833752102,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03531527104377104,
      "avg_duration_per_task_ms": 174484.34343434343,
      "avg_output_tokens_per_task": 29238.693602693602,
      "last_run_timestamp": "2026-09-19T08:07:24.139Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stealth/union-alpha",
      "model_slug": "stealth/union-alpha",
      "model_name": "Union Alpha",
      "provider_name": "auto-routing",
      "accuracy": 0.9090909999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": null,
      "avg_duration_per_task_ms": 146448.00505050505,
      "avg_output_tokens_per_task": 4567.823232323232,
      "last_run_timestamp": "2026-09-15T01:07:47.954Z"
    },
    {
      "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": 0.9090906666666667,
      "accuracy_std_dev": 0.0036369299629709197,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.037642313131313124,
      "avg_duration_per_task_ms": 128088.45959595959,
      "avg_output_tokens_per_task": 28275.836700336702,
      "last_run_timestamp": "2026-09-19T08:49:31.119Z"
    },
    {
      "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": 0.9087186329479767,
      "accuracy_std_dev": 0.004670283372078221,
      "run_count": 7,
      "total_questions": 1384,
      "avg_cost_per_task_usd": 0.0291304161849711,
      "avg_duration_per_task_ms": 405484.31502890174,
      "avg_output_tokens_per_task": 30120.609826589596,
      "last_run_timestamp": "2026-09-19T16:37:44.222Z"
    },
    {
      "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": 0.9086567723440133,
      "accuracy_std_dev": 0.002243205890386943,
      "run_count": 3,
      "total_questions": 593,
      "avg_cost_per_task_usd": 0.0342340826306914,
      "avg_duration_per_task_ms": 141097.99494097807,
      "avg_output_tokens_per_task": 28836.671163575043,
      "last_run_timestamp": "2026-09-19T08:47:59.078Z"
    },
    {
      "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": 0.907407,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00787879292929293,
      "avg_duration_per_task_ms": 237113.06565656565,
      "avg_output_tokens_per_task": 23447.141414141413,
      "last_run_timestamp": "2026-09-15T01:07:47.971Z"
    },
    {
      "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": 0.9070287642172525,
      "accuracy_std_dev": 0.016195054520979538,
      "run_count": 8,
      "total_questions": 1565,
      "avg_cost_per_task_usd": 0.1865306019169329,
      "avg_duration_per_task_ms": 516947.6555910543,
      "avg_output_tokens_per_task": 12424.835782747603,
      "last_run_timestamp": "2026-09-19T08:02:29.872Z"
    },
    {
      "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": 0.9067542764456982,
      "accuracy_std_dev": 0.021950605534673606,
      "run_count": 11,
      "total_questions": 2127,
      "avg_cost_per_task_usd": 0.0353538086506817,
      "avg_duration_per_task_ms": 260805.93888105312,
      "avg_output_tokens_per_task": 29657.3121767748,
      "last_run_timestamp": "2026-09-19T16:10:18.592Z"
    },
    {
      "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": 0.906285,
      "accuracy_std_dev": 0.004827531943619518,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03418770707070708,
      "avg_duration_per_task_ms": 132654.23905723906,
      "avg_output_tokens_per_task": 28802.200336700338,
      "last_run_timestamp": "2026-09-19T08:48:26.475Z"
    },
    {
      "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": 0.9056333457943925,
      "accuracy_std_dev": 0.02007868381757514,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.03042957359813084,
      "avg_duration_per_task_ms": 291576.496105919,
      "avg_output_tokens_per_task": 25313.03971962617,
      "last_run_timestamp": "2026-09-19T16:48:55.177Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Alibaba",
      "accuracy": 0.904882,
      "accuracy_std_dev": 0.020544122168315374,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.13592807828282827,
      "avg_duration_per_task_ms": 578809.478114478,
      "avg_output_tokens_per_task": 38384.35353535353,
      "last_run_timestamp": "2026-09-19T08:57:25.509Z"
    },
    {
      "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": 0.9042877457886676,
      "accuracy_std_dev": 0.034874145815887096,
      "run_count": 14,
      "total_questions": 2612,
      "avg_cost_per_task_usd": 0.1541255895865237,
      "avg_duration_per_task_ms": 517283.3177641654,
      "avg_output_tokens_per_task": 8744.607963246553,
      "last_run_timestamp": "2026-09-19T12:44:31.587Z"
    },
    {
      "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": 0.9041560076335877,
      "accuracy_std_dev": 0.003276000000000001,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.021076463104325698,
      "avg_duration_per_task_ms": 791723.9363867685,
      "avg_output_tokens_per_task": 17407.793893129772,
      "last_run_timestamp": "2026-09-19T09:03:39.099Z"
    },
    {
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Together",
      "accuracy": 0.9037548157099697,
      "accuracy_std_dev": 0.016887706798429505,
      "run_count": 12,
      "total_questions": 2317,
      "avg_cost_per_task_usd": 0.031995520932239965,
      "avg_duration_per_task_ms": 304192.94389296504,
      "avg_output_tokens_per_task": 26670.303409581356,
      "last_run_timestamp": "2026-09-19T16:45:47.346Z"
    },
    {
      "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": 0.9013081934015927,
      "accuracy_std_dev": 0.01190368377259848,
      "run_count": 9,
      "total_questions": 1758,
      "avg_cost_per_task_usd": 0.03069646018202503,
      "avg_duration_per_task_ms": 230610.76109215018,
      "avg_output_tokens_per_task": 25458.812855517634,
      "last_run_timestamp": "2026-09-19T16:27:17.903Z"
    },
    {
      "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": 0.9012855000000001,
      "accuracy_std_dev": 0.00564207676746779,
      "run_count": 22,
      "total_questions": 4356,
      "avg_cost_per_task_usd": 0.21591182460973365,
      "avg_duration_per_task_ms": 95090.06290174472,
      "avg_output_tokens_per_task": 7868.12924701561,
      "last_run_timestamp": "2026-09-19T08:52:57.055Z"
    },
    {
      "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": 0.9006734999999999,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03283414393939394,
      "avg_duration_per_task_ms": 255154.72727272726,
      "avg_output_tokens_per_task": 28959.247474747473,
      "last_run_timestamp": "2026-09-19T09:07:59.395Z"
    },
    {
      "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": 0.9006733333333333,
      "accuracy_std_dev": 0.007653512933432718,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.12550200336700337,
      "avg_duration_per_task_ms": 30223.45286195286,
      "avg_output_tokens_per_task": 2299.0185185185187,
      "last_run_timestamp": "2026-09-19T08:48:46.513Z"
    },
    {
      "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": 0.9006732857142857,
      "accuracy_std_dev": 0.005831732595140517,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.20125868181818185,
      "avg_duration_per_task_ms": 98286.09018759019,
      "avg_output_tokens_per_task": 7826.790764790765,
      "last_run_timestamp": "2026-09-19T08:55:19.818Z"
    },
    {
      "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": "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": 0.9003924999999999,
      "accuracy_std_dev": 0.005181281268502329,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.23008475925925925,
      "avg_duration_per_task_ms": 111602.88636363637,
      "avg_output_tokens_per_task": 9003.594276094276,
      "last_run_timestamp": "2026-09-19T08:48:22.873Z"
    },
    {
      "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": 0.9002196347031963,
      "accuracy_std_dev": 0.02634610017364999,
      "run_count": 10,
      "total_questions": 1971,
      "avg_cost_per_task_usd": 0.043411074581430745,
      "avg_duration_per_task_ms": 257713.3805175038,
      "avg_output_tokens_per_task": 31013.896499238966,
      "last_run_timestamp": "2026-09-19T16:06:20.530Z"
    },
    {
      "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": "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "StreamLake",
      "accuracy": 0.8992302857142858,
      "accuracy_std_dev": 0.03276088896279607,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.04223221067821068,
      "avg_duration_per_task_ms": 271453.88095238095,
      "avg_output_tokens_per_task": 31163.87950937951,
      "last_run_timestamp": "2026-09-19T16:00:30.869Z"
    },
    {
      "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": 0.8990668549618319,
      "accuracy_std_dev": 0.0016184999999999672,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.01465777099236641,
      "avg_duration_per_task_ms": 713790.7251908397,
      "avg_output_tokens_per_task": 29013.936386768448,
      "last_run_timestamp": "2026-09-19T09:05:22.577Z"
    },
    {
      "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": 0.8989897500000001,
      "accuracy_std_dev": 0.005831706820262806,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.008321901515151515,
      "avg_duration_per_task_ms": 193168.351010101,
      "avg_output_tokens_per_task": 25995.64835858586,
      "last_run_timestamp": "2026-09-19T09:01:52.624Z"
    },
    {
      "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": 0.8986529999999999,
      "accuracy_std_dev": 0.004441544821343102,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.03175560707070707,
      "avg_duration_per_task_ms": 205238.4111111111,
      "avg_output_tokens_per_task": 26427.771717171716,
      "last_run_timestamp": "2026-09-19T09:12:15.523Z"
    },
    {
      "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": 0.8985087142857142,
      "accuracy_std_dev": 0.025191571054701628,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.15058746031746031,
      "avg_duration_per_task_ms": 621838.7965367965,
      "avg_output_tokens_per_task": 38175.91341991342,
      "last_run_timestamp": "2026-09-19T09:15:37.835Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.898148,
      "accuracy_std_dev": 0.014408658178331523,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.035229258838383834,
      "avg_duration_per_task_ms": 168991.77904040404,
      "avg_output_tokens_per_task": 26420.739898989897,
      "last_run_timestamp": "2026-09-19T08:49:42.594Z"
    },
    {
      "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": 0.898148,
      "accuracy_std_dev": 0.04651015131398736,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.1383203263888889,
      "avg_duration_per_task_ms": 228916.913510101,
      "avg_output_tokens_per_task": 9157.267676767677,
      "last_run_timestamp": "2026-09-19T08:48:42.819Z"
    },
    {
      "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": 0.8978918260869566,
      "accuracy_std_dev": 0.006291373173206031,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.2227677929292929,
      "avg_duration_per_task_ms": 95189.05489679403,
      "avg_output_tokens_per_task": 8106.728590250329,
      "last_run_timestamp": "2026-09-19T08:48:23.715Z"
    },
    {
      "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": 0.8977272500000001,
      "accuracy_std_dev": 0.05811571909292271,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.09910069823232324,
      "avg_duration_per_task_ms": 237858.36363636365,
      "avg_output_tokens_per_task": 8860.338383838383,
      "last_run_timestamp": "2026-09-19T08:47:59.543Z"
    },
    {
      "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": 0.8970657142857142,
      "accuracy_std_dev": 0.01398216255207516,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.2014812243867244,
      "avg_duration_per_task_ms": 97566.17965367965,
      "avg_output_tokens_per_task": 7846.960317460317,
      "last_run_timestamp": "2026-09-19T08:48:23.322Z"
    },
    {
      "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": 0.8969693999999999,
      "accuracy_std_dev": 0.016713400846027716,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.027060835353535352,
      "avg_duration_per_task_ms": 212549.52424242423,
      "avg_output_tokens_per_task": 25322.420202020203,
      "last_run_timestamp": "2026-09-19T08:53:38.936Z"
    },
    {
      "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": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Azure",
      "accuracy": 0.8962534351585013,
      "accuracy_std_dev": 0.007566251387483333,
      "run_count": 16,
      "total_questions": 3123,
      "avg_cost_per_task_usd": 0.12569597342299071,
      "avg_duration_per_task_ms": 58925.71245597182,
      "avg_output_tokens_per_task": 4794.491194364393,
      "last_run_timestamp": "2026-09-19T09:01:07.300Z"
    },
    {
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8960617898734176,
      "accuracy_std_dev": 0.01049307968079437,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.011295574261603377,
      "avg_duration_per_task_ms": 279006.29071729956,
      "avg_output_tokens_per_task": 26796.40464135021,
      "last_run_timestamp": "2026-09-19T09:37:05.170Z"
    },
    {
      "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": 0.8960286613718412,
      "accuracy_std_dev": 0.061702303840823536,
      "run_count": 7,
      "total_questions": 1385,
      "avg_cost_per_task_usd": 0.14792670397111912,
      "avg_duration_per_task_ms": 306430.2418772563,
      "avg_output_tokens_per_task": 10222.844765342961,
      "last_run_timestamp": "2026-09-19T09:07:35.756Z"
    },
    {
      "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": 0.8959034999999999,
      "accuracy_std_dev": 0.009139365308925972,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.19779903198653198,
      "avg_duration_per_task_ms": 96138.98316498316,
      "avg_output_tokens_per_task": 7715.625420875421,
      "last_run_timestamp": "2026-09-19T08:50:35.115Z"
    },
    {
      "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": 0.8953344392265193,
      "accuracy_std_dev": 0.018557569482759252,
      "run_count": 11,
      "total_questions": 2172,
      "avg_cost_per_task_usd": 0.02011020073664825,
      "avg_duration_per_task_ms": 228595.4673112339,
      "avg_output_tokens_per_task": 26682.067219152854,
      "last_run_timestamp": "2026-09-19T08:54:33.078Z"
    },
    {
      "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": 0.8951176999999999,
      "accuracy_std_dev": 0.006210770774227599,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.20379388838383836,
      "avg_duration_per_task_ms": 100040.05353535354,
      "avg_output_tokens_per_task": 7943.0055555555555,
      "last_run_timestamp": "2026-09-19T08:51:26.369Z"
    },
    {
      "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": 0.894795448275862,
      "accuracy_std_dev": 0.021807656270218488,
      "run_count": 6,
      "total_questions": 1044,
      "avg_cost_per_task_usd": 0.16143976149425288,
      "avg_duration_per_task_ms": 585867.4003831417,
      "avg_output_tokens_per_task": 37446.46647509579,
      "last_run_timestamp": "2026-09-19T09:12:47.709Z"
    },
    {
      "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": 0.8942333620253164,
      "accuracy_std_dev": 0.016976412073488586,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.12454873333333333,
      "avg_duration_per_task_ms": 626537.8101265823,
      "avg_output_tokens_per_task": 36557.81012658228,
      "last_run_timestamp": "2026-09-19T09:12:43.627Z"
    },
    {
      "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": "AkashML",
      "accuracy": 0.893939,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03405869696969697,
      "avg_duration_per_task_ms": 153813.0883838384,
      "avg_output_tokens_per_task": 9711.138888888889,
      "last_run_timestamp": "2026-06-20T09:09:33.534Z"
    },
    {
      "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": 0.893728625,
      "accuracy_std_dev": 0.007454970371126546,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.1200770505050505,
      "avg_duration_per_task_ms": 58158.018939393936,
      "avg_output_tokens_per_task": 4592.037878787879,
      "last_run_timestamp": "2026-09-19T09:07:46.594Z"
    },
    {
      "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": 0.8932607767114523,
      "accuracy_std_dev": 0.024235313237801887,
      "run_count": 8,
      "total_questions": 1563,
      "avg_cost_per_task_usd": 0.043238984005118365,
      "avg_duration_per_task_ms": 591335.9296225208,
      "avg_output_tokens_per_task": 23011.071657069737,
      "last_run_timestamp": "2026-08-22T10:03:34.994Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "Novita",
      "accuracy": 0.8931731201011378,
      "accuracy_std_dev": 0.020887674587840063,
      "run_count": 4,
      "total_questions": 791,
      "avg_cost_per_task_usd": 0.03478943742098609,
      "avg_duration_per_task_ms": 165557.26927939316,
      "avg_output_tokens_per_task": 26087.041719342604,
      "last_run_timestamp": "2026-09-19T09:00:03.663Z"
    },
    {
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DeepInfra",
      "accuracy": 0.8928136074289877,
      "accuracy_std_dev": 0.017730075367313348,
      "run_count": 7,
      "total_questions": 1373,
      "avg_cost_per_task_usd": 0.11131419956300073,
      "avg_duration_per_task_ms": 377366.75018208305,
      "avg_output_tokens_per_task": 7606.375091041515,
      "last_run_timestamp": "2026-09-19T09:20:00.091Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "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": "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": 0.8924730132827325,
      "accuracy_std_dev": 0.008810976361192897,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.22325217647058823,
      "avg_duration_per_task_ms": 107407.18848829855,
      "avg_output_tokens_per_task": 8706.91587602783,
      "last_run_timestamp": "2026-09-19T08:46:15.881Z"
    },
    {
      "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": 0.892396125,
      "accuracy_std_dev": 0.00693984391822862,
      "run_count": 24,
      "total_questions": 4752,
      "avg_cost_per_task_usd": 0.20264107049663296,
      "avg_duration_per_task_ms": 70639.64036195286,
      "avg_output_tokens_per_task": 7379.623106060606,
      "last_run_timestamp": "2026-09-19T08:49:53.737Z"
    },
    {
      "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": "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": 0.8918245230352304,
      "accuracy_std_dev": 0.00608573196090658,
      "run_count": 4,
      "total_questions": 738,
      "avg_cost_per_task_usd": 0.0061554742547425476,
      "avg_duration_per_task_ms": 427061.5867208672,
      "avg_output_tokens_per_task": 12975.516260162602,
      "last_run_timestamp": "2026-09-19T09:12:08.423Z"
    },
    {
      "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": 0.8914630362257793,
      "accuracy_std_dev": 0.03662992869870515,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.12755125863521483,
      "avg_duration_per_task_ms": 562774.203032856,
      "avg_output_tokens_per_task": 38228.18028643639,
      "last_run_timestamp": "2026-09-19T09:07:20.465Z"
    },
    {
      "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": "Decart",
      "accuracy": 0.8914139999999999,
      "accuracy_std_dev": 0.007575999999999972,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03735867676767677,
      "avg_duration_per_task_ms": 78882.84090909091,
      "avg_output_tokens_per_task": 10904.39393939394,
      "last_run_timestamp": "2026-06-20T08:34:12.555Z"
    },
    {
      "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": 0.8908169457700652,
      "accuracy_std_dev": 0.02099999680855173,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.13300457700650758,
      "avg_duration_per_task_ms": 536629.229934924,
      "avg_output_tokens_per_task": 37314.125090383226,
      "last_run_timestamp": "2026-09-19T08:51:35.037Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Inceptron",
      "accuracy": 0.8905659347708894,
      "accuracy_std_dev": 0.017636022008661704,
      "run_count": 10,
      "total_questions": 1855,
      "avg_cost_per_task_usd": 0.09373495849056605,
      "avg_duration_per_task_ms": 523293.3725067385,
      "avg_output_tokens_per_task": 28466.21778975741,
      "last_run_timestamp": "2026-09-19T08:54:04.690Z"
    },
    {
      "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": 0.8903317142857142,
      "accuracy_std_dev": 0.024732969688912254,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.08410881962481961,
      "avg_duration_per_task_ms": 171127.59163059163,
      "avg_output_tokens_per_task": 28479.191197691198,
      "last_run_timestamp": "2026-09-19T16:11:37.780Z"
    },
    {
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Phala",
      "accuracy": 0.8893674137931035,
      "accuracy_std_dev": 0.02203700000000003,
      "run_count": 2,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.045426718390804595,
      "avg_duration_per_task_ms": 158494.20977011495,
      "avg_output_tokens_per_task": 32753.790229885057,
      "last_run_timestamp": "2026-09-19T09:02:12.681Z"
    },
    {
      "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": 0.8892205742320818,
      "accuracy_std_dev": 0.012673561529979113,
      "run_count": 12,
      "total_questions": 2344,
      "avg_cost_per_task_usd": 0.0036068412969283274,
      "avg_duration_per_task_ms": 283009.04991467576,
      "avg_output_tokens_per_task": 19743.632679180886,
      "last_run_timestamp": "2026-09-19T08:50:41.183Z"
    },
    {
      "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": 0.888735406077348,
      "accuracy_std_dev": 0.010037165766694254,
      "run_count": 11,
      "total_questions": 2172,
      "avg_cost_per_task_usd": 0.2139424162062615,
      "avg_duration_per_task_ms": 103563.50506445672,
      "avg_output_tokens_per_task": 8329.639042357274,
      "last_run_timestamp": "2026-09-19T08:49:51.557Z"
    },
    {
      "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": 0.8881117317228225,
      "accuracy_std_dev": 0.00876730099564649,
      "run_count": 24,
      "total_questions": 4719,
      "avg_cost_per_task_usd": 0.13360808010171646,
      "avg_duration_per_task_ms": 59682.03581267218,
      "avg_output_tokens_per_task": 4779.846789574062,
      "last_run_timestamp": "2026-09-19T08:56:54.905Z"
    },
    {
      "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": 0.8876260833333333,
      "accuracy_std_dev": 0.013293729934686846,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.011911787037037038,
      "avg_duration_per_task_ms": 214527.07491582492,
      "avg_output_tokens_per_task": 24222.33712121212,
      "last_run_timestamp": "2026-09-19T08:49:45.167Z"
    },
    {
      "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": 0.887310375,
      "accuracy_std_dev": 0.006640514916734627,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.23988401546717172,
      "avg_duration_per_task_ms": 84791.94412878787,
      "avg_output_tokens_per_task": 8924.59659090909,
      "last_run_timestamp": "2026-09-19T08:49:34.028Z"
    },
    {
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Reka",
      "accuracy": 0.8872051666666666,
      "accuracy_std_dev": 0.008961113052455555,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.01774598653198653,
      "avg_duration_per_task_ms": 138138.01430976432,
      "avg_output_tokens_per_task": 26885.06818181818,
      "last_run_timestamp": "2026-09-19T08:49:05.000Z"
    },
    {
      "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-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": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Azure",
      "accuracy": 0.886142711026616,
      "accuracy_std_dev": 0.009981418278950896,
      "run_count": 8,
      "total_questions": 1578,
      "avg_cost_per_task_usd": 0.2087646856780735,
      "avg_duration_per_task_ms": 102336.10709759188,
      "avg_output_tokens_per_task": 8124.274397972116,
      "last_run_timestamp": "2026-09-19T08:49:20.918Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Parasail",
      "accuracy": 0.8855216666666667,
      "accuracy_std_dev": 0.016666035428046132,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.042562936026936025,
      "avg_duration_per_task_ms": 323944.202020202,
      "avg_output_tokens_per_task": 15151.54377104377,
      "last_run_timestamp": "2026-05-23T08:52:13.739Z"
    },
    {
      "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": 0.8855215,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.026893237373737375,
      "avg_duration_per_task_ms": 167702.77272727274,
      "avg_output_tokens_per_task": 22217.964646464647,
      "last_run_timestamp": "2026-09-19T09:51:03.686Z"
    },
    {
      "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": 0.8846799999999999,
      "accuracy_std_dev": 0.01856306569508389,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.12417679966329967,
      "avg_duration_per_task_ms": 507255.2138047138,
      "avg_output_tokens_per_task": 31430.781144781144,
      "last_run_timestamp": "2026-09-19T09:49:23.372Z"
    },
    {
      "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": 0.8835445189873418,
      "accuracy_std_dev": 0.001982499999999998,
      "run_count": 2,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.012081701265822782,
      "avg_duration_per_task_ms": 336769.7392405063,
      "avg_output_tokens_per_task": 21488.718987341774,
      "last_run_timestamp": "2026-09-19T11:58:10.176Z"
    },
    {
      "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": 0.8834176976530144,
      "accuracy_std_dev": 0.013545671626692473,
      "run_count": 11,
      "total_questions": 2173,
      "avg_cost_per_task_usd": 0.007175545329038196,
      "avg_duration_per_task_ms": 505749.09387942933,
      "avg_output_tokens_per_task": 25759.638748274276,
      "last_run_timestamp": "2026-09-19T09:05:08.645Z"
    },
    {
      "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": 0.8833586219178081,
      "accuracy_std_dev": 0.04650342639234633,
      "run_count": 17,
      "total_questions": 3285,
      "avg_cost_per_task_usd": 0.07316709345509893,
      "avg_duration_per_task_ms": 692682.2697108067,
      "avg_output_tokens_per_task": 32513.825570776255,
      "last_run_timestamp": "2026-08-29T20:15:38.359Z"
    },
    {
      "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": 0.8829965,
      "accuracy_std_dev": 0.03256836877631423,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.12880823737373737,
      "avg_duration_per_task_ms": 39004.51136363636,
      "avg_output_tokens_per_task": 2367.6944444444443,
      "last_run_timestamp": "2026-09-19T08:49:32.428Z"
    },
    {
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Phala",
      "accuracy": 0.8827861249999999,
      "accuracy_std_dev": 0.010063667341450408,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.01960655429292929,
      "avg_duration_per_task_ms": 366878.014520202,
      "avg_output_tokens_per_task": 24088.609848484848,
      "last_run_timestamp": "2026-09-19T08:54:33.262Z"
    },
    {
      "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": 0.8825515716096324,
      "accuracy_std_dev": 0.020268991676360943,
      "run_count": 12,
      "total_questions": 2367,
      "avg_cost_per_task_usd": 0.1375574414871145,
      "avg_duration_per_task_ms": 860088.601182932,
      "avg_output_tokens_per_task": 38561.9945078158,
      "last_run_timestamp": "2026-08-29T18:18:56.485Z"
    },
    {
      "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": 0.8820443201861292,
      "accuracy_std_dev": 0.010098014499560479,
      "run_count": 24,
      "total_questions": 4513,
      "avg_cost_per_task_usd": 0.13465608021271883,
      "avg_duration_per_task_ms": 61985.66430312431,
      "avg_output_tokens_per_task": 4903.735430977177,
      "last_run_timestamp": "2026-09-19T08:56:11.770Z"
    },
    {
      "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": 0.8815657186311787,
      "accuracy_std_dev": 0.013649359425343822,
      "run_count": 12,
      "total_questions": 2367,
      "avg_cost_per_task_usd": 0.10104644275454162,
      "avg_duration_per_task_ms": 571611.1580059147,
      "avg_output_tokens_per_task": 26348.45078158006,
      "last_run_timestamp": "2026-09-19T09:12:25.011Z"
    },
    {
      "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": 0.8813909999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 163,
      "avg_cost_per_task_usd": 0.009804907975460122,
      "avg_duration_per_task_ms": 1011753.5766871165,
      "avg_output_tokens_per_task": 19559.061349693253,
      "last_run_timestamp": "2026-09-14T09:44:57.474Z"
    },
    {
      "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": "Modal",
      "accuracy": 0.881313,
      "accuracy_std_dev": 0.007576000000000027,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.027364550505050506,
      "avg_duration_per_task_ms": 154677.17424242425,
      "avg_output_tokens_per_task": 22608.323232323233,
      "last_run_timestamp": "2026-09-19T09:33:09.301Z"
    },
    {
      "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": 0.881313,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.029419752525252522,
      "avg_duration_per_task_ms": 107618.89141414141,
      "avg_output_tokens_per_task": 24385.08585858586,
      "last_run_timestamp": "2026-09-19T08:52:26.935Z"
    },
    {
      "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": 0.8812490225563911,
      "accuracy_std_dev": 0.01232224163211534,
      "run_count": 26,
      "total_questions": 5054,
      "avg_cost_per_task_usd": 0.17062608369608234,
      "avg_duration_per_task_ms": 343916.4879303522,
      "avg_output_tokens_per_task": 27421.04273842501,
      "last_run_timestamp": "2026-09-19T09:01:57.132Z"
    },
    {
      "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": 0.8805556673913044,
      "accuracy_std_dev": 0.0379554198818583,
      "run_count": 7,
      "total_questions": 1380,
      "avg_cost_per_task_usd": 0.13285437536231884,
      "avg_duration_per_task_ms": 766731.7239130435,
      "avg_output_tokens_per_task": 33250.103623188406,
      "last_run_timestamp": "2026-09-19T09:21:26.846Z"
    },
    {
      "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": "StreamLake",
      "accuracy": 0.880471125,
      "accuracy_std_dev": 0.005389907917522819,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.005834660353535354,
      "avg_duration_per_task_ms": 362869.7544191919,
      "avg_output_tokens_per_task": 28491.27462121212,
      "last_run_timestamp": "2026-09-19T08:50:07.571Z"
    },
    {
      "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": 0.8802532590673576,
      "accuracy_std_dev": 0.005991714074916775,
      "run_count": 3,
      "total_questions": 579,
      "avg_cost_per_task_usd": 0.14682200690846287,
      "avg_duration_per_task_ms": 338397.36614853196,
      "avg_output_tokens_per_task": 33348.9585492228,
      "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-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Decart",
      "accuracy": 0.8798644568527919,
      "accuracy_std_dev": 0.004527248931623837,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.0029938477157360403,
      "avg_duration_per_task_ms": 316818.90016920475,
      "avg_output_tokens_per_task": 18321.88663282572,
      "last_run_timestamp": "2026-08-22T10:20:26.941Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Inceptron",
      "accuracy": 0.879729960954447,
      "accuracy_std_dev": 0.011206573386241868,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.004548976861894432,
      "avg_duration_per_task_ms": 362426.65871294285,
      "avg_output_tokens_per_task": 18848.88069414317,
      "last_run_timestamp": "2026-09-19T11:00:11.912Z"
    },
    {
      "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": 0.8796295,
      "accuracy_std_dev": 0.006574181641086607,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.015792595959595958,
      "avg_duration_per_task_ms": 304804.172979798,
      "avg_output_tokens_per_task": 24118.43686868687,
      "last_run_timestamp": "2026-09-19T08:49:45.204Z"
    },
    {
      "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": 0.8796293333333333,
      "accuracy_std_dev": 0.013874043218262744,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.18146528114478117,
      "avg_duration_per_task_ms": 307037.70454545453,
      "avg_output_tokens_per_task": 45544.973905723906,
      "last_run_timestamp": "2026-09-19T08:48:01.358Z"
    },
    {
      "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": 0.8792085,
      "accuracy_std_dev": 0.026919662038183172,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.14250958080808082,
      "avg_duration_per_task_ms": 42415.96590909091,
      "avg_output_tokens_per_task": 2641.520202020202,
      "last_run_timestamp": "2026-09-19T08:47:22.067Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "WandB",
      "accuracy": 0.8791716692015209,
      "accuracy_std_dev": 0.009087349844013949,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.05804701140684411,
      "avg_duration_per_task_ms": 798447.8948035488,
      "avg_output_tokens_per_task": 16682.354879594423,
      "last_run_timestamp": "2026-07-11T08:42:46.205Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Fireworks",
      "accuracy": 0.8790546552363299,
      "accuracy_std_dev": 0.030844451113161732,
      "run_count": 6,
      "total_questions": 1079,
      "avg_cost_per_task_usd": 0.1346450278035218,
      "avg_duration_per_task_ms": 284974.0537534754,
      "avg_output_tokens_per_task": 31126.302131603337,
      "last_run_timestamp": "2026-09-19T09:14:09.704Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.8788689803571428,
      "accuracy_std_dev": 0.011915461634880886,
      "run_count": 17,
      "total_questions": 3360,
      "avg_cost_per_task_usd": 0.0034551410714285715,
      "avg_duration_per_task_ms": 175260.2726190476,
      "avg_output_tokens_per_task": 17584.270833333332,
      "last_run_timestamp": "2026-08-29T18:22:03.157Z"
    },
    {
      "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": "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-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "WandB",
      "accuracy": 0.8781917204968944,
      "accuracy_std_dev": 0.004843231342812369,
      "run_count": 5,
      "total_questions": 966,
      "avg_cost_per_task_usd": 0.005745680124223602,
      "avg_duration_per_task_ms": 442762.4254658385,
      "avg_output_tokens_per_task": 20256.71635610766,
      "last_run_timestamp": "2026-07-18T08:31:23.912Z"
    },
    {
      "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": 0.8781237785467129,
      "accuracy_std_dev": 0.021132034607833364,
      "run_count": 9,
      "total_questions": 1734,
      "avg_cost_per_task_usd": 0.12840682410611304,
      "avg_duration_per_task_ms": 62804.37370242215,
      "avg_output_tokens_per_task": 4913.2906574394465,
      "last_run_timestamp": "2026-09-19T08:47:57.110Z"
    },
    {
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AtlasCloud",
      "accuracy": 0.877945875,
      "accuracy_std_dev": 0.018131774468026423,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0226289154040404,
      "avg_duration_per_task_ms": 227816.96022727274,
      "avg_output_tokens_per_task": 25365.86047979798,
      "last_run_timestamp": "2026-09-19T09:08:32.908Z"
    },
    {
      "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": 0.8775250000000001,
      "accuracy_std_dev": 0.012731145529762818,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.021682643939393943,
      "avg_duration_per_task_ms": 244444.21085858587,
      "avg_output_tokens_per_task": 24838.689393939392,
      "last_run_timestamp": "2026-09-19T09:00:03.437Z"
    },
    {
      "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": 0.8771930050933787,
      "accuracy_std_dev": 0.007637161296363126,
      "run_count": 3,
      "total_questions": 589,
      "avg_cost_per_task_usd": 0.007654222410865875,
      "avg_duration_per_task_ms": 996245.6400679117,
      "avg_output_tokens_per_task": 30379.964346349745,
      "last_run_timestamp": "2026-09-19T08:50:20.358Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Nebius",
      "accuracy": 0.8771041666666666,
      "accuracy_std_dev": 0.010779529510707913,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0291337297979798,
      "avg_duration_per_task_ms": 93389.48653198653,
      "avg_output_tokens_per_task": 8046.875420875421,
      "last_run_timestamp": "2026-05-23T16:38:19.458Z"
    },
    {
      "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-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Novita",
      "accuracy": 0.8770844662731874,
      "accuracy_std_dev": 0.007890082234467962,
      "run_count": 18,
      "total_questions": 3558,
      "avg_cost_per_task_usd": 0.005329554525014053,
      "avg_duration_per_task_ms": 182892.29510961214,
      "avg_output_tokens_per_task": 18839.3898257448,
      "last_run_timestamp": "2026-08-29T19:12:25.584Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "BaseTen",
      "accuracy": 0.8766337242567858,
      "accuracy_std_dev": 0.03920192550002771,
      "run_count": 12,
      "total_questions": 2321,
      "avg_cost_per_task_usd": 0.0058388018095648425,
      "avg_duration_per_task_ms": 129752.1714778113,
      "avg_output_tokens_per_task": 22107.688496337785,
      "last_run_timestamp": "2026-09-19T08:56:52.111Z"
    },
    {
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8762625,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.032276419191919196,
      "avg_duration_per_task_ms": 263049.4646464646,
      "avg_output_tokens_per_task": 26729.401515151516,
      "last_run_timestamp": "2026-09-19T08:59:52.508Z"
    },
    {
      "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": 0.8761783907455013,
      "accuracy_std_dev": 0.007822499999999954,
      "run_count": 2,
      "total_questions": 389,
      "avg_cost_per_task_usd": 0.048952943444730085,
      "avg_duration_per_task_ms": 221005.3881748072,
      "avg_output_tokens_per_task": 16160.848329048844,
      "last_run_timestamp": "2026-09-19T08:57:27.117Z"
    },
    {
      "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": "Makora",
      "accuracy": 0.876137452982811,
      "accuracy_std_dev": 0.011389182123401155,
      "run_count": 5,
      "total_questions": 989,
      "avg_cost_per_task_usd": 0.004450751263902932,
      "avg_duration_per_task_ms": 154446.49140546005,
      "avg_output_tokens_per_task": 22487.894843276037,
      "last_run_timestamp": "2026-09-19T09:48:52.067Z"
    },
    {
      "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": 0.875421,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.022698737373737374,
      "avg_duration_per_task_ms": 167536.68686868687,
      "avg_output_tokens_per_task": 25022.651515151516,
      "last_run_timestamp": "2026-09-16T10:28:37.773Z"
    },
    {
      "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": 0.8751755374894692,
      "accuracy_std_dev": 0.015517134880153898,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.08454973546756529,
      "avg_duration_per_task_ms": 186945.5796124684,
      "avg_output_tokens_per_task": 21577.509688289807,
      "last_run_timestamp": "2026-09-19T09:33:49.568Z"
    },
    {
      "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": 0.875,
      "accuracy_std_dev": 0.012205000000000021,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.015933075757575757,
      "avg_duration_per_task_ms": 228129.89141414143,
      "avg_output_tokens_per_task": 23888.01767676768,
      "last_run_timestamp": "2026-09-19T10:55:25.751Z"
    },
    {
      "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": 0.875,
      "accuracy_std_dev": 0.004187691070267732,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07034678914141415,
      "avg_duration_per_task_ms": 185029.36994949495,
      "avg_output_tokens_per_task": 11452.73106060606,
      "last_run_timestamp": "2026-09-19T09:06:46.387Z"
    },
    {
      "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": 0.8745582826855123,
      "accuracy_std_dev": 0.047501460387912024,
      "run_count": 8,
      "total_questions": 1415,
      "avg_cost_per_task_usd": 0.07384584240282685,
      "avg_duration_per_task_ms": 519959.2473498233,
      "avg_output_tokens_per_task": 21893.836042402825,
      "last_run_timestamp": "2026-08-22T09:51:24.721Z"
    },
    {
      "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": 0.8744172980132449,
      "accuracy_std_dev": 0.016480704550770754,
      "run_count": 7,
      "total_questions": 1359,
      "avg_cost_per_task_usd": 0.005027246504782929,
      "avg_duration_per_task_ms": 379290.6011773363,
      "avg_output_tokens_per_task": 17886.896983075792,
      "last_run_timestamp": "2026-07-18T08:30:06.336Z"
    },
    {
      "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": 0.8743103404255319,
      "accuracy_std_dev": 0.09209312024616473,
      "run_count": 13,
      "total_questions": 2538,
      "avg_cost_per_task_usd": 0.02429028802206462,
      "avg_duration_per_task_ms": 194909.10914105596,
      "avg_output_tokens_per_task": 20523.450748620962,
      "last_run_timestamp": "2026-09-19T16:05:49.133Z"
    },
    {
      "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": 0.8740457236641223,
      "accuracy_std_dev": 0.03655758641704891,
      "run_count": 10,
      "total_questions": 1965,
      "avg_cost_per_task_usd": 0.12762344173027992,
      "avg_duration_per_task_ms": 372867.7043256998,
      "avg_output_tokens_per_task": 30124.432061068703,
      "last_run_timestamp": "2026-09-19T08:47:35.016Z"
    },
    {
      "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": 0.8739243333333333,
      "accuracy_std_dev": 0.016541355096981753,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.10989428507295174,
      "avg_duration_per_task_ms": 217232.23681257013,
      "avg_output_tokens_per_task": 31249.776094276094,
      "last_run_timestamp": "2026-09-19T08:49:55.345Z"
    },
    {
      "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": 0.8734939514185776,
      "accuracy_std_dev": 0.033103140956105115,
      "run_count": 13,
      "total_questions": 2573,
      "avg_cost_per_task_usd": 0.15498554061406922,
      "avg_duration_per_task_ms": 438624.14224640495,
      "avg_output_tokens_per_task": 11693.160124368442,
      "last_run_timestamp": "2026-09-19T08:57:31.387Z"
    },
    {
      "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": "Alibaba",
      "accuracy": 0.8728743269428278,
      "accuracy_std_dev": 0.022779113769318488,
      "run_count": 15,
      "total_questions": 2921,
      "avg_cost_per_task_usd": 0.060696329681615875,
      "avg_duration_per_task_ms": 329375.96234166384,
      "avg_output_tokens_per_task": 22546.172543649434,
      "last_run_timestamp": "2026-08-22T17:03:07.750Z"
    },
    {
      "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": 0.8728068947368423,
      "accuracy_std_dev": 0.01632482305148274,
      "run_count": 15,
      "total_questions": 2964,
      "avg_cost_per_task_usd": 0.004530426113360323,
      "avg_duration_per_task_ms": 167318.0539811066,
      "avg_output_tokens_per_task": 16025.914979757084,
      "last_run_timestamp": "2026-08-08T10:06:09.471Z"
    },
    {
      "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": 0.8727017508992805,
      "accuracy_std_dev": 0.010551952378424571,
      "run_count": 17,
      "total_questions": 3336,
      "avg_cost_per_task_usd": 0.08340600749400481,
      "avg_duration_per_task_ms": 697270.5386690648,
      "avg_output_tokens_per_task": 29384.263788968827,
      "last_run_timestamp": "2026-08-29T19:07:02.104Z"
    },
    {
      "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": 0.8724410619469029,
      "accuracy_std_dev": 0.010962697967207792,
      "run_count": 19,
      "total_questions": 3729,
      "avg_cost_per_task_usd": 0.0055334333601501735,
      "avg_duration_per_task_ms": 182246.17109144543,
      "avg_output_tokens_per_task": 19566.488871010995,
      "last_run_timestamp": "2026-08-29T18:42:44.868Z"
    },
    {
      "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": 0.8721201154090044,
      "accuracy_std_dev": 0.010011587092433206,
      "run_count": 8,
      "total_questions": 1577,
      "avg_cost_per_task_usd": 0.007726445783132531,
      "avg_duration_per_task_ms": 299427.08941027266,
      "avg_output_tokens_per_task": 27338.933417882054,
      "last_run_timestamp": "2026-09-19T08:58:27.800Z"
    },
    {
      "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": "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": 0.8720024103343466,
      "accuracy_std_dev": 0.014437391891889622,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.0073382553191489374,
      "avg_duration_per_task_ms": 410218.9604863222,
      "avg_output_tokens_per_task": 21333.504559270517,
      "last_run_timestamp": "2026-09-19T08:51:18.222Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Morph",
      "accuracy": 0.871875,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 160,
      "avg_cost_per_task_usd": 0.04507655,
      "avg_duration_per_task_ms": 1504372.61875,
      "avg_output_tokens_per_task": 18693.5375,
      "last_run_timestamp": "2026-08-27T15:00:05.074Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Wafer",
      "accuracy": 0.8718390517241379,
      "accuracy_std_dev": 0.07995783917690623,
      "run_count": 10,
      "total_questions": 1740,
      "avg_cost_per_task_usd": 0.009114119540229886,
      "avg_duration_per_task_ms": 138759.1695402299,
      "avg_output_tokens_per_task": 23586.4683908046,
      "last_run_timestamp": "2026-09-20T10:35:58.112Z"
    },
    {
      "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": 0.8718132857142857,
      "accuracy_std_dev": 0.01994854442032793,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.12823083044733047,
      "avg_duration_per_task_ms": 62432.78354978355,
      "avg_output_tokens_per_task": 4914.397546897547,
      "last_run_timestamp": "2026-09-19T08:48:19.751Z"
    },
    {
      "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": 0.8717413166815345,
      "accuracy_std_dev": 0.019041701395297674,
      "run_count": 17,
      "total_questions": 3363,
      "avg_cost_per_task_usd": 0.07283155159084152,
      "avg_duration_per_task_ms": 379368.3589057389,
      "avg_output_tokens_per_task": 21568.800475765685,
      "last_run_timestamp": "2026-08-29T18:39:19.202Z"
    },
    {
      "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": 0.8716215506756757,
      "accuracy_std_dev": 0.01729230466578475,
      "run_count": 18,
      "total_questions": 3552,
      "avg_cost_per_task_usd": 0.004909025900900901,
      "avg_duration_per_task_ms": 310490.77252252254,
      "avg_output_tokens_per_task": 17361.359797297297,
      "last_run_timestamp": "2026-08-29T18:13:33.416Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.871542434599156,
      "accuracy_std_dev": 0.007338207797953819,
      "run_count": 11,
      "total_questions": 2133,
      "avg_cost_per_task_usd": 0.0043919292076887015,
      "avg_duration_per_task_ms": 387373.37552742613,
      "avg_output_tokens_per_task": 23530.04781997187,
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Makora",
      "accuracy": 0.8715012977099236,
      "accuracy_std_dev": 0.0011400000000000299,
      "run_count": 2,
      "total_questions": 393,
      "avg_cost_per_task_usd": 0.04716040458015267,
      "avg_duration_per_task_ms": 197916.213740458,
      "avg_output_tokens_per_task": 39122.49618320611,
      "last_run_timestamp": "2026-09-19T10:12:37.404Z"
    },
    {
      "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": 0.8714924444444443,
      "accuracy_std_dev": 0.016244831438618754,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.007635035353535354,
      "avg_duration_per_task_ms": 200792.30808080808,
      "avg_output_tokens_per_task": 27092.780583613916,
      "last_run_timestamp": "2026-09-19T08:57:19.814Z"
    },
    {
      "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": "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": 0.8710768244274808,
      "accuracy_std_dev": 0.016176058048857262,
      "run_count": 4,
      "total_questions": 786,
      "avg_cost_per_task_usd": 0.009818778625954197,
      "avg_duration_per_task_ms": 366203.1501272265,
      "avg_output_tokens_per_task": 20813.51526717557,
      "last_run_timestamp": "2026-09-19T08:49:15.112Z"
    },
    {
      "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": 0.8710124489402697,
      "accuracy_std_dev": 0.014206989600055803,
      "run_count": 16,
      "total_questions": 3114,
      "avg_cost_per_task_usd": 0.005702225112395633,
      "avg_duration_per_task_ms": 187380.58863198457,
      "avg_output_tokens_per_task": 21072.890494540785,
      "last_run_timestamp": "2026-08-29T19:20:21.472Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8705363341869399,
      "accuracy_std_dev": 0.026338858318477553,
      "run_count": 12,
      "total_questions": 2343,
      "avg_cost_per_task_usd": 0.09737285574050364,
      "avg_duration_per_task_ms": 684757.7733674776,
      "avg_output_tokens_per_task": 25874.87921468203,
      "last_run_timestamp": "2026-09-19T08:51:08.574Z"
    },
    {
      "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-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Ionstream",
      "accuracy": 0.87037,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07184554292929292,
      "avg_duration_per_task_ms": 295285.3813131313,
      "avg_output_tokens_per_task": 23811.56313131313,
      "last_run_timestamp": "2026-09-12T10:35:19.982Z"
    },
    {
      "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": 0.8700592503816793,
      "accuracy_std_dev": 0.011904817070833141,
      "run_count": 10,
      "total_questions": 1965,
      "avg_cost_per_task_usd": 0.03175990890585242,
      "avg_duration_per_task_ms": 495176.59541984735,
      "avg_output_tokens_per_task": 27384.936386768448,
      "last_run_timestamp": "2026-08-29T19:08:19.007Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Alibaba",
      "accuracy": 0.8699310434782609,
      "accuracy_std_dev": 0.01348825804164903,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.02950167588932806,
      "avg_duration_per_task_ms": 201139.55204216074,
      "avg_output_tokens_per_task": 12529.4356609574,
      "last_run_timestamp": "2026-08-29T18:36:08.027Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Phala",
      "accuracy": 0.8698725770862801,
      "accuracy_std_dev": 0.02689058259409786,
      "run_count": 4,
      "total_questions": 707,
      "avg_cost_per_task_usd": 0.01008102828854314,
      "avg_duration_per_task_ms": 338323.4582743989,
      "avg_output_tokens_per_task": 19986.214992927864,
      "last_run_timestamp": "2026-09-19T08:51:05.252Z"
    },
    {
      "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": 0.8698646204481794,
      "accuracy_std_dev": 0.029668031806137727,
      "run_count": 8,
      "total_questions": 1428,
      "avg_cost_per_task_usd": 0.09692743837535013,
      "avg_duration_per_task_ms": 221464.25770308124,
      "avg_output_tokens_per_task": 26428.57492997199,
      "last_run_timestamp": "2026-09-19T08:45:54.842Z"
    },
    {
      "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": "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": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Google",
      "accuracy": 0.8691077500000002,
      "accuracy_std_dev": 0.03719680227099502,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.11411050757575758,
      "avg_duration_per_task_ms": 27732.646464646463,
      "avg_output_tokens_per_task": 2074.2247474747473,
      "last_run_timestamp": "2026-09-19T08:48:16.582Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "GMICloud",
      "accuracy": 0.8679501874999999,
      "accuracy_std_dev": 0.017033291102348476,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.03806877335858586,
      "avg_duration_per_task_ms": 189574.66571969696,
      "avg_output_tokens_per_task": 10526.75946969697,
      "last_run_timestamp": "2026-08-29T19:12:43.653Z"
    },
    {
      "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": 0.8676765999999999,
      "accuracy_std_dev": 0.004832456584388534,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.06976842626262626,
      "avg_duration_per_task_ms": 222247.99696969698,
      "avg_output_tokens_per_task": 11360.827272727272,
      "last_run_timestamp": "2026-09-19T08:56:34.140Z"
    },
    {
      "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": 0.8672080501968503,
      "accuracy_std_dev": 0.10319503473951104,
      "run_count": 12,
      "total_questions": 2032,
      "avg_cost_per_task_usd": 0.019514511811023625,
      "avg_duration_per_task_ms": 272320.6092519685,
      "avg_output_tokens_per_task": 16198,
      "last_run_timestamp": "2026-09-19T16:07:15.860Z"
    },
    {
      "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": 0.86700325,
      "accuracy_std_dev": 0.011963515314801908,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07822817297979798,
      "avg_duration_per_task_ms": 107799.15151515152,
      "avg_output_tokens_per_task": 12625.300505050505,
      "last_run_timestamp": "2026-09-19T08:50:00.917Z"
    },
    {
      "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": 0.8669794312026003,
      "accuracy_std_dev": 0.015914615486612767,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.06791193391115927,
      "avg_duration_per_task_ms": 196363.11339833876,
      "avg_output_tokens_per_task": 18755.92163235825,
      "last_run_timestamp": "2026-08-29T16:55:12.673Z"
    },
    {
      "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": 0.866898125,
      "accuracy_std_dev": 0.008034667532597418,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.006507090909090909,
      "avg_duration_per_task_ms": 263155.34027777775,
      "avg_output_tokens_per_task": 18228.347222222223,
      "last_run_timestamp": "2026-09-19T08:55:40.325Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Moonshot AI",
      "accuracy": 0.8665198811881186,
      "accuracy_std_dev": 0.012337540068582453,
      "run_count": 14,
      "total_questions": 2727,
      "avg_cost_per_task_usd": 0.14467475100843422,
      "avg_duration_per_task_ms": 785535.8151815182,
      "avg_output_tokens_per_task": 36269.02383571691,
      "last_run_timestamp": "2026-08-29T18:24:25.224Z"
    },
    {
      "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": 0.8664904348591549,
      "accuracy_std_dev": 0.01831953501771554,
      "run_count": 7,
      "total_questions": 1136,
      "avg_cost_per_task_usd": 0.0762857491197183,
      "avg_duration_per_task_ms": 481292.88996478874,
      "avg_output_tokens_per_task": 19103.125880281692,
      "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-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Moonshot AI",
      "accuracy": 0.8663463077140436,
      "accuracy_std_dev": 0.022216806823277722,
      "run_count": 18,
      "total_questions": 3539,
      "avg_cost_per_task_usd": 0.08553019892625037,
      "avg_duration_per_task_ms": 515037.7468211359,
      "avg_output_tokens_per_task": 28564.16643119525,
      "last_run_timestamp": "2026-08-29T18:25:54.650Z"
    },
    {
      "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": 0.8663298,
      "accuracy_std_dev": 0.011398071405286056,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.06935939393939393,
      "avg_duration_per_task_ms": 219280.1888888889,
      "avg_output_tokens_per_task": 11290.334343434344,
      "last_run_timestamp": "2026-09-19T09:01:59.104Z"
    },
    {
      "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": "Wafer",
      "accuracy": 0.8661615,
      "accuracy_std_dev": 0.012626499999999985,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0186262601010101,
      "avg_duration_per_task_ms": 28912.50505050505,
      "avg_output_tokens_per_task": 7135.919191919192,
      "last_run_timestamp": "2026-06-13T16:14:13.401Z"
    },
    {
      "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": 0.8660516956521741,
      "accuracy_std_dev": 0.012674417624678604,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.04402415107597715,
      "avg_duration_per_task_ms": 197822.74681598594,
      "avg_output_tokens_per_task": 12519.180720245937,
      "last_run_timestamp": "2026-08-29T18:39:14.500Z"
    },
    {
      "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": 0.8657404999999998,
      "accuracy_std_dev": 0.014549128315813283,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07900070454545453,
      "avg_duration_per_task_ms": 194604.26641414143,
      "avg_output_tokens_per_task": 12749.18813131313,
      "last_run_timestamp": "2026-09-19T08:51:22.058Z"
    },
    {
      "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": 0.8657200967741937,
      "accuracy_std_dev": 0.019433673652366045,
      "run_count": 30,
      "total_questions": 5177,
      "avg_cost_per_task_usd": 0.14950793490438477,
      "avg_duration_per_task_ms": 230585.83755070504,
      "avg_output_tokens_per_task": 27022.21209194514,
      "last_run_timestamp": "2026-09-19T09:13:15.177Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "GMICloud",
      "accuracy": 0.8649832,
      "accuracy_std_dev": 0.04797704545050687,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.10750342222222221,
      "avg_duration_per_task_ms": 531124.404040404,
      "avg_output_tokens_per_task": 27624.18686868687,
      "last_run_timestamp": "2026-09-19T08:57:20.148Z"
    },
    {
      "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": "Fireworks",
      "accuracy": 0.8646244400527008,
      "accuracy_std_dev": 0.0070903829233687965,
      "run_count": 8,
      "total_questions": 1518,
      "avg_cost_per_task_usd": 0.0050142358366271405,
      "avg_duration_per_task_ms": 140856.23188405798,
      "avg_output_tokens_per_task": 17854.836627140976,
      "last_run_timestamp": "2026-08-08T09:25:45.716Z"
    },
    {
      "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": 0.8643639115417018,
      "accuracy_std_dev": 0.014757024377299852,
      "run_count": 18,
      "total_questions": 3561,
      "avg_cost_per_task_usd": 0.04002094074698118,
      "avg_duration_per_task_ms": 443712.06318449875,
      "avg_output_tokens_per_task": 21300.27857343443,
      "last_run_timestamp": "2026-08-29T19:37:04.387Z"
    },
    {
      "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": "AkashML",
      "accuracy": 0.8643577142857142,
      "accuracy_std_dev": 0.0204579248543673,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.07737239971139971,
      "avg_duration_per_task_ms": 285766.67604617606,
      "avg_output_tokens_per_task": 19368.568542568544,
      "last_run_timestamp": "2026-06-13T09:38:25.896Z"
    },
    {
      "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": 0.8642561170313988,
      "accuracy_std_dev": 0.018479984718270198,
      "run_count": 16,
      "total_questions": 3153,
      "avg_cost_per_task_usd": 0.06389940437678403,
      "avg_duration_per_task_ms": 380883.4934982556,
      "avg_output_tokens_per_task": 19730.762765620046,
      "last_run_timestamp": "2026-08-29T19:18:11.839Z"
    },
    {
      "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": 0.8641172857142856,
      "accuracy_std_dev": 0.03201454919570288,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.10434612698412697,
      "avg_duration_per_task_ms": 510444.2525252525,
      "avg_output_tokens_per_task": 34101.318903318905,
      "last_run_timestamp": "2026-09-19T08:48:26.593Z"
    },
    {
      "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": 0.8640852666666665,
      "accuracy_std_dev": 0.013018156579519569,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.003674580134680135,
      "avg_duration_per_task_ms": 299779.9195286195,
      "avg_output_tokens_per_task": 17340.830976430978,
      "last_run_timestamp": "2026-08-22T17:16:03.168Z"
    },
    {
      "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": 0.8638886961538461,
      "accuracy_std_dev": 0.011642989955736238,
      "run_count": 8,
      "total_questions": 1560,
      "avg_cost_per_task_usd": 0.10131371474358974,
      "avg_duration_per_task_ms": 295529.557051282,
      "avg_output_tokens_per_task": 25493.882692307692,
      "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": "Baidu",
      "accuracy": 0.8638342436974789,
      "accuracy_std_dev": 0.020982236666300873,
      "run_count": 11,
      "total_questions": 2142,
      "avg_cost_per_task_usd": 0.12425809803921568,
      "avg_duration_per_task_ms": 840368.0368814192,
      "avg_output_tokens_per_task": 43681.42903828198,
      "last_run_timestamp": "2026-08-29T19:01:27.418Z"
    },
    {
      "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": 0.8636361428571427,
      "accuracy_std_dev": 0.00964989106420482,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.8530851695526697,
      "avg_duration_per_task_ms": 482874.69696969696,
      "avg_output_tokens_per_task": 33932.781385281385,
      "last_run_timestamp": "2026-09-19T08:55:16.214Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Nebius",
      "accuracy": 0.863636,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0938164393939394,
      "avg_duration_per_task_ms": 667365.3939393939,
      "avg_output_tokens_per_task": 23446.50505050505,
      "last_run_timestamp": "2026-09-10T01:44:54.935Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8632135551128817,
      "accuracy_std_dev": 0.07509086757181263,
      "run_count": 4,
      "total_questions": 753,
      "avg_cost_per_task_usd": 0.12150484462151397,
      "avg_duration_per_task_ms": 758279.0185922974,
      "avg_output_tokens_per_task": 31021.026560424965,
      "last_run_timestamp": "2026-09-19T08:59:57.944Z"
    },
    {
      "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": 0.8631552857142857,
      "accuracy_std_dev": 0.0099276693094788,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.8424338816738816,
      "avg_duration_per_task_ms": 415013.15367965365,
      "avg_output_tokens_per_task": 33506.266955266954,
      "last_run_timestamp": "2026-09-19T08:48:50.622Z"
    },
    {
      "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": 0.8630124096311478,
      "accuracy_std_dev": 0.020757306507348198,
      "run_count": 25,
      "total_questions": 4880,
      "avg_cost_per_task_usd": 0.24381632479508195,
      "avg_duration_per_task_ms": 339721.7473360656,
      "avg_output_tokens_per_task": 44446.558196721315,
      "last_run_timestamp": "2026-09-19T08:58:34.250Z"
    },
    {
      "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": 0.8626651566632756,
      "accuracy_std_dev": 0.015616368164205143,
      "run_count": 5,
      "total_questions": 983,
      "avg_cost_per_task_usd": 0.006657713123092574,
      "avg_duration_per_task_ms": 243535.13428280773,
      "avg_output_tokens_per_task": 13080.57985757884,
      "last_run_timestamp": "2026-09-19T09:17:42.829Z"
    },
    {
      "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": 0.862393223076923,
      "accuracy_std_dev": 0.009661955971101298,
      "run_count": 4,
      "total_questions": 780,
      "avg_cost_per_task_usd": 0.10223442820512822,
      "avg_duration_per_task_ms": 195837.56794871794,
      "avg_output_tokens_per_task": 26387.1858974359,
      "last_run_timestamp": "2026-09-19T08:47:19.823Z"
    },
    {
      "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": 0.8623737499999999,
      "accuracy_std_dev": 0.003834422340783562,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07836570454545455,
      "avg_duration_per_task_ms": 88417.3308080808,
      "avg_output_tokens_per_task": 12812.738636363636,
      "last_run_timestamp": "2026-09-19T08:48:01.553Z"
    },
    {
      "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": 0.8623736666666667,
      "accuracy_std_dev": 0.019269903780824178,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.08775729166666667,
      "avg_duration_per_task_ms": 407569.07954545453,
      "avg_output_tokens_per_task": 38211.61952861953,
      "last_run_timestamp": "2026-09-19T09:10:50.969Z"
    },
    {
      "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": "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": 0.8619530000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.023946904040404043,
      "avg_duration_per_task_ms": 411455.5404040404,
      "avg_output_tokens_per_task": 47654.88383838384,
      "last_run_timestamp": "2026-09-19T11:21:43.047Z"
    },
    {
      "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": "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-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Baidu",
      "accuracy": 0.8613799792531119,
      "accuracy_std_dev": 0.01086146499258382,
      "run_count": 11,
      "total_questions": 2169,
      "avg_cost_per_task_usd": 0.0035420834485938215,
      "avg_duration_per_task_ms": 187394.91378515444,
      "avg_output_tokens_per_task": 20972.25633932688,
      "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-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.8612793999999999,
      "accuracy_std_dev": 0.013988220381449522,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.006820060606060606,
      "avg_duration_per_task_ms": 354799.9131313131,
      "avg_output_tokens_per_task": 22824.536363636365,
      "last_run_timestamp": "2026-09-19T09:05:29.545Z"
    },
    {
      "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": "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": 0.8611109999999997,
      "accuracy_std_dev": 0.008672208830511384,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.8817781500721501,
      "avg_duration_per_task_ms": 416732.5119047619,
      "avg_output_tokens_per_task": 33399.694083694085,
      "last_run_timestamp": "2026-09-19T09:02:03.655Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Together",
      "accuracy": 0.8609416158536587,
      "accuracy_std_dev": 0.007965914224996389,
      "run_count": 5,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.016509900406504064,
      "avg_duration_per_task_ms": 333963.73475609755,
      "avg_output_tokens_per_task": 32804.009146341465,
      "last_run_timestamp": "2026-09-19T08:48:55.184Z"
    },
    {
      "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": 0.8608576565349543,
      "accuracy_std_dev": 0.016168604733866183,
      "run_count": 15,
      "total_questions": 2961,
      "avg_cost_per_task_usd": 0.015263644376899694,
      "avg_duration_per_task_ms": 308595.9824383654,
      "avg_output_tokens_per_task": 17709.83485309017,
      "last_run_timestamp": "2026-08-08T09:24:32.269Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Fireworks",
      "accuracy": 0.860810656081081,
      "accuracy_std_dev": 0.017596670139546283,
      "run_count": 8,
      "total_questions": 1480,
      "avg_cost_per_task_usd": 0.0684703912162162,
      "avg_duration_per_task_ms": 339840.025,
      "avg_output_tokens_per_task": 20257.95,
      "last_run_timestamp": "2026-07-18T08:30:34.727Z"
    },
    {
      "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": 0.86079326201373,
      "accuracy_std_dev": 0.028859126736614872,
      "run_count": 10,
      "total_questions": 1748,
      "avg_cost_per_task_usd": 0.010031786613272312,
      "avg_duration_per_task_ms": 339107.3318077803,
      "avg_output_tokens_per_task": 18845.827231121282,
      "last_run_timestamp": "2026-09-19T08:55:09.641Z"
    },
    {
      "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": 0.8606902500000001,
      "accuracy_std_dev": 0.01603212873225199,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.008977529040404041,
      "avg_duration_per_task_ms": 694440.691919192,
      "avg_output_tokens_per_task": 30908.54419191919,
      "last_run_timestamp": "2026-09-19T08:48:13.156Z"
    },
    {
      "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": 0.8605240633981405,
      "accuracy_std_dev": 0.05128936013142842,
      "run_count": 18,
      "total_questions": 3549,
      "avg_cost_per_task_usd": 0.00447025641025641,
      "avg_duration_per_task_ms": 218107.38489715412,
      "avg_output_tokens_per_task": 16009.49056072133,
      "last_run_timestamp": "2026-08-29T18:30:11.054Z"
    },
    {
      "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": 0.8604095833333332,
      "accuracy_std_dev": 0.019983922335460633,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.11414123358585856,
      "avg_duration_per_task_ms": 818760.5311447812,
      "avg_output_tokens_per_task": 36533.946127946125,
      "last_run_timestamp": "2026-08-29T18:43:04.938Z"
    },
    {
      "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": 0.8603764494556766,
      "accuracy_std_dev": 0.016565195248170166,
      "run_count": 10,
      "total_questions": 1929,
      "avg_cost_per_task_usd": 0.10031787610160703,
      "avg_duration_per_task_ms": 637584.9020217729,
      "avg_output_tokens_per_task": 34282.6588906169,
      "last_run_timestamp": "2026-09-19T08:55:10.654Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Inceptron",
      "accuracy": 0.8601241573604061,
      "accuracy_std_dev": 0.012372890590139204,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.10654743824027074,
      "avg_duration_per_task_ms": 370641.0829103215,
      "avg_output_tokens_per_task": 25074.472081218275,
      "last_run_timestamp": "2026-09-19T08:51:20.373Z"
    },
    {
      "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": "Novita",
      "accuracy": 0.8600690598563449,
      "accuracy_std_dev": 0.020027284386563163,
      "run_count": 19,
      "total_questions": 3759,
      "avg_cost_per_task_usd": 0.05487559829741953,
      "avg_duration_per_task_ms": 344231.89066241024,
      "avg_output_tokens_per_task": 20069.8159084863,
      "last_run_timestamp": "2026-08-29T19:21:48.221Z"
    },
    {
      "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": 0.86000405028463,
      "accuracy_std_dev": 0.023944982277086784,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.04536207748260594,
      "avg_duration_per_task_ms": 361772.752055661,
      "avg_output_tokens_per_task": 22068.88678051866,
      "last_run_timestamp": "2026-08-29T19:11:30.664Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Novita",
      "accuracy": 0.8594949584487533,
      "accuracy_std_dev": 0.016664788929539907,
      "run_count": 11,
      "total_questions": 2166,
      "avg_cost_per_task_usd": 0.14950503924284395,
      "avg_duration_per_task_ms": 850309.7391505078,
      "avg_output_tokens_per_task": 43660.16805170822,
      "last_run_timestamp": "2026-08-29T20:16:17.828Z"
    },
    {
      "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": 0.8591069047619047,
      "accuracy_std_dev": 0.034057120266518005,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.03376736123136124,
      "avg_duration_per_task_ms": 210716.60485810487,
      "avg_output_tokens_per_task": 11637.827561327562,
      "last_run_timestamp": "2026-08-15T15:55:02.752Z"
    },
    {
      "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": 0.8589856042216361,
      "accuracy_std_dev": 0.022745485097766183,
      "run_count": 23,
      "total_questions": 4548,
      "avg_cost_per_task_usd": 0.03739281288478453,
      "avg_duration_per_task_ms": 298793.5393579595,
      "avg_output_tokens_per_task": 12826.183597185576,
      "last_run_timestamp": "2026-08-29T19:19:13.521Z"
    },
    {
      "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": 0.858586,
      "accuracy_std_dev": 0.020201999999999998,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.025399686868686867,
      "avg_duration_per_task_ms": 99978.351010101,
      "avg_output_tokens_per_task": 16731.46212121212,
      "last_run_timestamp": "2026-09-19T09:57:08.766Z"
    },
    {
      "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": 0.8585858571428572,
      "accuracy_std_dev": 0.008098936198725138,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.8266230122655123,
      "avg_duration_per_task_ms": 471406.1378066378,
      "avg_output_tokens_per_task": 32893.57142857143,
      "last_run_timestamp": "2026-09-19T08:54:22.602Z"
    },
    {
      "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": 0.8583117441860466,
      "accuracy_std_dev": 0.020029378052937655,
      "run_count": 4,
      "total_questions": 774,
      "avg_cost_per_task_usd": 0.0894643488372093,
      "avg_duration_per_task_ms": 522454.9418604651,
      "avg_output_tokens_per_task": 23617.786821705427,
      "last_run_timestamp": "2026-09-19T09:16:39.168Z"
    },
    {
      "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": 0.8581650000000001,
      "accuracy_std_dev": 0.016424948356083192,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.10795924494949495,
      "avg_duration_per_task_ms": 770514.4431818182,
      "avg_output_tokens_per_task": 28271.43686868687,
      "last_run_timestamp": "2026-09-19T08:46:12.784Z"
    },
    {
      "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": 0.8580918152173914,
      "accuracy_std_dev": 0.03107799783834439,
      "run_count": 17,
      "total_questions": 3312,
      "avg_cost_per_task_usd": 0.06328980283816425,
      "avg_duration_per_task_ms": 802747.4701086957,
      "avg_output_tokens_per_task": 28225.76388888889,
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "StreamLake",
      "accuracy": 0.8576652329113923,
      "accuracy_std_dev": 0.03114981165807159,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.06335243122362869,
      "avg_duration_per_task_ms": 565982.5426160338,
      "avg_output_tokens_per_task": 40503.44135021097,
      "last_run_timestamp": "2026-09-19T08:47:25.115Z"
    },
    {
      "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": 0.8575758,
      "accuracy_std_dev": 0.00835660171122207,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0665558909090909,
      "avg_duration_per_task_ms": 212021.95252525253,
      "avg_output_tokens_per_task": 10824.585858585859,
      "last_run_timestamp": "2026-09-19T08:47:49.557Z"
    },
    {
      "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": 0.8575267943548386,
      "accuracy_std_dev": 0.0136809305307607,
      "run_count": 4,
      "total_questions": 744,
      "avg_cost_per_task_usd": 0.18561700403225806,
      "avg_duration_per_task_ms": 513052.8830645161,
      "avg_output_tokens_per_task": 55552.403225806454,
      "last_run_timestamp": "2026-07-18T08:26:48.377Z"
    },
    {
      "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": "BaseTen",
      "accuracy": 0.8574163891625615,
      "accuracy_std_dev": 0.01729438534295256,
      "run_count": 13,
      "total_questions": 2436,
      "avg_cost_per_task_usd": 0.14581712520525447,
      "avg_duration_per_task_ms": 191299.41174055828,
      "avg_output_tokens_per_task": 26193.477422003285,
      "last_run_timestamp": "2026-09-19T09:04:15.041Z"
    },
    {
      "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": 0.8573686660341556,
      "accuracy_std_dev": 0.03152328839896587,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.09650723655913979,
      "avg_duration_per_task_ms": 712555.4617330803,
      "avg_output_tokens_per_task": 37848.16002530044,
      "last_run_timestamp": "2026-09-12T08:54:06.518Z"
    },
    {
      "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": 0.8570308061104582,
      "accuracy_std_dev": 0.02562810322823216,
      "run_count": 13,
      "total_questions": 2553,
      "avg_cost_per_task_usd": 0.07877737015276144,
      "avg_duration_per_task_ms": 570287.8503721112,
      "avg_output_tokens_per_task": 34464.188405797104,
      "last_run_timestamp": "2026-09-19T08:48:29.546Z"
    },
    {
      "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": 0.856844813993174,
      "accuracy_std_dev": 0.04300853558120186,
      "run_count": 9,
      "total_questions": 1758,
      "avg_cost_per_task_usd": 0.08689430546075085,
      "avg_duration_per_task_ms": 500988.61945392494,
      "avg_output_tokens_per_task": 22774.62059158134,
      "last_run_timestamp": "2026-08-29T11:31:24.094Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8568008017241379,
      "accuracy_std_dev": 0.060206459555848316,
      "run_count": 4,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.011371310344827586,
      "avg_duration_per_task_ms": 603206.4712643678,
      "avg_output_tokens_per_task": 22523.2183908046,
      "last_run_timestamp": "2026-09-19T08:56:14.441Z"
    },
    {
      "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": 0.8567280117827869,
      "accuracy_std_dev": 0.019208364564689006,
      "run_count": 10,
      "total_questions": 1952,
      "avg_cost_per_task_usd": 0.009705317622950821,
      "avg_duration_per_task_ms": 263176.2423155738,
      "avg_output_tokens_per_task": 17464.506147540982,
      "last_run_timestamp": "2026-08-29T18:07:13.805Z"
    },
    {
      "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": 0.8564651992882563,
      "accuracy_std_dev": 0.004182357947378454,
      "run_count": 3,
      "total_questions": 562,
      "avg_cost_per_task_usd": 0.137316371886121,
      "avg_duration_per_task_ms": 228809.79359430604,
      "avg_output_tokens_per_task": 31529.871886120996,
      "last_run_timestamp": "2026-09-19T08:49:57.223Z"
    },
    {
      "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": 0.856341,
      "accuracy_std_dev": 0.010617665907973035,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.8414909006734006,
      "avg_duration_per_task_ms": 451224.2822671156,
      "avg_output_tokens_per_task": 33486.77272727273,
      "last_run_timestamp": "2026-09-19T08:45:30.738Z"
    },
    {
      "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-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Chutes",
      "accuracy": 0.8560606556149731,
      "accuracy_std_dev": 0.0169669220414311,
      "run_count": 10,
      "total_questions": 1870,
      "avg_cost_per_task_usd": 0.046046802673796794,
      "avg_duration_per_task_ms": 755872.1957219251,
      "avg_output_tokens_per_task": 23501.898395721924,
      "last_run_timestamp": "2026-07-18T16:22:39.277Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.8555779231547017,
      "accuracy_std_dev": 0.009816647688493245,
      "run_count": 5,
      "total_questions": 989,
      "avg_cost_per_task_usd": 0.016417139534883725,
      "avg_duration_per_task_ms": 493380.05561172904,
      "avg_output_tokens_per_task": 32598.366026289183,
      "last_run_timestamp": "2026-09-19T08:48:55.257Z"
    },
    {
      "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": 0.855499071018968,
      "accuracy_std_dev": 0.058276609033653,
      "run_count": 23,
      "total_questions": 4534,
      "avg_cost_per_task_usd": 0.04388444067048963,
      "avg_duration_per_task_ms": 191885.59814733127,
      "avg_output_tokens_per_task": 12114.957873842082,
      "last_run_timestamp": "2026-08-29T18:30:43.161Z"
    },
    {
      "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": 0.8554747697565143,
      "accuracy_std_dev": 0.03528402563356005,
      "run_count": 12,
      "total_questions": 2341,
      "avg_cost_per_task_usd": 0.16198979666809055,
      "avg_duration_per_task_ms": 293938.29090132425,
      "avg_output_tokens_per_task": 36709.38530542503,
      "last_run_timestamp": "2026-09-19T08:57:22.475Z"
    },
    {
      "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": 0.855459142857143,
      "accuracy_std_dev": 0.01202022619740304,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.7458562337662338,
      "avg_duration_per_task_ms": 493748.36435786437,
      "avg_output_tokens_per_task": 29676.530303030304,
      "last_run_timestamp": "2026-09-19T08:57:48.329Z"
    },
    {
      "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": 0.8552118345784417,
      "accuracy_std_dev": 0.013541336870486602,
      "run_count": 5,
      "total_questions": 937,
      "avg_cost_per_task_usd": 0.009773050160085377,
      "avg_duration_per_task_ms": 36346.31910352188,
      "avg_output_tokens_per_task": 8563.84204909285,
      "last_run_timestamp": "2026-08-01T18:26:19.887Z"
    },
    {
      "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": 0.8549781428571429,
      "accuracy_std_dev": 0.02627276912061813,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.007480496392496394,
      "avg_duration_per_task_ms": 490043.57647907647,
      "avg_output_tokens_per_task": 24203.634920634922,
      "last_run_timestamp": "2026-09-19T08:48:10.197Z"
    },
    {
      "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": 0.8546835012658227,
      "accuracy_std_dev": 0.02399739526567811,
      "run_count": 11,
      "total_questions": 1975,
      "avg_cost_per_task_usd": 0.12927437518987342,
      "avg_duration_per_task_ms": 439949.8698734177,
      "avg_output_tokens_per_task": 41561.01063291139,
      "last_run_timestamp": "2026-08-29T19:17:01.480Z"
    },
    {
      "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": 0.8546576666666666,
      "accuracy_std_dev": 0.005555300251911549,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.12576473232323232,
      "avg_duration_per_task_ms": 209920.99494949495,
      "avg_output_tokens_per_task": 28358.969696969696,
      "last_run_timestamp": "2026-09-19T08:49:27.482Z"
    },
    {
      "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": "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": 0.8543789735234216,
      "accuracy_std_dev": 0.012989646298494788,
      "run_count": 5,
      "total_questions": 982,
      "avg_cost_per_task_usd": 0.0073387342158859475,
      "avg_duration_per_task_ms": 333764.4317718941,
      "avg_output_tokens_per_task": 21754.163951120165,
      "last_run_timestamp": "2026-09-19T08:47:03.155Z"
    },
    {
      "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": 0.8539195888324873,
      "accuracy_std_dev": 0.008792242881591056,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.11491216243654823,
      "avg_duration_per_task_ms": 321574.34010152286,
      "avg_output_tokens_per_task": 29798.65143824027,
      "last_run_timestamp": "2026-09-19T08:46:08.188Z"
    },
    {
      "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": 0.853063631284916,
      "accuracy_std_dev": 0.013358496890952795,
      "run_count": 19,
      "total_questions": 3759,
      "avg_cost_per_task_usd": 0.017711467411545622,
      "avg_duration_per_task_ms": 155839.47140196862,
      "avg_output_tokens_per_task": 11900.042564511838,
      "last_run_timestamp": "2026-08-01T17:57:47.536Z"
    },
    {
      "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": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Azure",
      "accuracy": 0.8525731428571429,
      "accuracy_std_dev": 0.020985589732198676,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.8386075396825398,
      "avg_duration_per_task_ms": 455420.84199134196,
      "avg_output_tokens_per_task": 33379.707070707074,
      "last_run_timestamp": "2026-09-19T08:53:57.572Z"
    },
    {
      "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": 0.8525641846153847,
      "accuracy_std_dev": 0.0069550738853012715,
      "run_count": 4,
      "total_questions": 780,
      "avg_cost_per_task_usd": 0.1463613051282051,
      "avg_duration_per_task_ms": 253560.13461538462,
      "avg_output_tokens_per_task": 27986.41794871795,
      "last_run_timestamp": "2026-09-19T10:21:04.775Z"
    },
    {
      "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": 0.8524682087067861,
      "accuracy_std_dev": 0.02973171903175686,
      "run_count": 12,
      "total_questions": 2343,
      "avg_cost_per_task_usd": 0.02875936875800256,
      "avg_duration_per_task_ms": 363978.1912078532,
      "avg_output_tokens_per_task": 24367.118224498507,
      "last_run_timestamp": "2026-08-29T19:03:08.917Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Phala",
      "accuracy": 0.8520922857142859,
      "accuracy_std_dev": 0.06769645437022032,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.04623776310726311,
      "avg_duration_per_task_ms": 201066.01322751323,
      "avg_output_tokens_per_task": 13148.871332371333,
      "last_run_timestamp": "2026-08-29T20:00:02.637Z"
    },
    {
      "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": 0.85189953626943,
      "accuracy_std_dev": 0.030767744842407633,
      "run_count": 4,
      "total_questions": 772,
      "avg_cost_per_task_usd": 0.137902896373057,
      "avg_duration_per_task_ms": 375133.5362694301,
      "avg_output_tokens_per_task": 31541,
      "last_run_timestamp": "2026-09-19T08:53:57.386Z"
    },
    {
      "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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Novita",
      "accuracy": 0.8517631052631578,
      "accuracy_std_dev": 0.01675925147146067,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.013288521265284426,
      "avg_duration_per_task_ms": 130487.94444444444,
      "avg_output_tokens_per_task": 10986.221956406167,
      "last_run_timestamp": "2026-08-01T17:57:21.740Z"
    },
    {
      "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": 0.8516789626604435,
      "accuracy_std_dev": 0.034038116974527216,
      "run_count": 13,
      "total_questions": 2571,
      "avg_cost_per_task_usd": 0.05071204317386231,
      "avg_duration_per_task_ms": 213874.00194476856,
      "avg_output_tokens_per_task": 21039.041618047453,
      "last_run_timestamp": "2026-09-19T09:37:29.831Z"
    },
    {
      "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": 0.8515202907075875,
      "accuracy_std_dev": 0.012396151064835489,
      "run_count": 6,
      "total_questions": 1173,
      "avg_cost_per_task_usd": 0.12419511594202898,
      "avg_duration_per_task_ms": 467009.2557544757,
      "avg_output_tokens_per_task": 31591.142369991474,
      "last_run_timestamp": "2026-09-19T09:07:57.675Z"
    },
    {
      "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": 0.8513301536491676,
      "accuracy_std_dev": 0.05979411713817116,
      "run_count": 12,
      "total_questions": 2343,
      "avg_cost_per_task_usd": 0.06459188518992744,
      "avg_duration_per_task_ms": 616134.8638497653,
      "avg_output_tokens_per_task": 36840.9726845924,
      "last_run_timestamp": "2026-09-19T09:39:16.626Z"
    },
    {
      "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": 0.8512884334600759,
      "accuracy_std_dev": 0.0436746758115393,
      "run_count": 12,
      "total_questions": 2367,
      "avg_cost_per_task_usd": 0.11391814955640052,
      "avg_duration_per_task_ms": 499191.1106886354,
      "avg_output_tokens_per_task": 32549.244613434726,
      "last_run_timestamp": "2026-09-19T08:49:34.299Z"
    },
    {
      "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": 0.8509483597560974,
      "accuracy_std_dev": 0.02777727136059266,
      "run_count": 5,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.10137330792682926,
      "avg_duration_per_task_ms": 326460.5447154472,
      "avg_output_tokens_per_task": 22810.232723577235,
      "last_run_timestamp": "2026-09-19T08:57:56.121Z"
    },
    {
      "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": 0.8508814908551657,
      "accuracy_std_dev": 0.01345103975540779,
      "run_count": 31,
      "total_questions": 6069,
      "avg_cost_per_task_usd": 0.03916498055692866,
      "avg_duration_per_task_ms": 432081.76124567474,
      "avg_output_tokens_per_task": 22652.797495468774,
      "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": "Fireworks",
      "accuracy": 0.8505806741803278,
      "accuracy_std_dev": 0.015646816333043635,
      "run_count": 5,
      "total_questions": 976,
      "avg_cost_per_task_usd": 0.017172275614754096,
      "avg_duration_per_task_ms": 299428.9518442623,
      "avg_output_tokens_per_task": 34120.667008196724,
      "last_run_timestamp": "2026-09-19T08:49:36.471Z"
    },
    {
      "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": 0.8504337702909648,
      "accuracy_std_dev": 0.025173874720431894,
      "run_count": 10,
      "total_questions": 1959,
      "avg_cost_per_task_usd": 0.05186597600816743,
      "avg_duration_per_task_ms": 525779.0051046453,
      "avg_output_tokens_per_task": 20693.85911179173,
      "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-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Chutes",
      "accuracy": 0.8502667601600985,
      "accuracy_std_dev": 0.030732490705134455,
      "run_count": 17,
      "total_questions": 3248,
      "avg_cost_per_task_usd": 0.10197107050492611,
      "avg_duration_per_task_ms": 818392.5018472907,
      "avg_output_tokens_per_task": 31384.871613300493,
      "last_run_timestamp": "2026-08-29T18:13:11.491Z"
    },
    {
      "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": 0.8501585778210115,
      "accuracy_std_dev": 0.036849849445430474,
      "run_count": 27,
      "total_questions": 4626,
      "avg_cost_per_task_usd": 0.010481716817985303,
      "avg_duration_per_task_ms": 412689.4900562041,
      "avg_output_tokens_per_task": 15437.374621703415,
      "last_run_timestamp": "2026-09-19T08:48:53.177Z"
    },
    {
      "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": 0.8499811111111111,
      "accuracy_std_dev": 0.015580772491506915,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.7504352328843996,
      "avg_duration_per_task_ms": 500085.27890011226,
      "avg_output_tokens_per_task": 29943.05948372615,
      "last_run_timestamp": "2026-09-19T09:04:24.237Z"
    },
    {
      "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": 0.8499715520304567,
      "accuracy_std_dev": 0.014031811236486494,
      "run_count": 12,
      "total_questions": 2364,
      "avg_cost_per_task_usd": 0.027818640862944163,
      "avg_duration_per_task_ms": 497777.6065989848,
      "avg_output_tokens_per_task": 25887.815143824027,
      "last_run_timestamp": "2026-08-29T19:21:17.134Z"
    },
    {
      "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": 0.8497797692307693,
      "accuracy_std_dev": 0.008887516934129454,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.0686196585081585,
      "avg_duration_per_task_ms": 218386.79564879564,
      "avg_output_tokens_per_task": 21351.092074592074,
      "last_run_timestamp": "2026-09-19T09:06:57.641Z"
    },
    {
      "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": 0.8495459182879378,
      "accuracy_std_dev": 0.02771513034399036,
      "run_count": 8,
      "total_questions": 1542,
      "avg_cost_per_task_usd": 0.028309669909208815,
      "avg_duration_per_task_ms": 547916.742542153,
      "avg_output_tokens_per_task": 27038.87094682231,
      "last_run_timestamp": "2026-09-19T08:49:45.084Z"
    },
    {
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Phala",
      "accuracy": 0.8488110578734859,
      "accuracy_std_dev": 0.0374690880393425,
      "run_count": 4,
      "total_questions": 743,
      "avg_cost_per_task_usd": 0.12476096096904442,
      "avg_duration_per_task_ms": 559359.8277254375,
      "avg_output_tokens_per_task": 33151.71467025572,
      "last_run_timestamp": "2026-09-19T08:53:56.961Z"
    },
    {
      "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": 0.8487651666666665,
      "accuracy_std_dev": 0.03381563814971148,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.8327803324915825,
      "avg_duration_per_task_ms": 429212.7272727273,
      "avg_output_tokens_per_task": 33275.729797979795,
      "last_run_timestamp": "2026-09-19T08:47:59.678Z"
    },
    {
      "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": 0.8487252142857141,
      "accuracy_std_dev": 0.017377716902395323,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.7963357799422798,
      "avg_duration_per_task_ms": 477422.2904040404,
      "avg_output_tokens_per_task": 30185.76370851371,
      "last_run_timestamp": "2026-09-19T08:57:19.905Z"
    },
    {
      "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": 0.8486816237288136,
      "accuracy_std_dev": 0.03080583684772884,
      "run_count": 18,
      "total_questions": 3540,
      "avg_cost_per_task_usd": 0.21800511779661022,
      "avg_duration_per_task_ms": 505891.4338983051,
      "avg_output_tokens_per_task": 41719.584463276835,
      "last_run_timestamp": "2026-09-19T08:54:22.411Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "NextBit",
      "accuracy": 0.8484846666666668,
      "accuracy_std_dev": 0.022121414381142606,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.1076985303030303,
      "avg_duration_per_task_ms": 498790.32828282827,
      "avg_output_tokens_per_task": 31614.96127946128,
      "last_run_timestamp": "2026-08-29T19:12:34.610Z"
    },
    {
      "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": 0.8483078331388565,
      "accuracy_std_dev": 0.009457638031564467,
      "run_count": 13,
      "total_questions": 2571,
      "avg_cost_per_task_usd": 0.045122924542979385,
      "avg_duration_per_task_ms": 260125.00777907428,
      "avg_output_tokens_per_task": 41250.62932711007,
      "last_run_timestamp": "2026-09-19T16:06:22.304Z"
    },
    {
      "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": 0.8481442625127681,
      "accuracy_std_dev": 0.03159535831352445,
      "run_count": 15,
      "total_questions": 2937,
      "avg_cost_per_task_usd": 0.07499461457269323,
      "avg_duration_per_task_ms": 825143.8052434457,
      "avg_output_tokens_per_task": 32030.652706843717,
      "last_run_timestamp": "2026-08-29T18:42:20.541Z"
    },
    {
      "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": 0.8481386725888325,
      "accuracy_std_dev": 0.07204977338092783,
      "run_count": 12,
      "total_questions": 2364,
      "avg_cost_per_task_usd": 0.15121076776649744,
      "avg_duration_per_task_ms": 469840.51480541454,
      "avg_output_tokens_per_task": 34451.947969543144,
      "last_run_timestamp": "2026-09-19T08:48:42.422Z"
    },
    {
      "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": 0.8481244738853504,
      "accuracy_std_dev": 0.01045396470643235,
      "run_count": 12,
      "total_questions": 2355,
      "avg_cost_per_task_usd": 0.025721170276008492,
      "avg_duration_per_task_ms": 552477.8963906582,
      "avg_output_tokens_per_task": 26613.83906581741,
      "last_run_timestamp": "2026-09-19T09:19:53.890Z"
    },
    {
      "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": "Together",
      "accuracy": 0.8481239999999999,
      "accuracy_std_dev": 0.019416861332357486,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.0265733189033189,
      "avg_duration_per_task_ms": 50796.17424242424,
      "avg_output_tokens_per_task": 7342.486652236652,
      "last_run_timestamp": "2026-06-20T16:15:59.970Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.848103465060241,
      "accuracy_std_dev": 0.01673605276986749,
      "run_count": 19,
      "total_questions": 3735,
      "avg_cost_per_task_usd": 0.017738112449799195,
      "avg_duration_per_task_ms": 321964.4128514056,
      "avg_output_tokens_per_task": 14661.384471218205,
      "last_run_timestamp": "2026-08-29T18:16:39.840Z"
    },
    {
      "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": 0.8480308582278481,
      "accuracy_std_dev": 0.00841927700239488,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.035698729535864976,
      "avg_duration_per_task_ms": 353629.90632911393,
      "avg_output_tokens_per_task": 36124.08481012658,
      "last_run_timestamp": "2026-09-19T16:18:47.983Z"
    },
    {
      "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": 0.8480290637966805,
      "accuracy_std_dev": 0.06331280368811666,
      "run_count": 10,
      "total_questions": 1928,
      "avg_cost_per_task_usd": 0.05295942946058091,
      "avg_duration_per_task_ms": 591223.8205394191,
      "avg_output_tokens_per_task": 18578.95072614108,
      "last_run_timestamp": "2026-09-12T09:19:57.645Z"
    },
    {
      "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": 0.8480035714285714,
      "accuracy_std_dev": 0.016064940156903722,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.7519250108225108,
      "avg_duration_per_task_ms": 501563.3658008658,
      "avg_output_tokens_per_task": 29978.645743145742,
      "last_run_timestamp": "2026-09-19T08:50:15.420Z"
    },
    {
      "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": 0.8477077692307692,
      "accuracy_std_dev": 0.03663217256656703,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.0030205528360528365,
      "avg_duration_per_task_ms": 496322.17443667445,
      "avg_output_tokens_per_task": 11617.942113442114,
      "last_run_timestamp": "2026-07-18T08:42:46.001Z"
    },
    {
      "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": 0.847642875,
      "accuracy_std_dev": 0.03659437500640468,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.17585300694444442,
      "avg_duration_per_task_ms": 44732.63005050505,
      "avg_output_tokens_per_task": 3313.816919191919,
      "last_run_timestamp": "2026-09-19T08:59:16.818Z"
    },
    {
      "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": 0.8474537731821213,
      "accuracy_std_dev": 0.016586662965662614,
      "run_count": 23,
      "total_questions": 4497,
      "avg_cost_per_task_usd": 0.048555820102290405,
      "avg_duration_per_task_ms": 531371.0297976428,
      "avg_output_tokens_per_task": 24551.97487213698,
      "last_run_timestamp": "2026-08-30T00:19:02.095Z"
    },
    {
      "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": 0.8471111408,
      "accuracy_std_dev": 0.011562823272593973,
      "run_count": 19,
      "total_questions": 3750,
      "avg_cost_per_task_usd": 0.011580457866666668,
      "avg_duration_per_task_ms": 318015.93173333333,
      "avg_output_tokens_per_task": 13008.763733333333,
      "last_run_timestamp": "2026-08-15T14:57:08.773Z"
    },
    {
      "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": 0.8469509168474333,
      "accuracy_std_dev": 0.01813136080720607,
      "run_count": 21,
      "total_questions": 4149,
      "avg_cost_per_task_usd": 0.01589632562063148,
      "avg_duration_per_task_ms": 154447.76644974694,
      "avg_output_tokens_per_task": 13181.300795372379,
      "last_run_timestamp": "2026-08-29T18:12:43.943Z"
    },
    {
      "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-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8466981191037736,
      "accuracy_std_dev": 0.02482932381358783,
      "run_count": 5,
      "total_questions": 848,
      "avg_cost_per_task_usd": 0.009916996462264151,
      "avg_duration_per_task_ms": 317031.18396226416,
      "avg_output_tokens_per_task": 19623.935141509435,
      "last_run_timestamp": "2026-09-19T08:58:01.449Z"
    },
    {
      "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": 0.8465876167512689,
      "accuracy_std_dev": 0.014084137220399971,
      "run_count": 12,
      "total_questions": 2364,
      "avg_cost_per_task_usd": 0.10131879653130285,
      "avg_duration_per_task_ms": 853989.2901861252,
      "avg_output_tokens_per_task": 31653.363790186126,
      "last_run_timestamp": "2026-09-19T17:02:15.853Z"
    },
    {
      "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": 0.8464216922102598,
      "accuracy_std_dev": 0.016722260700560206,
      "run_count": 8,
      "total_questions": 1579,
      "avg_cost_per_task_usd": 0.08943012222925902,
      "avg_duration_per_task_ms": 420229.20645978465,
      "avg_output_tokens_per_task": 22253.619379354022,
      "last_run_timestamp": "2026-09-19T08:55:13.302Z"
    },
    {
      "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": 0.8462402666666666,
      "accuracy_std_dev": 0.020217635616681023,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.15834877643097642,
      "avg_duration_per_task_ms": 154755.59292929294,
      "avg_output_tokens_per_task": 15633.021548821549,
      "last_run_timestamp": "2026-09-19T08:49:28.907Z"
    },
    {
      "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": 0.8459593333333334,
      "accuracy_std_dev": 0.04023697333989667,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1740725378787879,
      "avg_duration_per_task_ms": 45559.39730639731,
      "avg_output_tokens_per_task": 3286.5547138047136,
      "last_run_timestamp": "2026-09-19T08:50:24.303Z"
    },
    {
      "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": 0.8459295011933173,
      "accuracy_std_dev": 0.01249384613206028,
      "run_count": 13,
      "total_questions": 2514,
      "avg_cost_per_task_usd": 0.023641319013524262,
      "avg_duration_per_task_ms": 567109.6607000795,
      "avg_output_tokens_per_task": 26995.523468575975,
      "last_run_timestamp": "2026-09-19T09:00:43.108Z"
    },
    {
      "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": 0.8458599302325581,
      "accuracy_std_dev": 0.025858269701844596,
      "run_count": 15,
      "total_questions": 2967,
      "avg_cost_per_task_usd": 0.012189807549713516,
      "avg_duration_per_task_ms": 182654.23222109876,
      "avg_output_tokens_per_task": 10138.908324907314,
      "last_run_timestamp": "2026-07-18T16:18:27.151Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Friendli",
      "accuracy": 0.8452890331905781,
      "accuracy_std_dev": 0.011290865996902098,
      "run_count": 5,
      "total_questions": 934,
      "avg_cost_per_task_usd": 0.01568154282655246,
      "avg_duration_per_task_ms": 212340.7505353319,
      "avg_output_tokens_per_task": 31157.209850107065,
      "last_run_timestamp": "2026-09-19T08:57:46.804Z"
    },
    {
      "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": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "auto-routing",
      "accuracy": 0.845117875,
      "accuracy_std_dev": 0.04165596517738336,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.17059957196969697,
      "avg_duration_per_task_ms": 43267.32512626263,
      "avg_output_tokens_per_task": 3208.438762626263,
      "last_run_timestamp": "2026-09-19T08:45:37.866Z"
    },
    {
      "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": 0.8443808801141771,
      "accuracy_std_dev": 0.0190832292537191,
      "run_count": 16,
      "total_questions": 3153,
      "avg_cost_per_task_usd": 0.02911775134792261,
      "avg_duration_per_task_ms": 432929.5179194418,
      "avg_output_tokens_per_task": 23367.47732318427,
      "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": "GMICloud",
      "accuracy": 0.8441076,
      "accuracy_std_dev": 0.01595676654714231,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.10875679292929294,
      "avg_duration_per_task_ms": 315962.43434343435,
      "avg_output_tokens_per_task": 27643.872727272726,
      "last_run_timestamp": "2026-09-19T08:54:22.565Z"
    },
    {
      "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": 0.843887868147448,
      "accuracy_std_dev": 0.023115623631383364,
      "run_count": 11,
      "total_questions": 2116,
      "avg_cost_per_task_usd": 0.14197860255198488,
      "avg_duration_per_task_ms": 310798.73534971644,
      "avg_output_tokens_per_task": 37568.326559546316,
      "last_run_timestamp": "2026-09-19T08:48:08.180Z"
    },
    {
      "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": 0.843855,
      "accuracy_std_dev": 0.020699899528741687,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.019635045454545454,
      "avg_duration_per_task_ms": 388426.77146464644,
      "avg_output_tokens_per_task": 35270.70202020202,
      "last_run_timestamp": "2026-09-19T08:58:24.996Z"
    },
    {
      "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": 0.8435248151724137,
      "accuracy_std_dev": 0.047063095761904726,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.04366690390804597,
      "avg_duration_per_task_ms": 260610.5043678161,
      "avg_output_tokens_per_task": 21718.556781609197,
      "last_run_timestamp": "2026-07-11T08:27:26.027Z"
    },
    {
      "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": 0.8435148007968127,
      "accuracy_std_dev": 0.009824077042964072,
      "run_count": 16,
      "total_questions": 3012,
      "avg_cost_per_task_usd": 0.0045806557104913674,
      "avg_duration_per_task_ms": 254608.46049136785,
      "avg_output_tokens_per_task": 11284.547476759628,
      "last_run_timestamp": "2026-08-29T18:39:38.845Z"
    },
    {
      "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": "Baidu",
      "accuracy": 0.8434342499999999,
      "accuracy_std_dev": 0.01988383838541994,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.00799521717171717,
      "avg_duration_per_task_ms": 74649.10037878787,
      "avg_output_tokens_per_task": 7345.006313131313,
      "last_run_timestamp": "2026-06-06T16:35:15.139Z"
    },
    {
      "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": 0.843434,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.02832048484848485,
      "avg_duration_per_task_ms": 290952.4191919192,
      "avg_output_tokens_per_task": 15039.919191919193,
      "last_run_timestamp": "2026-09-19T08:54:33.168Z"
    },
    {
      "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": 0.8434186912820514,
      "accuracy_std_dev": 0.048959906763890886,
      "run_count": 15,
      "total_questions": 2925,
      "avg_cost_per_task_usd": 0.12929724376068377,
      "avg_duration_per_task_ms": 178496.35042735044,
      "avg_output_tokens_per_task": 32231.285811965812,
      "last_run_timestamp": "2026-07-18T08:26:40.483Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DekaLLM",
      "accuracy": 0.8430610700389105,
      "accuracy_std_dev": 0.010186499999999987,
      "run_count": 2,
      "total_questions": 257,
      "avg_cost_per_task_usd": 0.03866427237354086,
      "avg_duration_per_task_ms": 140782.22178988328,
      "avg_output_tokens_per_task": 15417.210116731518,
      "last_run_timestamp": "2026-09-19T08:50:06.041Z"
    },
    {
      "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": 0.8427455454545453,
      "accuracy_std_dev": 0.01932638720348021,
      "run_count": 22,
      "total_questions": 4356,
      "avg_cost_per_task_usd": 0.16901573622589527,
      "avg_duration_per_task_ms": 157077.36363636365,
      "avg_output_tokens_per_task": 15869.985537190083,
      "last_run_timestamp": "2026-09-19T08:55:16.112Z"
    },
    {
      "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": 0.8426068009188362,
      "accuracy_std_dev": 0.022699620798815105,
      "run_count": 10,
      "total_questions": 1959,
      "avg_cost_per_task_usd": 0.12467511332312406,
      "avg_duration_per_task_ms": 432582.16334864724,
      "avg_output_tokens_per_task": 32390.03164880041,
      "last_run_timestamp": "2026-09-19T08:49:25.281Z"
    },
    {
      "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": 0.8425974369531652,
      "accuracy_std_dev": 0.020837639567858918,
      "run_count": 10,
      "total_questions": 1943,
      "avg_cost_per_task_usd": 0.18937706690684508,
      "avg_duration_per_task_ms": 350006.53216675244,
      "avg_output_tokens_per_task": 43348.177560473494,
      "last_run_timestamp": "2026-09-19T08:51:18.346Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Fireworks",
      "accuracy": 0.8425927500000001,
      "accuracy_std_dev": 0.014939580472272304,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.10931991035353535,
      "avg_duration_per_task_ms": 266096.30303030304,
      "avg_output_tokens_per_task": 24660.041666666668,
      "last_run_timestamp": "2026-09-12T08:53:18.149Z"
    },
    {
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Friendli",
      "accuracy": 0.8425808221092757,
      "accuracy_std_dev": 0.0344104863252307,
      "run_count": 12,
      "total_questions": 2361,
      "avg_cost_per_task_usd": 0.17626905887335875,
      "avg_duration_per_task_ms": 323040.35662854725,
      "avg_output_tokens_per_task": 40000.04659042779,
      "last_run_timestamp": "2026-09-19T09:14:29.020Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "StreamLake",
      "accuracy": 0.8423992077798862,
      "accuracy_std_dev": 0.022061600294686224,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.1114357827324478,
      "avg_duration_per_task_ms": 504012.30803289055,
      "avg_output_tokens_per_task": 34670.05819101835,
      "last_run_timestamp": "2026-08-29T20:19:40.446Z"
    },
    {
      "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": 0.8415372905296948,
      "accuracy_std_dev": 0.0365919096817196,
      "run_count": 19,
      "total_questions": 3738,
      "avg_cost_per_task_usd": 0.0653075548421616,
      "avg_duration_per_task_ms": 587777.9748528625,
      "avg_output_tokens_per_task": 26128.800428036382,
      "last_run_timestamp": "2026-08-29T18:40:35.044Z"
    },
    {
      "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": 0.8414075126536513,
      "accuracy_std_dev": 0.022108573391159597,
      "run_count": 21,
      "total_questions": 4149,
      "avg_cost_per_task_usd": 0.0148124528802121,
      "avg_duration_per_task_ms": 317232.769824054,
      "avg_output_tokens_per_task": 14106.06435285611,
      "last_run_timestamp": "2026-08-29T18:05:36.579Z"
    },
    {
      "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": 0.8411896666666666,
      "accuracy_std_dev": 0.00975199857579064,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.140150531986532,
      "avg_duration_per_task_ms": 165547.89562289562,
      "avg_output_tokens_per_task": 31680.003367003366,
      "last_run_timestamp": "2026-09-19T09:01:25.574Z"
    },
    {
      "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": 0.8410117804878049,
      "accuracy_std_dev": 0.0033504999999999785,
      "run_count": 2,
      "total_questions": 369,
      "avg_cost_per_task_usd": 0.07247923035230353,
      "avg_duration_per_task_ms": 402624.47967479675,
      "avg_output_tokens_per_task": 25758.56639566396,
      "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-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.840988087912088,
      "accuracy_std_dev": 0.024478415329178844,
      "run_count": 18,
      "total_questions": 3549,
      "avg_cost_per_task_usd": 0.12527278360101435,
      "avg_duration_per_task_ms": 593414.8410819949,
      "avg_output_tokens_per_task": 31740.369118061426,
      "last_run_timestamp": "2026-08-29T18:39:39.402Z"
    },
    {
      "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": "SiliconFlow",
      "accuracy": 0.8409090999999999,
      "accuracy_std_dev": 0.016326424363282965,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.029859258080808075,
      "avg_duration_per_task_ms": 210807.68282828282,
      "avg_output_tokens_per_task": 13522.216161616161,
      "last_run_timestamp": "2026-06-06T17:25:31.571Z"
    },
    {
      "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": 0.8408720981210859,
      "accuracy_std_dev": 0.02022286087457656,
      "run_count": 22,
      "total_questions": 4311,
      "avg_cost_per_task_usd": 0.05355351844119695,
      "avg_duration_per_task_ms": 553888.6455578752,
      "avg_output_tokens_per_task": 24570.292275574113,
      "last_run_timestamp": "2026-08-30T00:28:09.702Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Friendli",
      "accuracy": 0.8405005064184852,
      "accuracy_std_dev": 0.014666451484675618,
      "run_count": 16,
      "total_questions": 3116,
      "avg_cost_per_task_usd": 0.14646416527599487,
      "avg_duration_per_task_ms": 284641.20763799746,
      "avg_output_tokens_per_task": 33175.712451861364,
      "last_run_timestamp": "2026-08-29T19:00:02.941Z"
    },
    {
      "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": 0.8402252180451129,
      "accuracy_std_dev": 0.006276000000000004,
      "run_count": 2,
      "total_questions": 266,
      "avg_cost_per_task_usd": 0.00410540977443609,
      "avg_duration_per_task_ms": 368590.0263157895,
      "avg_output_tokens_per_task": 8011.255639097744,
      "last_run_timestamp": "2026-09-19T08:47:33.865Z"
    },
    {
      "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": 0.8400309335394126,
      "accuracy_std_dev": 0.029718375539176087,
      "run_count": 4,
      "total_questions": 647,
      "avg_cost_per_task_usd": 0.07242874188562597,
      "avg_duration_per_task_ms": 447585.9366306028,
      "avg_output_tokens_per_task": 23737.97372488408,
      "last_run_timestamp": "2026-09-19T08:52:28.200Z"
    },
    {
      "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": 0.8398567499999999,
      "accuracy_std_dev": 0.038088117043738205,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.17395044507575758,
      "avg_duration_per_task_ms": 45323.468434343435,
      "avg_output_tokens_per_task": 3273.729797979798,
      "last_run_timestamp": "2026-09-19T08:52:27.050Z"
    },
    {
      "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": 0.839856625,
      "accuracy_std_dev": 0.02003363270813297,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.15796164583333333,
      "avg_duration_per_task_ms": 153579.38446969696,
      "avg_output_tokens_per_task": 15584.020833333334,
      "last_run_timestamp": "2026-09-19T08:49:06.740Z"
    },
    {
      "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": 0.8397368503538927,
      "accuracy_std_dev": 0.015255152920898565,
      "run_count": 5,
      "total_questions": 989,
      "avg_cost_per_task_usd": 0.12783742972699697,
      "avg_duration_per_task_ms": 159433.03235591506,
      "avg_output_tokens_per_task": 28836.38321536906,
      "last_run_timestamp": "2026-09-19T09:00:01.729Z"
    },
    {
      "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": 0.839052991845056,
      "accuracy_std_dev": 0.019945151183070924,
      "run_count": 15,
      "total_questions": 2943,
      "avg_cost_per_task_usd": 0.053893335372069334,
      "avg_duration_per_task_ms": 628913.658851512,
      "avg_output_tokens_per_task": 30616.803601766904,
      "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": "Reka",
      "accuracy": 0.83880475,
      "accuracy_std_dev": 0.009249480644203778,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07424993055555555,
      "avg_duration_per_task_ms": 281356.54671717173,
      "avg_output_tokens_per_task": 20276.51388888889,
      "last_run_timestamp": "2026-09-19T08:49:08.732Z"
    },
    {
      "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": 0.8386599858635525,
      "accuracy_std_dev": 0.19405393544087168,
      "run_count": 10,
      "total_questions": 1627,
      "avg_cost_per_task_usd": 0.14442266195451753,
      "avg_duration_per_task_ms": 280013.19237861095,
      "avg_output_tokens_per_task": 11154.37738168408,
      "last_run_timestamp": "2026-09-20T10:39:59.883Z"
    },
    {
      "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": 0.8384512456608813,
      "accuracy_std_dev": 0.021133434301110997,
      "run_count": 8,
      "total_questions": 1498,
      "avg_cost_per_task_usd": 0.0535803865153538,
      "avg_duration_per_task_ms": 282678.4526034713,
      "avg_output_tokens_per_task": 22384.684245660883,
      "last_run_timestamp": "2026-09-19T10:08:51.605Z"
    },
    {
      "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": 0.8384059735973598,
      "accuracy_std_dev": 0.03372879005412596,
      "run_count": 14,
      "total_questions": 2727,
      "avg_cost_per_task_usd": 0.05032389952328565,
      "avg_duration_per_task_ms": 247684.96699669966,
      "avg_output_tokens_per_task": 22104.830583058305,
      "last_run_timestamp": "2026-09-19T09:24:21.238Z"
    },
    {
      "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": "Parasail",
      "accuracy": 0.8383835714285715,
      "accuracy_std_dev": 0.009733641278093865,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.031499565656565656,
      "avg_duration_per_task_ms": 532165.215007215,
      "avg_output_tokens_per_task": 15048.665223665224,
      "last_run_timestamp": "2026-06-06T16:02:18.201Z"
    },
    {
      "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": 0.8382395753791257,
      "accuracy_std_dev": 0.015894466289989074,
      "run_count": 17,
      "total_questions": 3363,
      "avg_cost_per_task_usd": 0.016586219744275946,
      "avg_duration_per_task_ms": 262022.0906928338,
      "avg_output_tokens_per_task": 15243.496580434135,
      "last_run_timestamp": "2026-08-29T18:09:03.379Z"
    },
    {
      "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": 0.837957142668428,
      "accuracy_std_dev": 0.006945026007870667,
      "run_count": 4,
      "total_questions": 757,
      "avg_cost_per_task_usd": 0.019369532364597094,
      "avg_duration_per_task_ms": 59399.14531043593,
      "avg_output_tokens_per_task": 15971.39894319683,
      "last_run_timestamp": "2026-09-19T09:12:31.094Z"
    },
    {
      "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": 0.8378685034013607,
      "accuracy_std_dev": 0.007616806900532539,
      "run_count": 5,
      "total_questions": 882,
      "avg_cost_per_task_usd": 0.018575892290249432,
      "avg_duration_per_task_ms": 593585.7845804988,
      "avg_output_tokens_per_task": 36956.14739229025,
      "last_run_timestamp": "2026-09-19T08:46:08.149Z"
    },
    {
      "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": 0.8377076351706036,
      "accuracy_std_dev": 0.022782891525221304,
      "run_count": 4,
      "total_questions": 762,
      "avg_cost_per_task_usd": 0.01399598031496063,
      "avg_duration_per_task_ms": 689440.656167979,
      "avg_output_tokens_per_task": 33978.16535433071,
      "last_run_timestamp": "2026-09-12T08:53:51.555Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8375420000000001,
      "accuracy_std_dev": 0.005646444323642987,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.08216196843434342,
      "avg_duration_per_task_ms": 323315.19444444444,
      "avg_output_tokens_per_task": 19594.77398989899,
      "last_run_timestamp": "2026-09-19T08:47:56.323Z"
    },
    {
      "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": 0.8374022479391249,
      "accuracy_std_dev": 0.08262800793142024,
      "run_count": 16,
      "total_questions": 3154,
      "avg_cost_per_task_usd": 0.09566040615091946,
      "avg_duration_per_task_ms": 471332.48065948003,
      "avg_output_tokens_per_task": 29376.78376664553,
      "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-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.8371893307626392,
      "accuracy_std_dev": 0.0077992202671491664,
      "run_count": 7,
      "total_questions": 1167,
      "avg_cost_per_task_usd": 0.01691520051413882,
      "avg_duration_per_task_ms": 189610.05741216795,
      "avg_output_tokens_per_task": 33640.42673521851,
      "last_run_timestamp": "2026-09-19T09:05:59.416Z"
    },
    {
      "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8369750858164482,
      "accuracy_std_dev": 0.010022153871541413,
      "run_count": 13,
      "total_questions": 2517,
      "avg_cost_per_task_usd": 0.06672872864521255,
      "avg_duration_per_task_ms": 709736.1537544696,
      "avg_output_tokens_per_task": 40652.02900278109,
      "last_run_timestamp": "2026-09-19T16:00:24.089Z"
    },
    {
      "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": 0.8368604243386243,
      "accuracy_std_dev": 0.053504834431884106,
      "run_count": 6,
      "total_questions": 945,
      "avg_cost_per_task_usd": 0.05630135555555556,
      "avg_duration_per_task_ms": 268944.617989418,
      "avg_output_tokens_per_task": 18648.48253968254,
      "last_run_timestamp": "2026-08-15T15:24:07.556Z"
    },
    {
      "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": 0.8367553969581748,
      "accuracy_std_dev": 0.016143329277057448,
      "run_count": 20,
      "total_questions": 3945,
      "avg_cost_per_task_usd": 0.04780176628643852,
      "avg_duration_per_task_ms": 445163.75487959443,
      "avg_output_tokens_per_task": 25765.407351077312,
      "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": "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Together",
      "accuracy": 0.8352059691011233,
      "accuracy_std_dev": 0.03754617901216079,
      "run_count": 11,
      "total_questions": 2136,
      "avg_cost_per_task_usd": 0.14707353417602997,
      "avg_duration_per_task_ms": 365510.97191011236,
      "avg_output_tokens_per_task": 33578.668539325845,
      "last_run_timestamp": "2026-08-15T08:57:45.830Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Alibaba",
      "accuracy": 0.8350164999999999,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.16845593434343434,
      "avg_duration_per_task_ms": 341569.6590909091,
      "avg_output_tokens_per_task": 41205.416666666664,
      "last_run_timestamp": "2026-09-19T08:58:05.571Z"
    },
    {
      "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": 0.8349114289772727,
      "accuracy_std_dev": 0.012684612209645646,
      "run_count": 11,
      "total_questions": 2112,
      "avg_cost_per_task_usd": 0.09840011221590908,
      "avg_duration_per_task_ms": 269461.8380681818,
      "avg_output_tokens_per_task": 24744.351799242424,
      "last_run_timestamp": "2026-07-18T08:29:37.313Z"
    },
    {
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "SambaNova",
      "accuracy": 0.8345957500000001,
      "accuracy_std_dev": 0.014451436422982317,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0074238952020202024,
      "avg_duration_per_task_ms": 38293.18308080808,
      "avg_output_tokens_per_task": 7280.891414141414,
      "last_run_timestamp": "2026-05-16T16:15:42.787Z"
    },
    {
      "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": 0.8344652224938874,
      "accuracy_std_dev": 0.01975287567409165,
      "run_count": 19,
      "total_questions": 3681,
      "avg_cost_per_task_usd": 0.003750411572942135,
      "avg_duration_per_task_ms": 316450.0222765553,
      "avg_output_tokens_per_task": 9605.333061668025,
      "last_run_timestamp": "2026-08-29T19:56:51.477Z"
    },
    {
      "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": "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/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "CoreWeave",
      "accuracy": 0.8341078565737051,
      "accuracy_std_dev": 0.01236126339861748,
      "run_count": 10,
      "total_questions": 1506,
      "avg_cost_per_task_usd": 0.03458700464807437,
      "avg_duration_per_task_ms": 234385.25099601594,
      "avg_output_tokens_per_task": 27567.501328021248,
      "last_run_timestamp": "2026-09-19T16:00:28.636Z"
    },
    {
      "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": 0.8340128348623852,
      "accuracy_std_dev": 0.01276366393869724,
      "run_count": 20,
      "total_questions": 3924,
      "avg_cost_per_task_usd": 0.004765704128440367,
      "avg_duration_per_task_ms": 173295.0881753313,
      "avg_output_tokens_per_task": 15701.070846075434,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Novita",
      "accuracy": 0.8340065999999999,
      "accuracy_std_dev": 0.029823943324785232,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.13819853030303028,
      "avg_duration_per_task_ms": 378353.51414141414,
      "avg_output_tokens_per_task": 35316.50404040404,
      "last_run_timestamp": "2026-09-19T08:49:31.209Z"
    },
    {
      "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": 0.8338945000000001,
      "accuracy_std_dev": 0.036102903838657247,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.6825328661616162,
      "avg_duration_per_task_ms": 471720.8164983165,
      "avg_output_tokens_per_task": 27444.19781144781,
      "last_run_timestamp": "2026-09-19T08:48:28.295Z"
    },
    {
      "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": "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": 0.833614,
      "accuracy_std_dev": 0.016210715242291687,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.012104537037037037,
      "avg_duration_per_task_ms": 473962.1228956229,
      "avg_output_tokens_per_task": 40122.851851851854,
      "last_run_timestamp": "2026-09-19T08:44:54.481Z"
    },
    {
      "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": 0.8334797826086957,
      "accuracy_std_dev": 0.038296180861868125,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.05379731664470795,
      "avg_duration_per_task_ms": 258855.73583662714,
      "avg_output_tokens_per_task": 12134.429732103645,
      "last_run_timestamp": "2026-08-29T18:50:46.982Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Parasail",
      "accuracy": 0.8332221724061321,
      "accuracy_std_dev": 0.13397161638673358,
      "run_count": 23,
      "total_questions": 4501,
      "avg_cost_per_task_usd": 0.03938373672517219,
      "avg_duration_per_task_ms": 180863.35369917797,
      "avg_output_tokens_per_task": 10881.63741390802,
      "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-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Novita",
      "accuracy": 0.833212137254902,
      "accuracy_std_dev": 0.02511813859250007,
      "run_count": 14,
      "total_questions": 2754,
      "avg_cost_per_task_usd": 0.06167101997095135,
      "avg_duration_per_task_ms": 884018.2875816993,
      "avg_output_tokens_per_task": 28448.484749455336,
      "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-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "auto-routing",
      "accuracy": 0.8329946475964793,
      "accuracy_std_dev": 0.03338342053015737,
      "run_count": 15,
      "total_questions": 2954,
      "avg_cost_per_task_usd": 0.11301010663507109,
      "avg_duration_per_task_ms": 707116.9400812457,
      "avg_output_tokens_per_task": 34269.73628977657,
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Mistral",
      "accuracy": 0.8328522857142855,
      "accuracy_std_dev": 0.011962448742559688,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.24642647763347764,
      "avg_duration_per_task_ms": 279472.34812409815,
      "avg_output_tokens_per_task": 54220.55988455989,
      "last_run_timestamp": "2026-09-17T02:17:44.342Z"
    },
    {
      "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": 0.8328282,
      "accuracy_std_dev": 0.014614290846975768,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.13095260505050504,
      "avg_duration_per_task_ms": 155759.8383838384,
      "avg_output_tokens_per_task": 29521.12626262626,
      "last_run_timestamp": "2026-09-19T08:58:36.320Z"
    },
    {
      "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": 0.8327721666666664,
      "accuracy_std_dev": 0.03409742452598958,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.014290039562289561,
      "avg_duration_per_task_ms": 292947.5042087542,
      "avg_output_tokens_per_task": 11998.566077441077,
      "last_run_timestamp": "2026-07-18T09:15:42.531Z"
    },
    {
      "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": 0.832677994757536,
      "accuracy_std_dev": 0.01790311723380873,
      "run_count": 4,
      "total_questions": 763,
      "avg_cost_per_task_usd": 0.14453915727391875,
      "avg_duration_per_task_ms": 303619.376146789,
      "avg_output_tokens_per_task": 32661.047182175622,
      "last_run_timestamp": "2026-09-19T08:49:14.892Z"
    },
    {
      "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": "Together",
      "accuracy": 0.8326117142857142,
      "accuracy_std_dev": 0.02081119738927845,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.05571586471861472,
      "avg_duration_per_task_ms": 232585.47619047618,
      "avg_output_tokens_per_task": 17366.68867243867,
      "last_run_timestamp": "2026-06-20T17:09:19.846Z"
    },
    {
      "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": 0.8324706727037516,
      "accuracy_std_dev": 0.015362295521421925,
      "run_count": 4,
      "total_questions": 773,
      "avg_cost_per_task_usd": 0.104053553686934,
      "avg_duration_per_task_ms": 725443.3247089263,
      "avg_output_tokens_per_task": 27644.671410090556,
      "last_run_timestamp": "2026-09-19T08:52:28.103Z"
    },
    {
      "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": 0.8319369960317462,
      "accuracy_std_dev": 0.026396632252184915,
      "run_count": 23,
      "total_questions": 4536,
      "avg_cost_per_task_usd": 0.04784231701940035,
      "avg_duration_per_task_ms": 507122.99823633156,
      "avg_output_tokens_per_task": 24258.074074074073,
      "last_run_timestamp": "2026-08-30T00:30:19.495Z"
    },
    {
      "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": 0.8318344082733814,
      "accuracy_std_dev": 0.03947741134153016,
      "run_count": 17,
      "total_questions": 3336,
      "avg_cost_per_task_usd": 0.14675382314148683,
      "avg_duration_per_task_ms": 651191.3980815348,
      "avg_output_tokens_per_task": 33576.36121103117,
      "last_run_timestamp": "2026-08-29T18:21:26.899Z"
    },
    {
      "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": 0.8316448145896658,
      "accuracy_std_dev": 0.03141045267820889,
      "run_count": 10,
      "total_questions": 1974,
      "avg_cost_per_task_usd": 0.11912360688956432,
      "avg_duration_per_task_ms": 503318.0197568389,
      "avg_output_tokens_per_task": 34925.666666666664,
      "last_run_timestamp": "2026-09-19T08:47:24.459Z"
    },
    {
      "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": 0.8313797864493997,
      "accuracy_std_dev": 0.024754194188330587,
      "run_count": 18,
      "total_questions": 3498,
      "avg_cost_per_task_usd": 0.04931393767867353,
      "avg_duration_per_task_ms": 846895.9456832475,
      "avg_output_tokens_per_task": 28170.789879931388,
      "last_run_timestamp": "2026-08-22T17:50:04.003Z"
    },
    {
      "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": 0.8313591965192169,
      "accuracy_std_dev": 0.02385695725514725,
      "run_count": 21,
      "total_questions": 4137,
      "avg_cost_per_task_usd": 0.015896513173797436,
      "avg_duration_per_task_ms": 285786.35170413344,
      "avg_output_tokens_per_task": 13215.3529127387,
      "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-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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "auto-routing",
      "accuracy": 0.8312575670329668,
      "accuracy_std_dev": 0.0356512774660921,
      "run_count": 14,
      "total_questions": 2730,
      "avg_cost_per_task_usd": 0.04862300952380953,
      "avg_duration_per_task_ms": 395101.81245421246,
      "avg_output_tokens_per_task": 19499.453846153847,
      "last_run_timestamp": "2026-08-01T17:53:10.275Z"
    },
    {
      "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": 0.8310886,
      "accuracy_std_dev": 0.05088763866506941,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.1784800898989899,
      "avg_duration_per_task_ms": 40054.59898989899,
      "avg_output_tokens_per_task": 3209.878451178451,
      "last_run_timestamp": "2026-09-19T09:08:31.785Z"
    },
    {
      "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": "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-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Crusoe",
      "accuracy": 0.8306985375494073,
      "accuracy_std_dev": 0.018572923858850742,
      "run_count": 3,
      "total_questions": 506,
      "avg_cost_per_task_usd": 0.020810217391304345,
      "avg_duration_per_task_ms": 244282.4347826087,
      "avg_output_tokens_per_task": 41406.6837944664,
      "last_run_timestamp": "2026-09-19T08:48:21.497Z"
    },
    {
      "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": 0.8304598896551725,
      "accuracy_std_dev": 0.003983459286711617,
      "run_count": 6,
      "total_questions": 1160,
      "avg_cost_per_task_usd": 0.03665691206896552,
      "avg_duration_per_task_ms": 205169.7853448276,
      "avg_output_tokens_per_task": 13753.440517241379,
      "last_run_timestamp": "2026-09-19T08:48:39.285Z"
    },
    {
      "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": 0.830442043478261,
      "accuracy_std_dev": 0.013460761850905967,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.060090737154150205,
      "avg_duration_per_task_ms": 218919.57114624506,
      "avg_output_tokens_per_task": 27589.693895476503,
      "last_run_timestamp": "2026-08-30T00:09:46.578Z"
    },
    {
      "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": 0.8303131074589128,
      "accuracy_std_dev": 0.009856971411895784,
      "run_count": 24,
      "total_questions": 4746,
      "avg_cost_per_task_usd": 0.00450906805731142,
      "avg_duration_per_task_ms": 297774.21533923305,
      "avg_output_tokens_per_task": 11688.514327855035,
      "last_run_timestamp": "2026-08-29T16:40:58.367Z"
    },
    {
      "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": 0.8302964372990355,
      "accuracy_std_dev": 0.05435128994016608,
      "run_count": 5,
      "total_questions": 933,
      "avg_cost_per_task_usd": 0.004838944265809216,
      "avg_duration_per_task_ms": 341319.115755627,
      "avg_output_tokens_per_task": 13605.640943193997,
      "last_run_timestamp": "2026-07-18T08:04:02.357Z"
    },
    {
      "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": 0.8299663333333334,
      "accuracy_std_dev": 0.009322650368263805,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.04014494781144781,
      "avg_duration_per_task_ms": 219758.45622895623,
      "avg_output_tokens_per_task": 14463.257575757576,
      "last_run_timestamp": "2026-09-19T08:47:20.012Z"
    },
    {
      "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": 0.8297885767530007,
      "accuracy_std_dev": 0.0250017324703876,
      "run_count": 24,
      "total_questions": 4749,
      "avg_cost_per_task_usd": 0.16577993451252898,
      "avg_duration_per_task_ms": 150407.92208886083,
      "avg_output_tokens_per_task": 15324.479679932618,
      "last_run_timestamp": "2026-09-19T08:54:35.988Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.8295164461407973,
      "accuracy_std_dev": 0.030217333777287914,
      "run_count": 18,
      "total_questions": 3537,
      "avg_cost_per_task_usd": 0.13597140797285837,
      "avg_duration_per_task_ms": 506852.9443030817,
      "avg_output_tokens_per_task": 33591.725190839694,
      "last_run_timestamp": "2026-08-29T18:09:15.344Z"
    },
    {
      "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": 0.8294639871023217,
      "accuracy_std_dev": 0.027475351240747944,
      "run_count": 18,
      "total_questions": 3489,
      "avg_cost_per_task_usd": 0.1216583038119805,
      "avg_duration_per_task_ms": 574849.784752078,
      "avg_output_tokens_per_task": 33861.400401261104,
      "last_run_timestamp": "2026-08-29T18:14:11.455Z"
    },
    {
      "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": 0.8294050555555555,
      "accuracy_std_dev": 0.021508202180853447,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.02960114337822672,
      "avg_duration_per_task_ms": 125113.44304152638,
      "avg_output_tokens_per_task": 16376.810606060606,
      "last_run_timestamp": "2026-08-29T18:13:32.148Z"
    },
    {
      "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": 0.829367646430954,
      "accuracy_std_dev": 0.05068897680083226,
      "run_count": 23,
      "total_questions": 4497,
      "avg_cost_per_task_usd": 0.10293063842561709,
      "avg_duration_per_task_ms": 624030.37914165,
      "avg_output_tokens_per_task": 32772.14164998888,
      "last_run_timestamp": "2026-08-29T18:11:51.472Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8293240541237112,
      "accuracy_std_dev": 0.05663663093125379,
      "run_count": 12,
      "total_questions": 2328,
      "avg_cost_per_task_usd": 0.04663492611683849,
      "avg_duration_per_task_ms": 552324.0841924399,
      "avg_output_tokens_per_task": 22329.278350515466,
      "last_run_timestamp": "2026-09-19T08:14:35.940Z"
    },
    {
      "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": 0.829273741542625,
      "accuracy_std_dev": 0.020525043792584238,
      "run_count": 4,
      "total_questions": 739,
      "avg_cost_per_task_usd": 0.023301357239512856,
      "avg_duration_per_task_ms": 305516.3085250338,
      "avg_output_tokens_per_task": 46400.958051420836,
      "last_run_timestamp": "2026-09-19T08:49:44.918Z"
    },
    {
      "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": 0.8291479228699551,
      "accuracy_std_dev": 0.046765034341613804,
      "run_count": 34,
      "total_questions": 6690,
      "avg_cost_per_task_usd": 0.03550391076233183,
      "avg_duration_per_task_ms": 277215.60224215244,
      "avg_output_tokens_per_task": 18441.03587443946,
      "last_run_timestamp": "2026-08-29T19:17:08.288Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AkashML",
      "accuracy": 0.8283502494845361,
      "accuracy_std_dev": 0.01123608353475534,
      "run_count": 5,
      "total_questions": 970,
      "avg_cost_per_task_usd": 0.12046294226804122,
      "avg_duration_per_task_ms": 260289.6525773196,
      "avg_output_tokens_per_task": 28436.055670103095,
      "last_run_timestamp": "2026-09-19T08:59:56.999Z"
    },
    {
      "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": "DekaLLM",
      "accuracy": 0.8282827500000001,
      "accuracy_std_dev": 0.02172313812936567,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0027490239898989898,
      "avg_duration_per_task_ms": 108260.87878787878,
      "avg_output_tokens_per_task": 2717.243686868687,
      "last_run_timestamp": "2026-05-16T08:55:42.302Z"
    },
    {
      "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": 0.82828275,
      "accuracy_std_dev": 0.014334603419261384,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.012366218434343434,
      "avg_duration_per_task_ms": 835296.3156565657,
      "avg_output_tokens_per_task": 36481.89646464647,
      "last_run_timestamp": "2026-09-19T08:15:51.634Z"
    },
    {
      "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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8276921538461538,
      "accuracy_std_dev": 0.027161692421496852,
      "run_count": 10,
      "total_questions": 1950,
      "avg_cost_per_task_usd": 0.032337288205128205,
      "avg_duration_per_task_ms": 326118.6758974359,
      "avg_output_tokens_per_task": 12882.684102564102,
      "last_run_timestamp": "2026-06-30T22:55:56.214Z"
    },
    {
      "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": 0.8275520710659898,
      "accuracy_std_dev": 0.0776517694741866,
      "run_count": 12,
      "total_questions": 2364,
      "avg_cost_per_task_usd": 0.07118247335025381,
      "avg_duration_per_task_ms": 485649.0562605753,
      "avg_output_tokens_per_task": 31102.425126903552,
      "last_run_timestamp": "2026-09-19T08:54:19.219Z"
    },
    {
      "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": 0.8273554038986354,
      "accuracy_std_dev": 0.018032038830334687,
      "run_count": 13,
      "total_questions": 2565,
      "avg_cost_per_task_usd": 0.020172549707602336,
      "avg_duration_per_task_ms": 335505.9828460039,
      "avg_output_tokens_per_task": 20130.481091617934,
      "last_run_timestamp": "2026-08-29T16:15:12.678Z"
    },
    {
      "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": 0.8272305,
      "accuracy_std_dev": 0.016036241150905673,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.06078435227272727,
      "avg_duration_per_task_ms": 492639.2809343434,
      "avg_output_tokens_per_task": 47927.3244949495,
      "last_run_timestamp": "2026-09-19T16:08:54.164Z"
    },
    {
      "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": 0.8271558080733944,
      "accuracy_std_dev": 0.040507064699838946,
      "run_count": 14,
      "total_questions": 2725,
      "avg_cost_per_task_usd": 0.016758301284403674,
      "avg_duration_per_task_ms": 220975.85504587155,
      "avg_output_tokens_per_task": 14210.697247706423,
      "last_run_timestamp": "2026-07-18T08:29:12.575Z"
    },
    {
      "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": 0.8267805538461538,
      "accuracy_std_dev": 0.020665297105829934,
      "run_count": 12,
      "total_questions": 2340,
      "avg_cost_per_task_usd": 0.08551232649572652,
      "avg_duration_per_task_ms": 577054.2572649573,
      "avg_output_tokens_per_task": 31689.464957264958,
      "last_run_timestamp": "2026-08-29T18:26:14.942Z"
    },
    {
      "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": 0.8267103169705471,
      "accuracy_std_dev": 0.014502002597532615,
      "run_count": 11,
      "total_questions": 2139,
      "avg_cost_per_task_usd": 0.009508418419822345,
      "avg_duration_per_task_ms": 526491.7872837775,
      "avg_output_tokens_per_task": 33720.50303880318,
      "last_run_timestamp": "2026-09-19T09:13:55.353Z"
    },
    {
      "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": "Cloudflare",
      "accuracy": 0.8263960270551509,
      "accuracy_std_dev": 0.04766683339597883,
      "run_count": 5,
      "total_questions": 961,
      "avg_cost_per_task_usd": 0.028155845993756503,
      "avg_duration_per_task_ms": 288712.7356919875,
      "avg_output_tokens_per_task": 8693.547346514048,
      "last_run_timestamp": "2026-09-19T09:20:52.921Z"
    },
    {
      "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": 0.8262626000000001,
      "accuracy_std_dev": 0.005773190715713467,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1255757282828283,
      "avg_duration_per_task_ms": 138406.39494949495,
      "avg_output_tokens_per_task": 30506.08282828283,
      "last_run_timestamp": "2026-09-19T08:50:32.999Z"
    },
    {
      "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": 0.82617825,
      "accuracy_std_dev": 0.007374361408793328,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.045437868686868686,
      "avg_duration_per_task_ms": 57534.18939393939,
      "avg_output_tokens_per_task": 11194.286616161617,
      "last_run_timestamp": "2026-09-19T08:49:40.730Z"
    },
    {
      "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": "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": 0.8259803298319327,
      "accuracy_std_dev": 0.013069264196579715,
      "run_count": 5,
      "total_questions": 952,
      "avg_cost_per_task_usd": 0.04310651365546219,
      "avg_duration_per_task_ms": 411042.2016806723,
      "avg_output_tokens_per_task": 42229.66176470588,
      "last_run_timestamp": "2026-09-19T08:50:24.150Z"
    },
    {
      "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": 0.8259176777899344,
      "accuracy_std_dev": 0.050464276054948014,
      "run_count": 28,
      "total_questions": 5484,
      "avg_cost_per_task_usd": 0.030212785740335524,
      "avg_duration_per_task_ms": 236266.8163749088,
      "avg_output_tokens_per_task": 15214.504011670313,
      "last_run_timestamp": "2026-08-01T17:44:01.464Z"
    },
    {
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8255112924187727,
      "accuracy_std_dev": 0.033062154820560075,
      "run_count": 21,
      "total_questions": 4155,
      "avg_cost_per_task_usd": 0.01651286329723225,
      "avg_duration_per_task_ms": 227671.96774969916,
      "avg_output_tokens_per_task": 13721.429602888087,
      "last_run_timestamp": "2026-08-29T18:09:57.600Z"
    },
    {
      "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": 0.825477,
      "accuracy_std_dev": 0.0028613344206273234,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04118768013468014,
      "avg_duration_per_task_ms": 215669.1313131313,
      "avg_output_tokens_per_task": 14940.681818181818,
      "last_run_timestamp": "2026-09-19T09:12:21.615Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8252867162300974,
      "accuracy_std_dev": 0.1181226298338754,
      "run_count": 21,
      "total_questions": 3894,
      "avg_cost_per_task_usd": 0.03139470313302517,
      "avg_duration_per_task_ms": 234154.43913713406,
      "avg_output_tokens_per_task": 9983.509758602979,
      "last_run_timestamp": "2026-08-29T18:43:58.355Z"
    },
    {
      "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": 0.8246641421568626,
      "accuracy_std_dev": 0.04102567192972371,
      "run_count": 19,
      "total_questions": 3672,
      "avg_cost_per_task_usd": 0.09416541448801743,
      "avg_duration_per_task_ms": 832671.316993464,
      "avg_output_tokens_per_task": 28688.930555555555,
      "last_run_timestamp": "2026-08-29T18:08:11.373Z"
    },
    {
      "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": 0.8245855497237569,
      "accuracy_std_dev": 0.05595579604141202,
      "run_count": 11,
      "total_questions": 2172,
      "avg_cost_per_task_usd": 0.02354478360957643,
      "avg_duration_per_task_ms": 541510.1505524862,
      "avg_output_tokens_per_task": 25036.35497237569,
      "last_run_timestamp": "2026-09-19T09:09:19.272Z"
    },
    {
      "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": 0.8245443432510885,
      "accuracy_std_dev": 0.014070989097480181,
      "run_count": 21,
      "total_questions": 4134,
      "avg_cost_per_task_usd": 0.0048930220125786164,
      "avg_duration_per_task_ms": 408825.40662796324,
      "avg_output_tokens_per_task": 11390.059748427673,
      "last_run_timestamp": "2026-08-29T18:40:24.101Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Novita",
      "accuracy": 0.8242424,
      "accuracy_std_dev": 0.011692885317149064,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.043399320202020204,
      "avg_duration_per_task_ms": 227400.42525252525,
      "avg_output_tokens_per_task": 14348.484848484848,
      "last_run_timestamp": "2026-09-19T08:47:45.478Z"
    },
    {
      "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": 0.823863625,
      "accuracy_std_dev": 0.014332832029448166,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.31506537247474753,
      "avg_duration_per_task_ms": 289259.720959596,
      "avg_output_tokens_per_task": 20815.732954545456,
      "last_run_timestamp": "2026-09-19T08:48:25.679Z"
    },
    {
      "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": 0.8234257166446501,
      "accuracy_std_dev": 0.04369321987022243,
      "run_count": 23,
      "total_questions": 4542,
      "avg_cost_per_task_usd": 0.06675309929546455,
      "avg_duration_per_task_ms": 567531.9035667107,
      "avg_output_tokens_per_task": 30278.626376045795,
      "last_run_timestamp": "2026-08-29T18:40:30.684Z"
    },
    {
      "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": 0.82339096128561,
      "accuracy_std_dev": 0.01836790238580393,
      "run_count": 21,
      "total_questions": 4107,
      "avg_cost_per_task_usd": 0.10042522181641102,
      "avg_duration_per_task_ms": 563721.646457268,
      "avg_output_tokens_per_task": 31885.672510348188,
      "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-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8232322727272725,
      "accuracy_std_dev": 0.03750024958969282,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.00558554453627181,
      "avg_duration_per_task_ms": 239700.90036730946,
      "avg_output_tokens_per_task": 13322.357208448118,
      "last_run_timestamp": "2026-08-01T20:03:34.768Z"
    },
    {
      "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": 0.823232,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.03873228787878788,
      "avg_duration_per_task_ms": 172542.09595959596,
      "avg_output_tokens_per_task": 15454.005050505051,
      "last_run_timestamp": "2026-09-20T21:05:40.652Z"
    },
    {
      "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": 0.8231507132003797,
      "accuracy_std_dev": 0.013738132759397979,
      "run_count": 16,
      "total_questions": 3159,
      "avg_cost_per_task_usd": 0.012798601772712881,
      "avg_duration_per_task_ms": 385156.8765432099,
      "avg_output_tokens_per_task": 12700.649889205444,
      "last_run_timestamp": "2026-08-29T18:35:57.732Z"
    },
    {
      "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": 0.823107677419355,
      "accuracy_std_dev": 0.0400712127273121,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.164333871600253,
      "avg_duration_per_task_ms": 161114.86337760912,
      "avg_output_tokens_per_task": 16212.869070208728,
      "last_run_timestamp": "2026-09-19T08:49:31.114Z"
    },
    {
      "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": 0.8222913416149069,
      "accuracy_std_dev": 0.020103371140184446,
      "run_count": 5,
      "total_questions": 966,
      "avg_cost_per_task_usd": 0.13681437163561078,
      "avg_duration_per_task_ms": 249682.33954451347,
      "avg_output_tokens_per_task": 30864.992753623188,
      "last_run_timestamp": "2026-09-19T08:50:10.014Z"
    },
    {
      "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": 0.8222221999999999,
      "accuracy_std_dev": 0.01398833591818555,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.034438841414141406,
      "avg_duration_per_task_ms": 232730.4090909091,
      "avg_output_tokens_per_task": 13762.19696969697,
      "last_run_timestamp": "2026-09-19T08:49:03.345Z"
    },
    {
      "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": 0.8221166532066507,
      "accuracy_std_dev": 0.015661128267637187,
      "run_count": 13,
      "total_questions": 2526,
      "avg_cost_per_task_usd": 0.03727458986539984,
      "avg_duration_per_task_ms": 329571.73871733964,
      "avg_output_tokens_per_task": 20674.40657165479,
      "last_run_timestamp": "2026-08-29T16:44:34.185Z"
    },
    {
      "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": 0.82211,
      "accuracy_std_dev": 0.02580143550915981,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.03809178872053872,
      "avg_duration_per_task_ms": 290291.4090909091,
      "avg_output_tokens_per_task": 14257.26936026936,
      "last_run_timestamp": "2026-09-19T08:52:03.747Z"
    },
    {
      "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": 0.8219697499999998,
      "accuracy_std_dev": 0.013060489125124707,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.03950909848484849,
      "avg_duration_per_task_ms": 229156.16414141413,
      "avg_output_tokens_per_task": 15405.453282828283,
      "last_run_timestamp": "2026-09-19T08:48:23.277Z"
    },
    {
      "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": 0.8219452102185287,
      "accuracy_std_dev": 0.034130220185287316,
      "run_count": 17,
      "total_questions": 3249,
      "avg_cost_per_task_usd": 0.08630355401662049,
      "avg_duration_per_task_ms": 387562.971683595,
      "avg_output_tokens_per_task": 22693.282856263464,
      "last_run_timestamp": "2026-08-22T16:49:15.579Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.8215312538860104,
      "accuracy_std_dev": 0.022946010202695863,
      "run_count": 3,
      "total_questions": 579,
      "avg_cost_per_task_usd": 0.08296633678756477,
      "avg_duration_per_task_ms": 239744.6183074266,
      "avg_output_tokens_per_task": 23823.246977547496,
      "last_run_timestamp": "2026-06-30T21:35:33.459Z"
    },
    {
      "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": 0.8214834459833794,
      "accuracy_std_dev": 0.011465056212986574,
      "run_count": 11,
      "total_questions": 2166,
      "avg_cost_per_task_usd": 0.005549661126500463,
      "avg_duration_per_task_ms": 487617.81809787627,
      "avg_output_tokens_per_task": 16572.74192059095,
      "last_run_timestamp": "2026-08-29T19:09:30.030Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Makora",
      "accuracy": 0.8213506405228759,
      "accuracy_std_dev": 0.02248216924587128,
      "run_count": 3,
      "total_questions": 459,
      "avg_cost_per_task_usd": 0.08758356862745097,
      "avg_duration_per_task_ms": 304618.9760348584,
      "avg_output_tokens_per_task": 34720.03050108933,
      "last_run_timestamp": "2026-08-29T10:57:24.797Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Parasail",
      "accuracy": 0.8212070710659898,
      "accuracy_std_dev": 0.011946281848340899,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.01807649915397631,
      "avg_duration_per_task_ms": 302940.0592216582,
      "avg_output_tokens_per_task": 16413.382402707277,
      "last_run_timestamp": "2026-08-29T09:59:35.603Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AkashML",
      "accuracy": 0.8210970962025316,
      "accuracy_std_dev": 0.02845442397599282,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.039231337130801694,
      "avg_duration_per_task_ms": 342482.6898734177,
      "avg_output_tokens_per_task": 40890.57890295359,
      "last_run_timestamp": "2026-09-19T16:14:18.749Z"
    },
    {
      "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": "Together",
      "accuracy": 0.8209874444444444,
      "accuracy_std_dev": 0.015449980784677811,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.07123941301907968,
      "avg_duration_per_task_ms": 202727.39393939395,
      "avg_output_tokens_per_task": 16398.18462401796,
      "last_run_timestamp": "2026-06-20T09:16:50.854Z"
    },
    {
      "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": 0.8209175000000001,
      "accuracy_std_dev": 0.025544439297036866,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.03782663825757576,
      "avg_duration_per_task_ms": 314167.50126262626,
      "avg_output_tokens_per_task": 39526.941287878784,
      "last_run_timestamp": "2026-09-19T16:25:45.279Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.8207071666666668,
      "accuracy_std_dev": 0.015360537180457651,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.008930407407407408,
      "avg_duration_per_task_ms": 167713.53535353535,
      "avg_output_tokens_per_task": 8891.933501683501,
      "last_run_timestamp": "2026-06-13T16:35:11.667Z"
    },
    {
      "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": 0.8206819729372937,
      "accuracy_std_dev": 0.01766511701862355,
      "run_count": 23,
      "total_questions": 4545,
      "avg_cost_per_task_usd": 0.004627064906490648,
      "avg_duration_per_task_ms": 269164.16699669964,
      "avg_output_tokens_per_task": 11500.384598459847,
      "last_run_timestamp": "2026-08-29T19:19:45.463Z"
    },
    {
      "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": 0.8204619306930694,
      "accuracy_std_dev": 0.012993991694579071,
      "run_count": 23,
      "total_questions": 4545,
      "avg_cost_per_task_usd": 0.011459436963696373,
      "avg_duration_per_task_ms": 354916.7643564356,
      "avg_output_tokens_per_task": 11342.301650165016,
      "last_run_timestamp": "2026-08-30T00:16:38.394Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "Xiaomi",
      "accuracy": 0.8204264444444443,
      "accuracy_std_dev": 0.015265342891014489,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0033981470258136926,
      "avg_duration_per_task_ms": 86880.46913580247,
      "avg_output_tokens_per_task": 11274.21492704826,
      "last_run_timestamp": "2026-06-13T08:05:57.280Z"
    },
    {
      "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": 0.8203543500992719,
      "accuracy_std_dev": 0.047076899102709124,
      "run_count": 23,
      "total_questions": 4533,
      "avg_cost_per_task_usd": 0.07291643834105449,
      "avg_duration_per_task_ms": 483587.120891242,
      "avg_output_tokens_per_task": 28482.810500772117,
      "last_run_timestamp": "2026-08-29T19:11:37.135Z"
    },
    {
      "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": 0.8203043478260869,
      "accuracy_std_dev": 0.021973884023913672,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.004347063241106719,
      "avg_duration_per_task_ms": 319883.9997804128,
      "avg_output_tokens_per_task": 11111.253184014053,
      "last_run_timestamp": "2026-08-29T19:12:36.104Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "AkashML",
      "accuracy": 0.8201456666666667,
      "accuracy_std_dev": 0.021829497258424348,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.03759047053872054,
      "avg_duration_per_task_ms": 252471.02188552188,
      "avg_output_tokens_per_task": 14715.675084175084,
      "last_run_timestamp": "2026-09-19T08:47:29.837Z"
    },
    {
      "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": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Z.AI",
      "accuracy": 0.8198026983870969,
      "accuracy_std_dev": 0.05250346254129803,
      "run_count": 19,
      "total_questions": 3720,
      "avg_cost_per_task_usd": 0.10939475376344085,
      "avg_duration_per_task_ms": 699399.8586021506,
      "avg_output_tokens_per_task": 34933.458333333336,
      "last_run_timestamp": "2026-08-29T18:49:03.323Z"
    },
    {
      "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": "CoreWeave",
      "accuracy": 0.8195846666666666,
      "accuracy_std_dev": 0.016130608974520736,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.11832865488215488,
      "avg_duration_per_task_ms": 298591.92424242425,
      "avg_output_tokens_per_task": 48880.583333333336,
      "last_run_timestamp": "2026-09-19T08:45:26.530Z"
    },
    {
      "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": 0.819234,
      "accuracy_std_dev": 0.018593792606673883,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.011218165404040404,
      "avg_duration_per_task_ms": 587595.9393939395,
      "avg_output_tokens_per_task": 33133.1470959596,
      "last_run_timestamp": "2026-09-19T09:14:17.409Z"
    },
    {
      "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": 0.8192089661016949,
      "accuracy_std_dev": 0.16613254756324589,
      "run_count": 12,
      "total_questions": 2301,
      "avg_cost_per_task_usd": 0.08146611951325512,
      "avg_duration_per_task_ms": 457624.593220339,
      "avg_output_tokens_per_task": 22201.0308561495,
      "last_run_timestamp": "2026-09-19T09:05:59.509Z"
    },
    {
      "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": 0.8188576490610326,
      "accuracy_std_dev": 0.0725401109287967,
      "run_count": 13,
      "total_questions": 1704,
      "avg_cost_per_task_usd": 0.020530700117370893,
      "avg_duration_per_task_ms": 66110.05575117371,
      "avg_output_tokens_per_task": 20662.859154929578,
      "last_run_timestamp": "2026-08-29T19:44:46.543Z"
    },
    {
      "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": 0.8187381511299434,
      "accuracy_std_dev": 0.020219033179017607,
      "run_count": 11,
      "total_questions": 2124,
      "avg_cost_per_task_usd": 0.052499902542372884,
      "avg_duration_per_task_ms": 261839.05320150658,
      "avg_output_tokens_per_task": 21842.697740112995,
      "last_run_timestamp": "2026-08-01T17:52:55.742Z"
    },
    {
      "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": 0.818182,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14423544949494949,
      "avg_duration_per_task_ms": 273234.71717171714,
      "avg_output_tokens_per_task": 28813.959595959597,
      "last_run_timestamp": "2026-09-19T08:47:57.150Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8181816666666667,
      "accuracy_std_dev": 0.08468970304522794,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.007531994107744108,
      "avg_duration_per_task_ms": 771784.1237373737,
      "avg_output_tokens_per_task": 30879.069865319867,
      "last_run_timestamp": "2026-09-19T13:19:37.492Z"
    },
    {
      "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": 0.8180828300653594,
      "accuracy_std_dev": 0.1016204485105719,
      "run_count": 14,
      "total_questions": 2754,
      "avg_cost_per_task_usd": 0.04520774546114742,
      "avg_duration_per_task_ms": 188140.4183006536,
      "avg_output_tokens_per_task": 21641.156136528687,
      "last_run_timestamp": "2026-09-19T08:53:56.863Z"
    },
    {
      "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": 0.8180825764931331,
      "accuracy_std_dev": 0.05580919908720086,
      "run_count": 33,
      "total_questions": 6262,
      "avg_cost_per_task_usd": 0.005215132864899393,
      "avg_duration_per_task_ms": 358805.9053018205,
      "avg_output_tokens_per_task": 12457.650590865538,
      "last_run_timestamp": "2026-08-29T20:12:46.798Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Baidu",
      "accuracy": 0.8172269616858236,
      "accuracy_std_dev": 0.059023830851696865,
      "run_count": 12,
      "total_questions": 2349,
      "avg_cost_per_task_usd": 0.1289344618986803,
      "avg_duration_per_task_ms": 699975.4461472967,
      "avg_output_tokens_per_task": 43383.50404427416,
      "last_run_timestamp": "2026-08-29T19:22:50.890Z"
    },
    {
      "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": 0.8169413157894736,
      "accuracy_std_dev": 0.012465325718516053,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.3349828732057416,
      "avg_duration_per_task_ms": 301917.04891015415,
      "avg_output_tokens_per_task": 21250.364167995747,
      "last_run_timestamp": "2026-09-19T08:51:02.860Z"
    },
    {
      "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": 0.8168373000310271,
      "accuracy_std_dev": 0.052236706041768964,
      "run_count": 19,
      "total_questions": 3223,
      "avg_cost_per_task_usd": 0.0374196242631089,
      "avg_duration_per_task_ms": 545799.8814768849,
      "avg_output_tokens_per_task": 31060.19795221843,
      "last_run_timestamp": "2026-08-29T19:07:57.500Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.8164983333333333,
      "accuracy_std_dev": 0.010378007269649053,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.002570441077441077,
      "avg_duration_per_task_ms": 557971.7020202021,
      "avg_output_tokens_per_task": 4954.074074074074,
      "last_run_timestamp": "2026-06-13T08:04:34.111Z"
    },
    {
      "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": 0.8161555499021526,
      "accuracy_std_dev": 0.033245258304521796,
      "run_count": 8,
      "total_questions": 1533,
      "avg_cost_per_task_usd": 0.01330266210045662,
      "avg_duration_per_task_ms": 93162.18590998043,
      "avg_output_tokens_per_task": 17543.737769080235,
      "last_run_timestamp": "2026-08-29T18:36:28.592Z"
    },
    {
      "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": 0.8161240555555556,
      "accuracy_std_dev": 0.03828528953567215,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.018823198372615035,
      "avg_duration_per_task_ms": 237443.63439955105,
      "avg_output_tokens_per_task": 17323.3285634119,
      "last_run_timestamp": "2026-08-29T19:18:38.408Z"
    },
    {
      "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": 0.8159271462140993,
      "accuracy_std_dev": 0.011945687451440362,
      "run_count": 4,
      "total_questions": 766,
      "avg_cost_per_task_usd": 0.10126304177545693,
      "avg_duration_per_task_ms": 221714.96997389034,
      "avg_output_tokens_per_task": 28424.819843342037,
      "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-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8157936203059803,
      "accuracy_std_dev": 0.02454612802005358,
      "run_count": 11,
      "total_questions": 2157,
      "avg_cost_per_task_usd": 0.00473010709318498,
      "avg_duration_per_task_ms": 443979.5808993973,
      "avg_output_tokens_per_task": 11275.669448307835,
      "last_run_timestamp": "2026-08-01T18:25:37.046Z"
    },
    {
      "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": 0.8157768571428571,
      "accuracy_std_dev": 0.021818009359429742,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.04211656493506493,
      "avg_duration_per_task_ms": 262574.5,
      "avg_output_tokens_per_task": 14107.66883116883,
      "last_run_timestamp": "2026-09-19T08:55:42.602Z"
    },
    {
      "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": 0.8157223200612557,
      "accuracy_std_dev": 0.0849169018041167,
      "run_count": 10,
      "total_questions": 1959,
      "avg_cost_per_task_usd": 0.10062858907605922,
      "avg_duration_per_task_ms": 941154.5890760592,
      "avg_output_tokens_per_task": 42235.08728943339,
      "last_run_timestamp": "2026-09-19T08:57:20.373Z"
    },
    {
      "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": 0.8148881320132014,
      "accuracy_std_dev": 0.01824134523693167,
      "run_count": 23,
      "total_questions": 4545,
      "avg_cost_per_task_usd": 0.06443867590759077,
      "avg_duration_per_task_ms": 437849.9201320132,
      "avg_output_tokens_per_task": 32018.359075907592,
      "last_run_timestamp": "2026-08-29T19:17:49.455Z"
    },
    {
      "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": 0.8146043749999999,
      "accuracy_std_dev": 0.010793946798292766,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.36479117045454545,
      "avg_duration_per_task_ms": 359805.747790404,
      "avg_output_tokens_per_task": 22977.79356060606,
      "last_run_timestamp": "2026-09-19T08:53:08.894Z"
    },
    {
      "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": 0.8144500956822106,
      "accuracy_std_dev": 0.08783913519448809,
      "run_count": 15,
      "total_questions": 2895,
      "avg_cost_per_task_usd": 0.018499234196891194,
      "avg_duration_per_task_ms": 92763.99827288429,
      "avg_output_tokens_per_task": 15429.552677029362,
      "last_run_timestamp": "2026-07-18T08:31:24.271Z"
    },
    {
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Z.AI",
      "accuracy": 0.8141788521373509,
      "accuracy_std_dev": 0.09118468393428555,
      "run_count": 15,
      "total_questions": 2854,
      "avg_cost_per_task_usd": 0.05950021793973371,
      "avg_duration_per_task_ms": 877552.4621583743,
      "avg_output_tokens_per_task": 27412.963910301332,
      "last_run_timestamp": "2026-08-29T19:12:33.941Z"
    },
    {
      "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": 0.813853,
      "accuracy_std_dev": 0.0071343829445860135,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.025065300144300147,
      "avg_duration_per_task_ms": 187376.2683982684,
      "avg_output_tokens_per_task": 16614.633477633477,
      "last_run_timestamp": "2026-09-19T08:53:39.948Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.813825179389313,
      "accuracy_std_dev": 0.015856752133850578,
      "run_count": 8,
      "total_questions": 1572,
      "avg_cost_per_task_usd": 0.02502054643765903,
      "avg_duration_per_task_ms": 242441.94974554708,
      "avg_output_tokens_per_task": 10837.982824427481,
      "last_run_timestamp": "2026-07-01T04:42:33.719Z"
    },
    {
      "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": 0.8137254984162897,
      "accuracy_std_dev": 0.08405639786637797,
      "run_count": 24,
      "total_questions": 4420,
      "avg_cost_per_task_usd": 0.010863166515837101,
      "avg_duration_per_task_ms": 478036.79049773753,
      "avg_output_tokens_per_task": 11973.086425339367,
      "last_run_timestamp": "2026-08-29T18:29:51.235Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Chutes",
      "accuracy": 0.812989225241824,
      "accuracy_std_dev": 0.04159005714835348,
      "run_count": 11,
      "total_questions": 2171,
      "avg_cost_per_task_usd": 0.05707970474435743,
      "avg_duration_per_task_ms": 931231.9829571625,
      "avg_output_tokens_per_task": 29294.633348687243,
      "last_run_timestamp": "2026-09-19T17:02:33.762Z"
    },
    {
      "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": 0.8127893802083332,
      "accuracy_std_dev": 0.02323940439151189,
      "run_count": 12,
      "total_questions": 2304,
      "avg_cost_per_task_usd": 0.014347588107638889,
      "avg_duration_per_task_ms": 360100.9900173611,
      "avg_output_tokens_per_task": 13752.019097222223,
      "last_run_timestamp": "2026-08-29T16:10:33.070Z"
    },
    {
      "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": 0.8116520828488373,
      "accuracy_std_dev": 0.02238695596611053,
      "run_count": 7,
      "total_questions": 1376,
      "avg_cost_per_task_usd": 0.021895043604651165,
      "avg_duration_per_task_ms": 295867.60247093026,
      "avg_output_tokens_per_task": 18000.89898255814,
      "last_run_timestamp": "2026-09-19T08:44:55.432Z"
    },
    {
      "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": 0.8111882789927104,
      "accuracy_std_dev": 0.14234185480055578,
      "run_count": 16,
      "total_questions": 3018,
      "avg_cost_per_task_usd": 0.003854944665341285,
      "avg_duration_per_task_ms": 404801.41351888666,
      "avg_output_tokens_per_task": 9504.889993373095,
      "last_run_timestamp": "2026-08-29T18:15:38.703Z"
    },
    {
      "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": "AkashML",
      "accuracy": 0.8109296153846153,
      "accuracy_std_dev": 0.14891345125817648,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.008492456099456099,
      "avg_duration_per_task_ms": 161496.31857031857,
      "avg_output_tokens_per_task": 7371.756021756022,
      "last_run_timestamp": "2026-07-01T03:19:17.846Z"
    },
    {
      "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": "WandB",
      "accuracy": 0.8106060000000002,
      "accuracy_std_dev": 0.016940078320362043,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.000941689393939394,
      "avg_duration_per_task_ms": 123178.65404040404,
      "avg_output_tokens_per_task": 9148.02398989899,
      "last_run_timestamp": "2026-06-20T16:49:31.612Z"
    },
    {
      "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": 0.8100816932907348,
      "accuracy_std_dev": 0.08830772854780453,
      "run_count": 11,
      "total_questions": 1878,
      "avg_cost_per_task_usd": 0.08196660915867944,
      "avg_duration_per_task_ms": 525045.6922257721,
      "avg_output_tokens_per_task": 25551.769968051118,
      "last_run_timestamp": "2026-09-19T16:07:59.654Z"
    },
    {
      "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": 0.8097645,
      "accuracy_std_dev": 0.0622895,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.034729030303030296,
      "avg_duration_per_task_ms": 253034.54292929292,
      "avg_output_tokens_per_task": 31317.588383838385,
      "last_run_timestamp": "2026-09-15T01:07:48.063Z"
    },
    {
      "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": 0.8093432500000001,
      "accuracy_std_dev": 0.008279566757234316,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.03340588888888889,
      "avg_duration_per_task_ms": 671635.9166666666,
      "avg_output_tokens_per_task": 15467.433080808081,
      "last_run_timestamp": "2026-09-19T09:35:13.634Z"
    },
    {
      "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": 0.8091227713684209,
      "accuracy_std_dev": 0.08205896778163711,
      "run_count": 14,
      "total_questions": 2375,
      "avg_cost_per_task_usd": 0.0624333372631579,
      "avg_duration_per_task_ms": 697771.9258947369,
      "avg_output_tokens_per_task": 19861.26063157895,
      "last_run_timestamp": "2026-08-01T18:16:39.158Z"
    },
    {
      "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": 0.8089819438325991,
      "accuracy_std_dev": 0.06057756732336092,
      "run_count": 14,
      "total_questions": 2724,
      "avg_cost_per_task_usd": 0.15691310939794417,
      "avg_duration_per_task_ms": 677800.3755506608,
      "avg_output_tokens_per_task": 35807.34140969163,
      "last_run_timestamp": "2026-08-29T18:07:06.751Z"
    },
    {
      "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": 0.808541063084112,
      "accuracy_std_dev": 0.015819760309078213,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.011975492601246106,
      "avg_duration_per_task_ms": 326659.34890965733,
      "avg_output_tokens_per_task": 12473.692757009347,
      "last_run_timestamp": "2026-08-30T00:01:09.438Z"
    },
    {
      "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": 0.8079965500000003,
      "accuracy_std_dev": 0.018621334421772776,
      "run_count": 20,
      "total_questions": 3960,
      "avg_cost_per_task_usd": 0.01976742373737374,
      "avg_duration_per_task_ms": 154039.40732323233,
      "avg_output_tokens_per_task": 15688.315909090908,
      "last_run_timestamp": "2026-08-29T19:24:09.844Z"
    },
    {
      "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": 0.8079122642559109,
      "accuracy_std_dev": 0.15281497299086352,
      "run_count": 11,
      "total_questions": 2157,
      "avg_cost_per_task_usd": 0.062700123783032,
      "avg_duration_per_task_ms": 394470.919796013,
      "avg_output_tokens_per_task": 17556.69772832638,
      "last_run_timestamp": "2026-09-19T09:15:56.025Z"
    },
    {
      "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": 0.8076977522123893,
      "accuracy_std_dev": 0.09983296230241988,
      "run_count": 12,
      "total_questions": 2373,
      "avg_cost_per_task_usd": 0.08567701222081754,
      "avg_duration_per_task_ms": 492014.607248209,
      "avg_output_tokens_per_task": 31610.124736620313,
      "last_run_timestamp": "2026-09-19T16:24:52.704Z"
    },
    {
      "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": 0.8075196666666667,
      "accuracy_std_dev": 0.04008333166735965,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.16043005976430974,
      "avg_duration_per_task_ms": 159846.56397306398,
      "avg_output_tokens_per_task": 15912.45622895623,
      "last_run_timestamp": "2026-09-19T08:44:53.138Z"
    },
    {
      "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": 0.8073155454545455,
      "accuracy_std_dev": 0.016232793247361044,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.3389105385674931,
      "avg_duration_per_task_ms": 354028.78879706154,
      "avg_output_tokens_per_task": 22430.448576675848,
      "last_run_timestamp": "2026-09-19T08:54:14.389Z"
    },
    {
      "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": 0.80698645,
      "accuracy_std_dev": 0.018297436846386447,
      "run_count": 20,
      "total_questions": 3960,
      "avg_cost_per_task_usd": 0.026827777777777784,
      "avg_duration_per_task_ms": 285701.8515151515,
      "avg_output_tokens_per_task": 15377.552777777777,
      "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-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8069215714285713,
      "accuracy_std_dev": 0.01495832495843468,
      "run_count": 13,
      "total_questions": 2562,
      "avg_cost_per_task_usd": 0.011023931303669009,
      "avg_duration_per_task_ms": 293511.45472287276,
      "avg_output_tokens_per_task": 11473.932084309134,
      "last_run_timestamp": "2026-08-29T19:12:57.906Z"
    },
    {
      "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": "Together",
      "accuracy": 0.806818,
      "accuracy_std_dev": 0.027977996970834047,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.04117596338383839,
      "avg_duration_per_task_ms": 339680.07196969696,
      "avg_output_tokens_per_task": 15286.914141414141,
      "last_run_timestamp": "2026-05-16T10:24:57.481Z"
    },
    {
      "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": "Fireworks",
      "accuracy": 0.8063973333333334,
      "accuracy_std_dev": 0.07629808423472303,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0736860505050505,
      "avg_duration_per_task_ms": 283613.16077441076,
      "avg_output_tokens_per_task": 18366.69107744108,
      "last_run_timestamp": "2026-05-23T11:07:21.864Z"
    },
    {
      "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.8060605000000001,
      "accuracy_std_dev": 0.034859352280987646,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.003273143939393939,
      "avg_duration_per_task_ms": 84028.83181818182,
      "avg_output_tokens_per_task": 10856.877272727274,
      "last_run_timestamp": "2026-06-13T08:37:07.102Z"
    },
    {
      "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": 0.8058781554959786,
      "accuracy_std_dev": 0.01700166220410292,
      "run_count": 17,
      "total_questions": 3357,
      "avg_cost_per_task_usd": 0.014998186773905272,
      "avg_duration_per_task_ms": 223628.16324098897,
      "avg_output_tokens_per_task": 13349.141197497765,
      "last_run_timestamp": "2026-08-29T18:59:02.809Z"
    },
    {
      "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": 0.805836,
      "accuracy_std_dev": 0.0155108889386349,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.022988649831649832,
      "avg_duration_per_task_ms": 268127.5387205387,
      "avg_output_tokens_per_task": 17683.01936026936,
      "last_run_timestamp": "2026-09-19T08:47:58.850Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Chutes",
      "accuracy": 0.805556,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 192,
      "avg_cost_per_task_usd": 0.16585933333333333,
      "avg_duration_per_task_ms": 688743.8177083334,
      "avg_output_tokens_per_task": 42426.447916666664,
      "last_run_timestamp": "2026-07-31T00:21:35.641Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DekaLLM",
      "accuracy": 0.805555546875,
      "accuracy_std_dev": 0.01288499999999998,
      "run_count": 2,
      "total_questions": 384,
      "avg_cost_per_task_usd": 0.0362872109375,
      "avg_duration_per_task_ms": 729621.7864583334,
      "avg_output_tokens_per_task": 36271.864583333336,
      "last_run_timestamp": "2026-09-12T16:45:54.034Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8055555000000001,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.01909540909090909,
      "avg_duration_per_task_ms": 512069.0252525252,
      "avg_output_tokens_per_task": 9996.752525252525,
      "last_run_timestamp": "2026-05-09T11:21:48.466Z"
    },
    {
      "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": "DekaLLM",
      "accuracy": 0.8055555,
      "accuracy_std_dev": 0.012626499999999985,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0015357878787878787,
      "avg_duration_per_task_ms": 249503.77525252526,
      "avg_output_tokens_per_task": 825.6868686868687,
      "last_run_timestamp": "2026-05-02T20:26:38.249Z"
    },
    {
      "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": 0.8052697686375321,
      "accuracy_std_dev": 0.05350092420872278,
      "run_count": 4,
      "total_questions": 778,
      "avg_cost_per_task_usd": 0.08564073007712082,
      "avg_duration_per_task_ms": 699962.8868894601,
      "avg_output_tokens_per_task": 25474.38560411311,
      "last_run_timestamp": "2026-09-05T09:27:38.360Z"
    },
    {
      "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": 0.8052215698062792,
      "accuracy_std_dev": 0.07724919018062586,
      "run_count": 16,
      "total_questions": 2994,
      "avg_cost_per_task_usd": 0.1436293523714095,
      "avg_duration_per_task_ms": 319160.0223780895,
      "avg_output_tokens_per_task": 29533.40046760187,
      "last_run_timestamp": "2026-09-19T09:11:52.131Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.8045346501901139,
      "accuracy_std_dev": 0.045705742477164614,
      "run_count": 12,
      "total_questions": 2367,
      "avg_cost_per_task_usd": 0.07353121546261089,
      "avg_duration_per_task_ms": 676893.81791297,
      "avg_output_tokens_per_task": 28532.3481199831,
      "last_run_timestamp": "2026-09-19T16:47:45.805Z"
    },
    {
      "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": 0.803999108108108,
      "accuracy_std_dev": 0.01725557388613565,
      "run_count": 23,
      "total_questions": 4551,
      "avg_cost_per_task_usd": 0.016347447374203468,
      "avg_duration_per_task_ms": 162155.09140848165,
      "avg_output_tokens_per_task": 10785.491100856954,
      "last_run_timestamp": "2026-08-29T18:07:47.743Z"
    },
    {
      "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": 0.8039509370988447,
      "accuracy_std_dev": 0.04606482842170186,
      "run_count": 12,
      "total_questions": 2337,
      "avg_cost_per_task_usd": 0.004562442875481386,
      "avg_duration_per_task_ms": 529537.7351305092,
      "avg_output_tokens_per_task": 12022.224646983312,
      "last_run_timestamp": "2026-08-29T19:15:43.723Z"
    },
    {
      "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": 0.8036616249999999,
      "accuracy_std_dev": 0.02419106182010981,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.338616759469697,
      "avg_duration_per_task_ms": 355080.81313131313,
      "avg_output_tokens_per_task": 22428.503156565657,
      "last_run_timestamp": "2026-09-19T08:51:20.836Z"
    },
    {
      "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": 0.8027495,
      "accuracy_std_dev": 0.02711787599075562,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.3324898122895622,
      "avg_duration_per_task_ms": 351546.73316498316,
      "avg_output_tokens_per_task": 22079.203703703704,
      "last_run_timestamp": "2026-09-19T08:49:06.868Z"
    },
    {
      "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": 0.8025378463620981,
      "accuracy_std_dev": 0.137748568405128,
      "run_count": 15,
      "total_questions": 2955,
      "avg_cost_per_task_usd": 0.004375704230118444,
      "avg_duration_per_task_ms": 379301.62267343485,
      "avg_output_tokens_per_task": 10823.406091370558,
      "last_run_timestamp": "2026-08-29T19:18:32.746Z"
    },
    {
      "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": 0.8025315620253164,
      "accuracy_std_dev": 0.041046181356000994,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.0376030194092827,
      "avg_duration_per_task_ms": 384444.73713080166,
      "avg_output_tokens_per_task": 37258.89915611815,
      "last_run_timestamp": "2026-09-19T16:23:22.280Z"
    },
    {
      "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": 0.8023188347826088,
      "accuracy_std_dev": 0.008868734307792898,
      "run_count": 3,
      "total_questions": 575,
      "avg_cost_per_task_usd": 0.05968593391304348,
      "avg_duration_per_task_ms": 239074.65391304347,
      "avg_output_tokens_per_task": 15240.28,
      "last_run_timestamp": "2026-09-19T08:48:26.677Z"
    },
    {
      "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": 0.8022664747948416,
      "accuracy_std_dev": 0.03416052727843685,
      "run_count": 13,
      "total_questions": 2559,
      "avg_cost_per_task_usd": 0.10685159788980071,
      "avg_duration_per_task_ms": 697410.7983587339,
      "avg_output_tokens_per_task": 34433.65064478312,
      "last_run_timestamp": "2026-08-29T18:40:24.764Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Novita",
      "accuracy": 0.8020385792266869,
      "accuracy_std_dev": 0.0530296415355365,
      "run_count": 20,
      "total_questions": 3957,
      "avg_cost_per_task_usd": 0.003907585039171089,
      "avg_duration_per_task_ms": 215449.7402072277,
      "avg_output_tokens_per_task": 9523.579732120294,
      "last_run_timestamp": "2026-08-29T19:20:50.265Z"
    },
    {
      "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": 0.8019233052236256,
      "accuracy_std_dev": 0.09230759113302772,
      "run_count": 15,
      "total_questions": 2929,
      "avg_cost_per_task_usd": 0.010292674291567086,
      "avg_duration_per_task_ms": 247238.68316831684,
      "avg_output_tokens_per_task": 8923.839877091157,
      "last_run_timestamp": "2026-07-11T15:51:15.190Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Io Net",
      "accuracy": 0.8016302663043479,
      "accuracy_std_dev": 0.016866645256903415,
      "run_count": 6,
      "total_questions": 1104,
      "avg_cost_per_task_usd": 0.053515479166666664,
      "avg_duration_per_task_ms": 272256.60507246375,
      "avg_output_tokens_per_task": 12863.78713768116,
      "last_run_timestamp": "2026-07-18T08:30:38.228Z"
    },
    {
      "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": 0.8015763636363635,
      "accuracy_std_dev": 0.05357946338302784,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.143211629476584,
      "avg_duration_per_task_ms": 721714.1303948576,
      "avg_output_tokens_per_task": 46550.70293847567,
      "last_run_timestamp": "2026-09-19T09:25:42.553Z"
    },
    {
      "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": 0.80138718727491,
      "accuracy_std_dev": 0.013694984002920067,
      "run_count": 13,
      "total_questions": 2499,
      "avg_cost_per_task_usd": 0.009386521408563426,
      "avg_duration_per_task_ms": 310251.2785114046,
      "avg_output_tokens_per_task": 9778.357743097238,
      "last_run_timestamp": "2026-08-15T14:56:55.269Z"
    },
    {
      "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "auto-routing",
      "accuracy": 0.8011413739967898,
      "accuracy_std_dev": 0.026788015723950006,
      "run_count": 19,
      "total_questions": 3738,
      "avg_cost_per_task_usd": 0.010181774745853397,
      "avg_duration_per_task_ms": 304238.04440877476,
      "avg_output_tokens_per_task": 10343.559390048154,
      "last_run_timestamp": "2026-08-29T18:21:38.666Z"
    },
    {
      "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": 0.80042085,
      "accuracy_std_dev": 0.09910293924565254,
      "run_count": 20,
      "total_questions": 3960,
      "avg_cost_per_task_usd": 0.02063269419191919,
      "avg_duration_per_task_ms": 99441.73484848485,
      "avg_output_tokens_per_task": 15808.599242424243,
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Sail Research",
      "accuracy": 0.799896012993763,
      "accuracy_std_dev": 0.11770242543873084,
      "run_count": 10,
      "total_questions": 1924,
      "avg_cost_per_task_usd": 0.006078587837837837,
      "avg_duration_per_task_ms": 630170.6242203743,
      "avg_output_tokens_per_task": 26380.861226611225,
      "last_run_timestamp": "2026-09-19T09:04:04.808Z"
    },
    {
      "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": 0.7993420703124999,
      "accuracy_std_dev": 0.19162084340952992,
      "run_count": 14,
      "total_questions": 2432,
      "avg_cost_per_task_usd": 0.16367626192434215,
      "avg_duration_per_task_ms": 295251.3548519737,
      "avg_output_tokens_per_task": 28199.261924342107,
      "last_run_timestamp": "2026-09-20T10:34:11.013Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Parasail",
      "accuracy": 0.7984771862955031,
      "accuracy_std_dev": 0.15755601228064434,
      "run_count": 17,
      "total_questions": 2802,
      "avg_cost_per_task_usd": 0.07837114346895076,
      "avg_duration_per_task_ms": 375817.0189150607,
      "avg_output_tokens_per_task": 18242.825124910778,
      "last_run_timestamp": "2026-08-22T17:29:49.714Z"
    },
    {
      "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-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7976528739352641,
      "accuracy_std_dev": 0.01137849398787464,
      "run_count": 18,
      "total_questions": 3522,
      "avg_cost_per_task_usd": 0.03818711385576377,
      "avg_duration_per_task_ms": 869626.9395229983,
      "avg_output_tokens_per_task": 17421.532935831914,
      "last_run_timestamp": "2026-08-29T20:02:23.660Z"
    },
    {
      "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": 0.79727825,
      "accuracy_std_dev": 0.08339832626730288,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.08855118055555555,
      "avg_duration_per_task_ms": 505679.5707070707,
      "avg_output_tokens_per_task": 28090.66245791246,
      "last_run_timestamp": "2026-09-19T16:34:54.602Z"
    },
    {
      "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": 0.7971869670886074,
      "accuracy_std_dev": 0.03513963815261782,
      "run_count": 18,
      "total_questions": 3555,
      "avg_cost_per_task_usd": 0.031226099296765116,
      "avg_duration_per_task_ms": 382925.899859353,
      "avg_output_tokens_per_task": 15698.530239099859,
      "last_run_timestamp": "2026-08-22T17:09:41.940Z"
    },
    {
      "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": 0.7969371849642005,
      "accuracy_std_dev": 0.11399068476884858,
      "run_count": 13,
      "total_questions": 2514,
      "avg_cost_per_task_usd": 0.1032379212410501,
      "avg_duration_per_task_ms": 717491.7549721559,
      "avg_output_tokens_per_task": 30670.05369928401,
      "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-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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Cloudflare",
      "accuracy": 0.796213666369313,
      "accuracy_std_dev": 0.06626010818847784,
      "run_count": 17,
      "total_questions": 3363,
      "avg_cost_per_task_usd": 0.07408058043413619,
      "avg_duration_per_task_ms": 465867.30954504904,
      "avg_output_tokens_per_task": 23193.086827237585,
      "last_run_timestamp": "2026-08-22T17:17:31.993Z"
    },
    {
      "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": 0.7956016344772546,
      "accuracy_std_dev": 0.026298589487996727,
      "run_count": 19,
      "total_questions": 3759,
      "avg_cost_per_task_usd": 0.009070496408619311,
      "avg_duration_per_task_ms": 398032.09071561584,
      "avg_output_tokens_per_task": 22891.514232508645,
      "last_run_timestamp": "2026-08-29T18:49:24.022Z"
    },
    {
      "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": 0.7955027719182596,
      "accuracy_std_dev": 0.03582202847933504,
      "run_count": 23,
      "total_questions": 4551,
      "avg_cost_per_task_usd": 0.010505668424522085,
      "avg_duration_per_task_ms": 330984.9740716326,
      "avg_output_tokens_per_task": 10159.019116677653,
      "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-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DigitalOcean",
      "accuracy": 0.795170489361702,
      "accuracy_std_dev": 0.03970564182833467,
      "run_count": 10,
      "total_questions": 1974,
      "avg_cost_per_task_usd": 0.03361843617021277,
      "avg_duration_per_task_ms": 752189.3389057751,
      "avg_output_tokens_per_task": 19508.68237082067,
      "last_run_timestamp": "2026-09-19T08:51:21.034Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "WandB",
      "accuracy": 0.794507625,
      "accuracy_std_dev": 0.01694477015879457,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.010556673295454547,
      "avg_duration_per_task_ms": 162269.5918560606,
      "avg_output_tokens_per_task": 6275.243371212121,
      "last_run_timestamp": "2026-07-18T16:16:42.667Z"
    },
    {
      "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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Parasail",
      "accuracy": 0.7929361286060933,
      "accuracy_std_dev": 0.09695873156687938,
      "run_count": 19,
      "total_questions": 3709,
      "avg_cost_per_task_usd": 0.015247492315988135,
      "avg_duration_per_task_ms": 206557.6931787544,
      "avg_output_tokens_per_task": 15414.885144243732,
      "last_run_timestamp": "2026-08-29T20:00:43.878Z"
    },
    {
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.7928263074433656,
      "accuracy_std_dev": 0.023434100042370905,
      "run_count": 17,
      "total_questions": 3090,
      "avg_cost_per_task_usd": 0.03394864174757281,
      "avg_duration_per_task_ms": 579008.496763754,
      "avg_output_tokens_per_task": 18524.763754045307,
      "last_run_timestamp": "2026-08-29T18:31:11.241Z"
    },
    {
      "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": 0.7910053138664271,
      "accuracy_std_dev": 0.15895006299418277,
      "run_count": 17,
      "total_questions": 3339,
      "avg_cost_per_task_usd": 0.011569411799940102,
      "avg_duration_per_task_ms": 139037.51362683438,
      "avg_output_tokens_per_task": 9589.577118897874,
      "last_run_timestamp": "2026-08-01T18:21:26.176Z"
    },
    {
      "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": 0.7909605118644066,
      "accuracy_std_dev": 0.012465345763974308,
      "run_count": 18,
      "total_questions": 3540,
      "avg_cost_per_task_usd": 0.03681927146892655,
      "avg_duration_per_task_ms": 671335.5409604519,
      "avg_output_tokens_per_task": 16802.62627118644,
      "last_run_timestamp": "2026-08-29T19:47:52.368Z"
    },
    {
      "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": 0.7906814256756757,
      "accuracy_std_dev": 0.02348428081895062,
      "run_count": 6,
      "total_questions": 1184,
      "avg_cost_per_task_usd": 0.059051968749999996,
      "avg_duration_per_task_ms": 165937.2972972973,
      "avg_output_tokens_per_task": 18641.25929054054,
      "last_run_timestamp": "2026-09-19T09:21:23.718Z"
    },
    {
      "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": "Fireworks",
      "accuracy": 0.7904040000000001,
      "accuracy_std_dev": 0.013598748159297637,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.006979520202020202,
      "avg_duration_per_task_ms": 73170.09974747474,
      "avg_output_tokens_per_task": 4078.4949494949497,
      "last_run_timestamp": "2026-05-09T16:48:20.198Z"
    },
    {
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.790014,
      "accuracy_std_dev": 0.07589864878664765,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.016462866666666666,
      "avg_duration_per_task_ms": 474335.2464135021,
      "avg_output_tokens_per_task": 17867.735021097047,
      "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": "AtlasCloud",
      "accuracy": 0.7887197076142133,
      "accuracy_std_dev": 0.015079397638131601,
      "run_count": 15,
      "total_questions": 2955,
      "avg_cost_per_task_usd": 0.12096143113367173,
      "avg_duration_per_task_ms": 644827.7421319797,
      "avg_output_tokens_per_task": 30231.451776649745,
      "last_run_timestamp": "2026-08-29T20:08:40.362Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Morph",
      "accuracy": 0.7883901041069723,
      "accuracy_std_dev": 0.19387690688581088,
      "run_count": 16,
      "total_questions": 3141,
      "avg_cost_per_task_usd": 0.02842944317096465,
      "avg_duration_per_task_ms": 121865.52976758993,
      "avg_output_tokens_per_task": 8107.963387456224,
      "last_run_timestamp": "2026-07-04T18:14:56.425Z"
    },
    {
      "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": 0.7882558207964601,
      "accuracy_std_dev": 0.04689700002782695,
      "run_count": 10,
      "total_questions": 1808,
      "avg_cost_per_task_usd": 0.00581368860619469,
      "avg_duration_per_task_ms": 499550.75,
      "avg_output_tokens_per_task": 13199.183628318584,
      "last_run_timestamp": "2026-08-29T19:08:46.030Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.7879646298811546,
      "accuracy_std_dev": 0.023042252221345448,
      "run_count": 9,
      "total_questions": 1767,
      "avg_cost_per_task_usd": 0.023100724391624222,
      "avg_duration_per_task_ms": 400432.5291454443,
      "avg_output_tokens_per_task": 10710.642331635541,
      "last_run_timestamp": "2026-07-01T04:44:07.324Z"
    },
    {
      "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": 0.7879190221533008,
      "accuracy_std_dev": 0.05144838052109178,
      "run_count": 12,
      "total_questions": 2257,
      "avg_cost_per_task_usd": 0.04170270757642889,
      "avg_duration_per_task_ms": 437887.0739920248,
      "avg_output_tokens_per_task": 20993.67213114754,
      "last_run_timestamp": "2026-09-19T10:02:08.258Z"
    },
    {
      "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": "Chutes",
      "accuracy": 0.787879,
      "accuracy_std_dev": 0.010101000000000027,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.004192080808080808,
      "avg_duration_per_task_ms": 367628.1565656566,
      "avg_output_tokens_per_task": 4131.532828282829,
      "last_run_timestamp": "2026-05-02T19:00:01.558Z"
    },
    {
      "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": 0.7873578066075747,
      "accuracy_std_dev": 0.03703259018651595,
      "run_count": 19,
      "total_questions": 3723,
      "avg_cost_per_task_usd": 0.06184803760408272,
      "avg_duration_per_task_ms": 348744.5423045931,
      "avg_output_tokens_per_task": 18125.925597636317,
      "last_run_timestamp": "2026-08-29T18:54:50.382Z"
    },
    {
      "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": 0.7867736181619257,
      "accuracy_std_dev": 0.03349025786215252,
      "run_count": 14,
      "total_questions": 2742,
      "avg_cost_per_task_usd": 0.004660726112326768,
      "avg_duration_per_task_ms": 483380.77534646244,
      "avg_output_tokens_per_task": 11108.198395331874,
      "last_run_timestamp": "2026-08-01T18:07:34.003Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "NextBit",
      "accuracy": 0.7863751692015208,
      "accuracy_std_dev": 0.12285667045366319,
      "run_count": 15,
      "total_questions": 2630,
      "avg_cost_per_task_usd": 0.0199307855513308,
      "avg_duration_per_task_ms": 423150.044486692,
      "avg_output_tokens_per_task": 12234.871102661597,
      "last_run_timestamp": "2026-08-01T23:27:37.975Z"
    },
    {
      "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": 0.7862293764980027,
      "accuracy_std_dev": 0.1666251748521213,
      "run_count": 16,
      "total_questions": 3004,
      "avg_cost_per_task_usd": 0.00942010652463382,
      "avg_duration_per_task_ms": 248834.9277629827,
      "avg_output_tokens_per_task": 7858.000332889481,
      "last_run_timestamp": "2026-07-18T16:53:38.018Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "WandB",
      "accuracy": 0.7861595487316422,
      "accuracy_std_dev": 0.02054339937029649,
      "run_count": 4,
      "total_questions": 749,
      "avg_cost_per_task_usd": 0.023644552736982648,
      "avg_duration_per_task_ms": 116214.79706275034,
      "avg_output_tokens_per_task": 18836.48865153538,
      "last_run_timestamp": "2026-07-18T08:04:57.082Z"
    },
    {
      "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": 0.7859751167002013,
      "accuracy_std_dev": 0.06046117739911878,
      "run_count": 23,
      "total_questions": 4473,
      "avg_cost_per_task_usd": 0.07098338184663536,
      "avg_duration_per_task_ms": 540645.976302258,
      "avg_output_tokens_per_task": 31720.033310976974,
      "last_run_timestamp": "2026-08-29T18:23:09.250Z"
    },
    {
      "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": 0.785920719858156,
      "accuracy_std_dev": 0.021394714621875175,
      "run_count": 13,
      "total_questions": 2538,
      "avg_cost_per_task_usd": 0.04721227226162333,
      "avg_duration_per_task_ms": 569592.5594956658,
      "avg_output_tokens_per_task": 20532.268321513002,
      "last_run_timestamp": "2026-08-29T16:48:15.193Z"
    },
    {
      "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": 0.7856621441048034,
      "accuracy_std_dev": 0.00664559449898186,
      "run_count": 7,
      "total_questions": 1374,
      "avg_cost_per_task_usd": 0.018259436681222704,
      "avg_duration_per_task_ms": 150221.00873362445,
      "avg_output_tokens_per_task": 15062.550218340612,
      "last_run_timestamp": "2026-09-19T09:05:04.503Z"
    },
    {
      "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": "Wafer",
      "accuracy": 0.7853535,
      "accuracy_std_dev": 0.02272750000000001,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.002243229797979798,
      "avg_duration_per_task_ms": 474421.00757575757,
      "avg_output_tokens_per_task": 14645.75,
      "last_run_timestamp": "2026-06-20T12:44:34.391Z"
    },
    {
      "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": 0.7829818899297424,
      "accuracy_std_dev": 0.05788395901278767,
      "run_count": 13,
      "total_questions": 2562,
      "avg_cost_per_task_usd": 0.012435911787665888,
      "avg_duration_per_task_ms": 294718.2923497268,
      "avg_output_tokens_per_task": 12376.93169398907,
      "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-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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Friendli",
      "accuracy": 0.7820426666666666,
      "accuracy_std_dev": 0.18849682087457664,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.056621786195286204,
      "avg_duration_per_task_ms": 177410.45185185186,
      "avg_output_tokens_per_task": 17616.97272727273,
      "last_run_timestamp": "2026-07-01T03:02:14.071Z"
    },
    {
      "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": 0.7820227306590256,
      "accuracy_std_dev": 0.04532488089752691,
      "run_count": 16,
      "total_questions": 3141,
      "avg_cost_per_task_usd": 0.03846658134352117,
      "avg_duration_per_task_ms": 642155.4489016237,
      "avg_output_tokens_per_task": 20074.63387456224,
      "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-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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7810580866666665,
      "accuracy_std_dev": 0.02055413818726277,
      "run_count": 16,
      "total_questions": 3150,
      "avg_cost_per_task_usd": 0.04090850158730158,
      "avg_duration_per_task_ms": 425804.21428571426,
      "avg_output_tokens_per_task": 18703.222222222223,
      "last_run_timestamp": "2026-08-29T18:26:43.144Z"
    },
    {
      "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": "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-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.7805645299890948,
      "accuracy_std_dev": 0.02929019713259585,
      "run_count": 14,
      "total_questions": 2751,
      "avg_cost_per_task_usd": 0.004841514721919302,
      "avg_duration_per_task_ms": 633261.4282079244,
      "avg_output_tokens_per_task": 11578.931297709923,
      "last_run_timestamp": "2026-08-01T22:57:01.346Z"
    },
    {
      "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": "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": 0.7783505180412369,
      "accuracy_std_dev": 0.01799990454372838,
      "run_count": 12,
      "total_questions": 2328,
      "avg_cost_per_task_usd": 0.012542361254295534,
      "avg_duration_per_task_ms": 614270.4780927835,
      "avg_output_tokens_per_task": 31292.053694158076,
      "last_run_timestamp": "2026-08-29T16:45:34.440Z"
    },
    {
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "auto-routing",
      "accuracy": 0.7756354601542417,
      "accuracy_std_dev": 0.02561398969421896,
      "run_count": 12,
      "total_questions": 2334,
      "avg_cost_per_task_usd": 0.005398404884318766,
      "avg_duration_per_task_ms": 660748.7202227935,
      "avg_output_tokens_per_task": 33855.62853470437,
      "last_run_timestamp": "2026-09-19T17:12:41.248Z"
    },
    {
      "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": 0.7753727142857143,
      "accuracy_std_dev": 0.16596390492747365,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.01666139538239538,
      "avg_duration_per_task_ms": 580879.6435786436,
      "avg_output_tokens_per_task": 26434.363636363636,
      "last_run_timestamp": "2026-09-19T08:12:48.581Z"
    },
    {
      "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": 0.774832191940615,
      "accuracy_std_dev": 0.02630010557108088,
      "run_count": 10,
      "total_questions": 1886,
      "avg_cost_per_task_usd": 0.009038662778366914,
      "avg_duration_per_task_ms": 449340.6447507953,
      "avg_output_tokens_per_task": 36069.23966065748,
      "last_run_timestamp": "2026-09-19T17:14:53.367Z"
    },
    {
      "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": 0.7733209195128408,
      "accuracy_std_dev": 0.08870509019372362,
      "run_count": 21,
      "total_questions": 3777,
      "avg_cost_per_task_usd": 0.01586187847498014,
      "avg_duration_per_task_ms": 345375.779719354,
      "avg_output_tokens_per_task": 16140.08895949166,
      "last_run_timestamp": "2026-08-29T18:51:07.939Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7727855930680358,
      "accuracy_std_dev": 0.05706427864424032,
      "run_count": 12,
      "total_questions": 2337,
      "avg_cost_per_task_usd": 0.0018226653829696191,
      "avg_duration_per_task_ms": 478170.8801882756,
      "avg_output_tokens_per_task": 11197.934103551563,
      "last_run_timestamp": "2026-08-22T17:47:50.063Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Chutes",
      "accuracy": 0.7721462095719381,
      "accuracy_std_dev": 0.16044205147254612,
      "run_count": 18,
      "total_questions": 3364,
      "avg_cost_per_task_usd": 0.04585099791914388,
      "avg_duration_per_task_ms": 421607.292508918,
      "avg_output_tokens_per_task": 18189.730677764564,
      "last_run_timestamp": "2026-07-18T16:43:10.588Z"
    },
    {
      "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": 0.7721407260997067,
      "accuracy_std_dev": 0.028224906684324803,
      "run_count": 11,
      "total_questions": 1705,
      "avg_cost_per_task_usd": 0.009176458064516129,
      "avg_duration_per_task_ms": 41666.574780058654,
      "avg_output_tokens_per_task": 12170.292082111437,
      "last_run_timestamp": "2026-08-29T19:07:43.372Z"
    },
    {
      "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": 0.7715098846153846,
      "accuracy_std_dev": 0.09651313170593676,
      "run_count": 12,
      "total_questions": 2340,
      "avg_cost_per_task_usd": 0.09523880213675213,
      "avg_duration_per_task_ms": 1068926.4589743589,
      "avg_output_tokens_per_task": 33358.00341880342,
      "last_run_timestamp": "2026-08-29T18:41:58.319Z"
    },
    {
      "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": 0.7714939999999999,
      "accuracy_std_dev": 0.029762842455652656,
      "run_count": 5,
      "total_questions": 946,
      "avg_cost_per_task_usd": 0.04608473361522198,
      "avg_duration_per_task_ms": 257712.6194503171,
      "avg_output_tokens_per_task": 17355.710359408033,
      "last_run_timestamp": "2026-09-19T09:19:09.687Z"
    },
    {
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Venice",
      "accuracy": 0.7704213310344826,
      "accuracy_std_dev": 0.045856057411631505,
      "run_count": 11,
      "total_questions": 2175,
      "avg_cost_per_task_usd": 0.0051286381609195394,
      "avg_duration_per_task_ms": 279309.07172413793,
      "avg_output_tokens_per_task": 33671.028045977015,
      "last_run_timestamp": "2026-09-19T16:37:35.043Z"
    },
    {
      "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": 0.7701633959390862,
      "accuracy_std_dev": 0.05197684190953593,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.007365169204737733,
      "avg_duration_per_task_ms": 438945.5972927242,
      "avg_output_tokens_per_task": 14498.280879864637,
      "last_run_timestamp": "2026-09-19T08:57:55.272Z"
    },
    {
      "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": 0.7698924683870967,
      "accuracy_std_dev": 0.033487219191307406,
      "run_count": 12,
      "total_questions": 2325,
      "avg_cost_per_task_usd": 0.011860219354838712,
      "avg_duration_per_task_ms": 404374.8103225806,
      "avg_output_tokens_per_task": 29903.957849462364,
      "last_run_timestamp": "2026-08-29T16:43:11.936Z"
    },
    {
      "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": 0.7693605,
      "accuracy_std_dev": 0.13636349999999997,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0882362904040404,
      "avg_duration_per_task_ms": 1341629.558080808,
      "avg_output_tokens_per_task": 22000.722222222223,
      "last_run_timestamp": "2026-08-30T15:00:10.690Z"
    },
    {
      "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": "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Groq",
      "accuracy": 0.7680775714285716,
      "accuracy_std_dev": 0.014666364043105512,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.011148957431457432,
      "avg_duration_per_task_ms": 51087.674122174125,
      "avg_output_tokens_per_task": 18441.45286195286,
      "last_run_timestamp": "2026-08-29T18:36:51.876Z"
    },
    {
      "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": 0.7679573333333333,
      "accuracy_std_dev": 0.010259553071270784,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.016145424242424242,
      "avg_duration_per_task_ms": 98341.22727272728,
      "avg_output_tokens_per_task": 6033.04797979798,
      "last_run_timestamp": "2026-09-19T08:57:22.040Z"
    },
    {
      "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": 0.76767675,
      "accuracy_std_dev": 0.03241562151667774,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.012590233585858583,
      "avg_duration_per_task_ms": 407671.7121212121,
      "avg_output_tokens_per_task": 21713.55460858586,
      "last_run_timestamp": "2026-08-29T18:36:10.280Z"
    },
    {
      "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": 0.767378624023886,
      "accuracy_std_dev": 0.043028318846837645,
      "run_count": 22,
      "total_questions": 4354,
      "avg_cost_per_task_usd": 0.012152718649517686,
      "avg_duration_per_task_ms": 79257.83968764354,
      "avg_output_tokens_per_task": 7012.955672944419,
      "last_run_timestamp": "2026-08-22T17:12:08.919Z"
    },
    {
      "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": "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7668169148936171,
      "accuracy_std_dev": 0.053587367069901595,
      "run_count": 8,
      "total_questions": 1551,
      "avg_cost_per_task_usd": 0.010489795615731785,
      "avg_duration_per_task_ms": 700478.4300451322,
      "avg_output_tokens_per_task": 31633.707930367505,
      "last_run_timestamp": "2026-09-12T09:07:23.668Z"
    },
    {
      "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": "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7659437959183671,
      "accuracy_std_dev": 0.0759970281637527,
      "run_count": 12,
      "total_questions": 2352,
      "avg_cost_per_task_usd": 0.004940951955782313,
      "avg_duration_per_task_ms": 720500.3579931973,
      "avg_output_tokens_per_task": 32852.24107142857,
      "last_run_timestamp": "2026-09-19T18:05:31.436Z"
    },
    {
      "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": 0.765512238095238,
      "accuracy_std_dev": 0.012007319862344141,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.011864828763828762,
      "avg_duration_per_task_ms": 112663.68951418951,
      "avg_output_tokens_per_task": 19644.012265512265,
      "last_run_timestamp": "2026-08-29T20:01:12.707Z"
    },
    {
      "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": "WandB",
      "accuracy": 0.7651513333333334,
      "accuracy_std_dev": 0.022539240875021092,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.00031351936026936025,
      "avg_duration_per_task_ms": 33882.18939393939,
      "avg_output_tokens_per_task": 2865.460437710438,
      "last_run_timestamp": "2026-06-20T16:16:12.961Z"
    },
    {
      "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": 0.7642404121475055,
      "accuracy_std_dev": 0.012311945545143383,
      "run_count": 21,
      "total_questions": 4149,
      "avg_cost_per_task_usd": 0.009483342251144855,
      "avg_duration_per_task_ms": 98709.46348517716,
      "avg_output_tokens_per_task": 18943.602795854422,
      "last_run_timestamp": "2026-08-29T18:39:41.257Z"
    },
    {
      "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": 0.7638135731310943,
      "accuracy_std_dev": 0.022048436994096723,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.005248553990610328,
      "avg_duration_per_task_ms": 264357.4124232575,
      "avg_output_tokens_per_task": 16758.418201516793,
      "last_run_timestamp": "2026-07-18T16:28:34.215Z"
    },
    {
      "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": 0.7636680000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 189,
      "avg_cost_per_task_usd": 0.14155542328042328,
      "avg_duration_per_task_ms": 2288029.2486772486,
      "avg_output_tokens_per_task": 37245.06349206349,
      "last_run_timestamp": "2026-08-23T18:19:25.499Z"
    },
    {
      "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": "SiliconFlow",
      "accuracy": 0.7626265,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.005752090909090909,
      "avg_duration_per_task_ms": 193402.83585858587,
      "avg_output_tokens_per_task": 2823.3358585858587,
      "last_run_timestamp": "2026-05-09T09:35:10.004Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.7613916000000002,
      "accuracy_std_dev": 0.18676951595653934,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.005808141077441077,
      "avg_duration_per_task_ms": 42452.30505050505,
      "avg_output_tokens_per_task": 6068.625925925926,
      "last_run_timestamp": "2026-07-01T02:27:21.176Z"
    },
    {
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "SambaNova",
      "accuracy": 0.7608648513980522,
      "accuracy_std_dev": 0.21997558889516092,
      "run_count": 19,
      "total_questions": 3183,
      "avg_cost_per_task_usd": 0.04201378196669809,
      "avg_duration_per_task_ms": 59751.10775997487,
      "avg_output_tokens_per_task": 14315.729814640277,
      "last_run_timestamp": "2026-08-15T14:54:40.470Z"
    },
    {
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "Novita",
      "accuracy": 0.7606466373507059,
      "accuracy_std_dev": 0.023479472398289566,
      "run_count": 14,
      "total_questions": 2763,
      "avg_cost_per_task_usd": 0.010367346362649294,
      "avg_duration_per_task_ms": 144206.35830618892,
      "avg_output_tokens_per_task": 4038.1357220412597,
      "last_run_timestamp": "2026-08-01T17:44:11.144Z"
    },
    {
      "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": 0.7605543076923077,
      "accuracy_std_dev": 0.017867230930843605,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.010017935897435899,
      "avg_duration_per_task_ms": 107941.90792540793,
      "avg_output_tokens_per_task": 3732.981351981352,
      "last_run_timestamp": "2026-08-01T17:45:27.258Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.7605394662379422,
      "accuracy_std_dev": 0.10958240222843778,
      "run_count": 19,
      "total_questions": 3732,
      "avg_cost_per_task_usd": 0.002762414523043944,
      "avg_duration_per_task_ms": 395715.09458735265,
      "avg_output_tokens_per_task": 14065.52706323687,
      "last_run_timestamp": "2026-08-29T19:07:25.263Z"
    },
    {
      "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": 0.7595521475625824,
      "accuracy_std_dev": 0.019571763499602556,
      "run_count": 8,
      "total_questions": 1518,
      "avg_cost_per_task_usd": 0.013668990118577072,
      "avg_duration_per_task_ms": 155525.03227931488,
      "avg_output_tokens_per_task": 18166.727272727272,
      "last_run_timestamp": "2026-08-29T18:11:51.470Z"
    },
    {
      "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": "Amazon Bedrock",
      "accuracy": 0.7587190035005834,
      "accuracy_std_dev": 0.01366534537530361,
      "run_count": 26,
      "total_questions": 5142,
      "avg_cost_per_task_usd": 0.015162635355892653,
      "avg_duration_per_task_ms": 166747.76215480358,
      "avg_output_tokens_per_task": 25091.703033838974,
      "last_run_timestamp": "2026-08-29T16:18:21.978Z"
    },
    {
      "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": 0.7585374615196078,
      "accuracy_std_dev": 0.038599957465404885,
      "run_count": 21,
      "total_questions": 4080,
      "avg_cost_per_task_usd": 0.01587172034313725,
      "avg_duration_per_task_ms": 44875.42181372549,
      "avg_output_tokens_per_task": 16664.9625,
      "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-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.7577766369409167,
      "accuracy_std_dev": 0.02573710862237637,
      "run_count": 20,
      "total_questions": 3622,
      "avg_cost_per_task_usd": 0.00409766344561016,
      "avg_duration_per_task_ms": 651759.315847598,
      "avg_output_tokens_per_task": 10571.302319160684,
      "last_run_timestamp": "2026-08-29T19:19:08.877Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.7569231911314984,
      "accuracy_std_dev": 0.07227409752980317,
      "run_count": 20,
      "total_questions": 3924,
      "avg_cost_per_task_usd": 0.006370367991845058,
      "avg_duration_per_task_ms": 612171.6885830784,
      "avg_output_tokens_per_task": 7236.809887869521,
      "last_run_timestamp": "2026-08-30T01:20:44.266Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Phala",
      "accuracy": 0.7559212403433475,
      "accuracy_std_dev": 0.12618386614116156,
      "run_count": 11,
      "total_questions": 2097,
      "avg_cost_per_task_usd": 0.07370895374344302,
      "avg_duration_per_task_ms": 833401.029089175,
      "avg_output_tokens_per_task": 27619.250357653793,
      "last_run_timestamp": "2026-09-19T17:25:48.622Z"
    },
    {
      "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": 0.755892,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0020983333333333336,
      "avg_duration_per_task_ms": 164183.51515151514,
      "avg_output_tokens_per_task": 7127.272727272727,
      "last_run_timestamp": "2026-08-21T21:34:38.624Z"
    },
    {
      "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": 0.7552889455999998,
      "accuracy_std_dev": 0.020781831603270767,
      "run_count": 19,
      "total_questions": 3750,
      "avg_cost_per_task_usd": 0.010816569600000001,
      "avg_duration_per_task_ms": 208301.1416,
      "avg_output_tokens_per_task": 18668.501333333334,
      "last_run_timestamp": "2026-08-29T18:30:03.855Z"
    },
    {
      "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": "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": 0.7551499775280899,
      "accuracy_std_dev": 0.03338976368963997,
      "run_count": 6,
      "total_questions": 1068,
      "avg_cost_per_task_usd": 0.006923780898876404,
      "avg_duration_per_task_ms": 301563.05337078654,
      "avg_output_tokens_per_task": 21238.912921348314,
      "last_run_timestamp": "2026-07-18T08:04:08.574Z"
    },
    {
      "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": 0.7533753845344554,
      "accuracy_std_dev": 0.11640605414937877,
      "run_count": 11,
      "total_questions": 1901,
      "avg_cost_per_task_usd": 0.04040272172540768,
      "avg_duration_per_task_ms": 597850.3240399789,
      "avg_output_tokens_per_task": 17550.840610205156,
      "last_run_timestamp": "2026-08-22T16:06:14.894Z"
    },
    {
      "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": "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Inceptron",
      "accuracy": 0.7518907958115183,
      "accuracy_std_dev": 0.056756617575006676,
      "run_count": 6,
      "total_questions": 1146,
      "avg_cost_per_task_usd": 0.0764649572425829,
      "avg_duration_per_task_ms": 465218.3132635253,
      "avg_output_tokens_per_task": 17503.08987783595,
      "last_run_timestamp": "2026-06-30T23:33:15.910Z"
    },
    {
      "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": 0.75178689122807,
      "accuracy_std_dev": 0.06483392638966569,
      "run_count": 13,
      "total_questions": 2565,
      "avg_cost_per_task_usd": 0.01753287212475633,
      "avg_duration_per_task_ms": 827358.4358674464,
      "avg_output_tokens_per_task": 10936.669005847953,
      "last_run_timestamp": "2026-08-22T17:28:27.797Z"
    },
    {
      "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": 0.7515384892307692,
      "accuracy_std_dev": 0.10704096328246257,
      "run_count": 8,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.002935932307692308,
      "avg_duration_per_task_ms": 200717.69076923077,
      "avg_output_tokens_per_task": 13572.203076923077,
      "last_run_timestamp": "2026-09-12T16:00:20.960Z"
    },
    {
      "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": 0.7510363518653986,
      "accuracy_std_dev": 0.11063154442009523,
      "run_count": 7,
      "total_questions": 1367,
      "avg_cost_per_task_usd": 0.05419231089978055,
      "avg_duration_per_task_ms": 442904.11777615215,
      "avg_output_tokens_per_task": 22733.470373079737,
      "last_run_timestamp": "2026-08-15T10:07:37.662Z"
    },
    {
      "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": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "DeepInfra",
      "accuracy": 0.7504692996781979,
      "accuracy_std_dev": 0.12884102665725183,
      "run_count": 13,
      "total_questions": 2486,
      "avg_cost_per_task_usd": 0.03235196661303299,
      "avg_duration_per_task_ms": 504543.08045052295,
      "avg_output_tokens_per_task": 14950.808125502816,
      "last_run_timestamp": "2026-08-29T16:22:26.395Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.7496631999999999,
      "accuracy_std_dev": 0.012666643389627747,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.00575224191919192,
      "avg_duration_per_task_ms": 96461.17525252525,
      "avg_output_tokens_per_task": 2226.588888888889,
      "last_run_timestamp": "2026-07-01T05:04:09.747Z"
    },
    {
      "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": "Cerebras",
      "accuracy": 0.7494948,
      "accuracy_std_dev": 0.018515427928082036,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0035561919191919194,
      "avg_duration_per_task_ms": 3605.560606060606,
      "avg_output_tokens_per_task": 2828.088888888889,
      "last_run_timestamp": "2026-05-23T08:06:03.679Z"
    },
    {
      "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": 0.7493782527075812,
      "accuracy_std_dev": 0.028699053688207896,
      "run_count": 21,
      "total_questions": 4155,
      "avg_cost_per_task_usd": 0.007139869554753309,
      "avg_duration_per_task_ms": 177604.2055354994,
      "avg_output_tokens_per_task": 14341.533814681106,
      "last_run_timestamp": "2026-08-29T19:17:18.768Z"
    },
    {
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Venice",
      "accuracy": 0.7485421870794079,
      "accuracy_std_dev": 0.10163071553147703,
      "run_count": 5,
      "total_questions": 743,
      "avg_cost_per_task_usd": 0.014177663526244952,
      "avg_duration_per_task_ms": 490544.0363391655,
      "avg_output_tokens_per_task": 26085.409152086137,
      "last_run_timestamp": "2026-09-15T00:30:59.087Z"
    },
    {
      "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": 0.7479971828113618,
      "accuracy_std_dev": 0.07910913250219596,
      "run_count": 14,
      "total_questions": 2746,
      "avg_cost_per_task_usd": 0.006062376911871813,
      "avg_duration_per_task_ms": 285154.7793153678,
      "avg_output_tokens_per_task": 21130.692279679533,
      "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-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Together",
      "accuracy": 0.7474749999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.004966555555555556,
      "avg_duration_per_task_ms": 39085.368686868685,
      "avg_output_tokens_per_task": 4025.959595959596,
      "last_run_timestamp": "2026-05-09T08:06:12.753Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Phala",
      "accuracy": 0.7473080875303925,
      "accuracy_std_dev": 0.15339754617572723,
      "run_count": 16,
      "total_questions": 2879,
      "avg_cost_per_task_usd": 0.06448924591872178,
      "avg_duration_per_task_ms": 699576.7933310177,
      "avg_output_tokens_per_task": 22910.511983327546,
      "last_run_timestamp": "2026-08-29T18:25:52.753Z"
    },
    {
      "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": 0.7472975263157895,
      "accuracy_std_dev": 0.014827052609720271,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.0034353149920255174,
      "avg_duration_per_task_ms": 77071.67570441254,
      "avg_output_tokens_per_task": 5605.975544922913,
      "last_run_timestamp": "2026-08-29T18:15:48.874Z"
    },
    {
      "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": 0.7467733333333335,
      "accuracy_std_dev": 0.012587939799224065,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.011146377104377103,
      "avg_duration_per_task_ms": 81852.9377104377,
      "avg_output_tokens_per_task": 12261.164983164983,
      "last_run_timestamp": "2026-08-29T18:17:28.541Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-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": "Amazon Bedrock",
      "accuracy": 0.7464074370245138,
      "accuracy_std_dev": 0.026169100495032988,
      "run_count": 18,
      "total_questions": 3549,
      "avg_cost_per_task_usd": 0.10053423443223442,
      "avg_duration_per_task_ms": 535768.9887292195,
      "avg_output_tokens_per_task": 31443.99887292195,
      "last_run_timestamp": "2026-08-29T18:15:24.940Z"
    },
    {
      "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": 0.7463525,
      "accuracy_std_dev": 0.013479678966050276,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.010016893658810323,
      "avg_duration_per_task_ms": 77371.40852974186,
      "avg_output_tokens_per_task": 9879.324354657689,
      "last_run_timestamp": "2026-08-29T18:11:58.511Z"
    },
    {
      "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/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7460692307032589,
      "accuracy_std_dev": 0.11382303442613109,
      "run_count": 12,
      "total_questions": 2332,
      "avg_cost_per_task_usd": 0.03359197898799315,
      "avg_duration_per_task_ms": 1194136.795883362,
      "avg_output_tokens_per_task": 13708.225986277874,
      "last_run_timestamp": "2026-08-29T20:16:12.181Z"
    },
    {
      "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": "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": 0.7450696428571427,
      "accuracy_std_dev": 0.014631101665420764,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.00583946608946609,
      "avg_duration_per_task_ms": 147663.13888888888,
      "avg_output_tokens_per_task": 6788.261904761905,
      "last_run_timestamp": "2026-08-29T18:27:42.956Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.744804408618582,
      "accuracy_std_dev": 0.052815519367177405,
      "run_count": 18,
      "total_questions": 3272,
      "avg_cost_per_task_usd": 0.07897924847188266,
      "avg_duration_per_task_ms": 486054.8694987775,
      "avg_output_tokens_per_task": 22570.938569682152,
      "last_run_timestamp": "2026-08-22T17:30:52.646Z"
    },
    {
      "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": 0.7427059912472648,
      "accuracy_std_dev": 0.04677682194604984,
      "run_count": 14,
      "total_questions": 2742,
      "avg_cost_per_task_usd": 0.008213210430342814,
      "avg_duration_per_task_ms": 681410.7862873814,
      "avg_output_tokens_per_task": 20787.104668125456,
      "last_run_timestamp": "2026-08-22T10:02:15.507Z"
    },
    {
      "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-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Wafer",
      "accuracy": 0.7424240000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.048281424242424244,
      "avg_duration_per_task_ms": 987580.0707070708,
      "avg_output_tokens_per_task": 21326.60606060606,
      "last_run_timestamp": "2026-06-13T08:04:32.525Z"
    },
    {
      "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": 0.7399168897876645,
      "accuracy_std_dev": 0.01871797550235247,
      "run_count": 15,
      "total_questions": 2967,
      "avg_cost_per_task_usd": 0.003766104819683181,
      "avg_duration_per_task_ms": 178216.75362318842,
      "avg_output_tokens_per_task": 5955.366363329963,
      "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": "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "Groq",
      "accuracy": 0.7397306666666669,
      "accuracy_std_dev": 0.023056894603470685,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.009337984175084173,
      "avg_duration_per_task_ms": 13785.80404040404,
      "avg_output_tokens_per_task": 2990.9208754208753,
      "last_run_timestamp": "2026-08-01T17:45:49.025Z"
    },
    {
      "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": 0.7393956351351351,
      "accuracy_std_dev": 0.03984266574715486,
      "run_count": 18,
      "total_questions": 3552,
      "avg_cost_per_task_usd": 0.00809140793918919,
      "avg_duration_per_task_ms": 458526.92088963964,
      "avg_output_tokens_per_task": 21780.53462837838,
      "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": "Novita",
      "accuracy": 0.7388981686340642,
      "accuracy_std_dev": 0.02112011938163336,
      "run_count": 18,
      "total_questions": 3558,
      "avg_cost_per_task_usd": 0.003153482855536819,
      "avg_duration_per_task_ms": 127292.86734120292,
      "avg_output_tokens_per_task": 5353.168071950534,
      "last_run_timestamp": "2026-08-29T18:16:39.843Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Parasail",
      "accuracy": 0.738764951772013,
      "accuracy_std_dev": 0.25012599709265915,
      "run_count": 14,
      "total_questions": 2737,
      "avg_cost_per_task_usd": 0.060706641213006944,
      "avg_duration_per_task_ms": 511177.83814395324,
      "avg_output_tokens_per_task": 17354.046401169162,
      "last_run_timestamp": "2026-08-29T18:36:08.021Z"
    },
    {
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Google",
      "accuracy": 0.7384040983827492,
      "accuracy_std_dev": 0.01718365966906416,
      "run_count": 30,
      "total_questions": 5936,
      "avg_cost_per_task_usd": 0.005306772237196765,
      "avg_duration_per_task_ms": 67835.73685983827,
      "avg_output_tokens_per_task": 5569.043800539083,
      "last_run_timestamp": "2026-08-29T19:15:44.037Z"
    },
    {
      "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": 0.7381102500000001,
      "accuracy_std_dev": 0.012944321677380398,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.015707310606060607,
      "avg_duration_per_task_ms": 67722.01136363637,
      "avg_output_tokens_per_task": 10394.812184343435,
      "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": "Together",
      "accuracy": 0.7380050000000001,
      "accuracy_std_dev": 0.01915680847636162,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0026956635101010105,
      "avg_duration_per_task_ms": 123729.69255050505,
      "avg_output_tokens_per_task": 4372.988005050505,
      "last_run_timestamp": "2026-07-04T17:56:44.437Z"
    },
    {
      "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": "SiliconFlow",
      "accuracy": 0.7373739999999999,
      "accuracy_std_dev": 0.010101000000000027,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.001737270202020202,
      "avg_duration_per_task_ms": 127098.16161616161,
      "avg_output_tokens_per_task": 2854.8282828282827,
      "last_run_timestamp": "2026-05-09T09:51:37.028Z"
    },
    {
      "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": "Nebius",
      "accuracy": 0.7373736666666667,
      "accuracy_std_dev": 0.18223849985542448,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04463949831649832,
      "avg_duration_per_task_ms": 228965.81313131313,
      "avg_output_tokens_per_task": 14731.36026936027,
      "last_run_timestamp": "2026-06-30T21:13:44.266Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Venice",
      "accuracy": 0.7358002289156627,
      "accuracy_std_dev": 0.04992948859920135,
      "run_count": 9,
      "total_questions": 1743,
      "avg_cost_per_task_usd": 0.0068313775100401625,
      "avg_duration_per_task_ms": 1022403.1669535284,
      "avg_output_tokens_per_task": 18090.09982788296,
      "last_run_timestamp": "2026-08-29T20:11:32.700Z"
    },
    {
      "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": 0.7357838333333334,
      "accuracy_std_dev": 0.03274764050745574,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.004765584175084174,
      "avg_duration_per_task_ms": 89248.37345679013,
      "avg_output_tokens_per_task": 5293.844837261504,
      "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": "DekaLLM",
      "accuracy": 0.735337781737194,
      "accuracy_std_dev": 0.017059479919011102,
      "run_count": 14,
      "total_questions": 2694,
      "avg_cost_per_task_usd": 0.0013270267260579063,
      "avg_duration_per_task_ms": 313987.4647364514,
      "avg_output_tokens_per_task": 7917.988864142539,
      "last_run_timestamp": "2026-07-18T16:27:55.986Z"
    },
    {
      "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": 0.7349234689468948,
      "accuracy_std_dev": 0.22406624102384445,
      "run_count": 6,
      "total_questions": 1111,
      "avg_cost_per_task_usd": 0.07603830513051305,
      "avg_duration_per_task_ms": 255861.0288028803,
      "avg_output_tokens_per_task": 17676.750675067506,
      "last_run_timestamp": "2026-07-18T08:32:17.535Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Phala",
      "accuracy": 0.7338664379562044,
      "accuracy_std_dev": 0.09881106660279618,
      "run_count": 15,
      "total_questions": 2877,
      "avg_cost_per_task_usd": 0.12776717935349324,
      "avg_duration_per_task_ms": 902541.7400069517,
      "avg_output_tokens_per_task": 30238.822036843936,
      "last_run_timestamp": "2026-08-29T18:42:16.346Z"
    },
    {
      "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": 0.7330631323430115,
      "accuracy_std_dev": 0.07952342617106217,
      "run_count": 8,
      "total_questions": 1481,
      "avg_cost_per_task_usd": 0.03091582241728561,
      "avg_duration_per_task_ms": 150438.8203916273,
      "avg_output_tokens_per_task": 27968.933153274815,
      "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": "Amazon Bedrock",
      "accuracy": 0.7321035652173912,
      "accuracy_std_dev": 0.012294195363546603,
      "run_count": 23,
      "total_questions": 4554,
      "avg_cost_per_task_usd": 0.20428436868686872,
      "avg_duration_per_task_ms": 241980.82740447958,
      "avg_output_tokens_per_task": 38170.10320597277,
      "last_run_timestamp": "2026-09-19T08:49:06.253Z"
    },
    {
      "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": 0.7315366610778443,
      "accuracy_std_dev": 0.23240685150709303,
      "run_count": 5,
      "total_questions": 835,
      "avg_cost_per_task_usd": 0.05666666227544911,
      "avg_duration_per_task_ms": 351949.8934131737,
      "avg_output_tokens_per_task": 18214.095808383234,
      "last_run_timestamp": "2026-08-15T08:47:51.369Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Io Net",
      "accuracy": 0.731313,
      "accuracy_std_dev": 0.03591182134618071,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.00554679898989899,
      "avg_duration_per_task_ms": 199026.8303030303,
      "avg_output_tokens_per_task": 11014.675757575758,
      "last_run_timestamp": "2026-05-23T09:41:28.552Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.7310988181818181,
      "accuracy_std_dev": 0.012035232012857314,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.006938019742883379,
      "avg_duration_per_task_ms": 43950.472910927456,
      "avg_output_tokens_per_task": 5102.792011019284,
      "last_run_timestamp": "2026-07-01T02:32:36.197Z"
    },
    {
      "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": 0.7296560582226762,
      "accuracy_std_dev": 0.15560230205957754,
      "run_count": 5,
      "total_questions": 979,
      "avg_cost_per_task_usd": 0.01869288764044944,
      "avg_duration_per_task_ms": 96519.51685393258,
      "avg_output_tokens_per_task": 14860.048008171603,
      "last_run_timestamp": "2026-07-18T16:00:08.986Z"
    },
    {
      "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": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.729377,
      "accuracy_std_dev": 0.01822436147029575,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.18529492234848485,
      "avg_duration_per_task_ms": 288734.0814393939,
      "avg_output_tokens_per_task": 36901.592803030304,
      "last_run_timestamp": "2026-09-19T08:54:09.617Z"
    },
    {
      "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": "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": 0.726431125,
      "accuracy_std_dev": 0.010170955675814077,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.19067559027777775,
      "avg_duration_per_task_ms": 240794.80744949495,
      "avg_output_tokens_per_task": 37944.43686868687,
      "last_run_timestamp": "2026-09-19T08:48:18.111Z"
    },
    {
      "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": 0.7263065279007972,
      "accuracy_std_dev": 0.1217642983358514,
      "run_count": 6,
      "total_questions": 1129,
      "avg_cost_per_task_usd": 0.2174638077945084,
      "avg_duration_per_task_ms": 1661844.5757307352,
      "avg_output_tokens_per_task": 59283.423383525245,
      "last_run_timestamp": "2026-08-22T09:29:12.890Z"
    },
    {
      "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": 0.7261392373678025,
      "accuracy_std_dev": 0.007451069289848932,
      "run_count": 13,
      "total_questions": 2553,
      "avg_cost_per_task_usd": 0.0023729573051312187,
      "avg_duration_per_task_ms": 400351.9271445358,
      "avg_output_tokens_per_task": 6460.609479044261,
      "last_run_timestamp": "2026-08-29T18:39:40.767Z"
    },
    {
      "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": 0.7260814529667148,
      "accuracy_std_dev": 0.020878527218346984,
      "run_count": 21,
      "total_questions": 4146,
      "avg_cost_per_task_usd": 0.004243377713458756,
      "avg_duration_per_task_ms": 227026.06222865413,
      "avg_output_tokens_per_task": 16857.46454413893,
      "last_run_timestamp": "2026-08-29T18:36:32.803Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7249342617387262,
      "accuracy_std_dev": 0.20867571941350305,
      "run_count": 12,
      "total_questions": 2151,
      "avg_cost_per_task_usd": 0.004015295676429567,
      "avg_duration_per_task_ms": 907421.850302185,
      "avg_output_tokens_per_task": 26470.74058577406,
      "last_run_timestamp": "2026-09-19T20:39:06.442Z"
    },
    {
      "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": 0.7244675919165579,
      "accuracy_std_dev": 0.009807494652528479,
      "run_count": 8,
      "total_questions": 1534,
      "avg_cost_per_task_usd": 0.18618542242503258,
      "avg_duration_per_task_ms": 349648.9543676662,
      "avg_output_tokens_per_task": 37047.215775749675,
      "last_run_timestamp": "2026-09-19T08:49:55.182Z"
    },
    {
      "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": 0.7230691676829271,
      "accuracy_std_dev": 0.01328152332415223,
      "run_count": 10,
      "total_questions": 1968,
      "avg_cost_per_task_usd": 0.005270103150406503,
      "avg_duration_per_task_ms": 358290.0543699187,
      "avg_output_tokens_per_task": 7038.59349593496,
      "last_run_timestamp": "2026-08-22T16:50:17.574Z"
    },
    {
      "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": 0.7223346,
      "accuracy_std_dev": 0.07253641034202524,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.1090764063973064,
      "avg_duration_per_task_ms": 826522.4053872054,
      "avg_output_tokens_per_task": 27681.187205387207,
      "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-5.2-20260616",
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Phala",
      "accuracy": 0.7198153664772725,
      "accuracy_std_dev": 0.16422235834886253,
      "run_count": 15,
      "total_questions": 2816,
      "avg_cost_per_task_usd": 0.008276323508522728,
      "avg_duration_per_task_ms": 230801.09623579544,
      "avg_output_tokens_per_task": 6590.78515625,
      "last_run_timestamp": "2026-07-18T16:47:35.300Z"
    },
    {
      "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": 0.7197963406113538,
      "accuracy_std_dev": 0.10733758179611058,
      "run_count": 7,
      "total_questions": 1374,
      "avg_cost_per_task_usd": 0.00507840538573508,
      "avg_duration_per_task_ms": 429791.4192139738,
      "avg_output_tokens_per_task": 22080.005094614266,
      "last_run_timestamp": "2026-09-19T09:16:40.511Z"
    },
    {
      "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": 0.7197671666666666,
      "accuracy_std_dev": 0.010411737109574402,
      "run_count": 24,
      "total_questions": 4752,
      "avg_cost_per_task_usd": 0.1969247567340067,
      "avg_duration_per_task_ms": 258757.54566498316,
      "avg_output_tokens_per_task": 36726.52988215488,
      "last_run_timestamp": "2026-09-19T08:56:57.786Z"
    },
    {
      "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": "DeepInfra",
      "accuracy": 0.7196582307692306,
      "accuracy_std_dev": 0.022147329696075465,
      "run_count": 3,
      "total_questions": 585,
      "avg_cost_per_task_usd": 0.03072831965811966,
      "avg_duration_per_task_ms": 160744.22905982906,
      "avg_output_tokens_per_task": 13887.176068376068,
      "last_run_timestamp": "2026-06-30T21:06:56.500Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Io Net",
      "accuracy": 0.7181841912777354,
      "accuracy_std_dev": 0.06555394037846128,
      "run_count": 7,
      "total_questions": 1307,
      "avg_cost_per_task_usd": 0.060895005355776584,
      "avg_duration_per_task_ms": 455243.765876052,
      "avg_output_tokens_per_task": 23693.538638102524,
      "last_run_timestamp": "2026-08-15T09:29:41.209Z"
    },
    {
      "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/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7166106666666667,
      "accuracy_std_dev": 0.0055553002519115975,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.006570582491582491,
      "avg_duration_per_task_ms": 1280458.1801346801,
      "avg_output_tokens_per_task": 54335.32491582492,
      "last_run_timestamp": "2026-09-19T16:57:17.316Z"
    },
    {
      "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": 0.715615295838926,
      "accuracy_std_dev": 0.03314273943226165,
      "run_count": 19,
      "total_questions": 3725,
      "avg_cost_per_task_usd": 0.003583049395973154,
      "avg_duration_per_task_ms": 213584.61583892617,
      "avg_output_tokens_per_task": 9282.251812080536,
      "last_run_timestamp": "2026-08-29T19:25:16.673Z"
    },
    {
      "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": 0.7151083124999998,
      "accuracy_std_dev": 0.01761938659339479,
      "run_count": 20,
      "total_questions": 3936,
      "avg_cost_per_task_usd": 0.005333751270325202,
      "avg_duration_per_task_ms": 47488.129827235774,
      "avg_output_tokens_per_task": 6775.859756097561,
      "last_run_timestamp": "2026-08-29T18:12:18.859Z"
    },
    {
      "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": 0.7146379609952925,
      "accuracy_std_dev": 0.12348955006372564,
      "run_count": 23,
      "total_questions": 4461,
      "avg_cost_per_task_usd": 0.0694516646491818,
      "avg_duration_per_task_ms": 609642.5270118808,
      "avg_output_tokens_per_task": 37654.42793095719,
      "last_run_timestamp": "2026-08-29T19:10:33.460Z"
    },
    {
      "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": 0.7145656108440172,
      "accuracy_std_dev": 0.017584165899374688,
      "run_count": 19,
      "total_questions": 3744,
      "avg_cost_per_task_usd": 0.006881101762820511,
      "avg_duration_per_task_ms": 63918.18936965812,
      "avg_output_tokens_per_task": 6296.184561965812,
      "last_run_timestamp": "2026-08-29T19:10:03.342Z"
    },
    {
      "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": 0.7144979411764706,
      "accuracy_std_dev": 0.017256114730285467,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.007874328282828283,
      "avg_duration_per_task_ms": 136479.90077243018,
      "avg_output_tokens_per_task": 4099.4212715389185,
      "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": "DeepInfra",
      "accuracy": 0.713069165234002,
      "accuracy_std_dev": 0.0550740987322616,
      "run_count": 32,
      "total_questions": 6282,
      "avg_cost_per_task_usd": 0.005810328876154091,
      "avg_duration_per_task_ms": 328744.6372174467,
      "avg_output_tokens_per_task": 17988.781916587075,
      "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-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Ionstream",
      "accuracy": 0.7115813800243607,
      "accuracy_std_dev": 0.0647214547058296,
      "run_count": 17,
      "total_questions": 3284,
      "avg_cost_per_task_usd": 0.004965959805115713,
      "avg_duration_per_task_ms": 87058.84226552983,
      "avg_output_tokens_per_task": 6155.969853836785,
      "last_run_timestamp": "2026-08-08T16:46:11.805Z"
    },
    {
      "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": 0.7114771734605377,
      "accuracy_std_dev": 0.08659653364462867,
      "run_count": 7,
      "total_questions": 1153,
      "avg_cost_per_task_usd": 0.007400922810060712,
      "avg_duration_per_task_ms": 492060.8178664354,
      "avg_output_tokens_per_task": 32850.75888985256,
      "last_run_timestamp": "2026-08-29T12:30:09.617Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Parasail",
      "accuracy": 0.7105995910791368,
      "accuracy_std_dev": 0.04851107627346435,
      "run_count": 18,
      "total_questions": 3475,
      "avg_cost_per_task_usd": 0.0055933033093525176,
      "avg_duration_per_task_ms": 116405.79597122302,
      "avg_output_tokens_per_task": 7306.719136690647,
      "last_run_timestamp": "2026-08-29T18:11:53.163Z"
    },
    {
      "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": 0.7105179047619047,
      "accuracy_std_dev": 0.016159891819253525,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.00377856685906686,
      "avg_duration_per_task_ms": 69215.27994227994,
      "avg_output_tokens_per_task": 3508.088023088023,
      "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-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.7104377647058823,
      "accuracy_std_dev": 0.016291381852087287,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.006191515745692216,
      "avg_duration_per_task_ms": 101266.32917409387,
      "avg_output_tokens_per_task": 4016.169043374926,
      "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": "WandB",
      "accuracy": 0.709930669222344,
      "accuracy_std_dev": 0.17059505917300186,
      "run_count": 5,
      "total_questions": 913,
      "avg_cost_per_task_usd": 0.05822560240963856,
      "avg_duration_per_task_ms": 183232.56188389924,
      "avg_output_tokens_per_task": 16113.98247535597,
      "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": "AkashML",
      "accuracy": 0.70928025,
      "accuracy_std_dev": 0.2309020441067759,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.015585212121212121,
      "avg_duration_per_task_ms": 184205.19981060605,
      "avg_output_tokens_per_task": 14971.618055555555,
      "last_run_timestamp": "2026-08-15T15:24:35.389Z"
    },
    {
      "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/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Novita",
      "accuracy": 0.7083626384778011,
      "accuracy_std_dev": 0.1958727186126973,
      "run_count": 15,
      "total_questions": 2838,
      "avg_cost_per_task_usd": 0.040317193798449606,
      "avg_duration_per_task_ms": 587592.7716701903,
      "avg_output_tokens_per_task": 13532.868217054263,
      "last_run_timestamp": "2026-08-22T17:45:06.645Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.7073907882352942,
      "accuracy_std_dev": 0.022165627566230677,
      "run_count": 17,
      "total_questions": 3315,
      "avg_cost_per_task_usd": 0.011149934238310712,
      "avg_duration_per_task_ms": 62511.60935143288,
      "avg_output_tokens_per_task": 7378.323076923077,
      "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": "AtlasCloud",
      "accuracy": 0.7072108333333335,
      "accuracy_std_dev": 0.016372077555161485,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.006544036195286195,
      "avg_duration_per_task_ms": 37236.955808080806,
      "avg_output_tokens_per_task": 4331.103535353535,
      "last_run_timestamp": "2026-06-30T19:34:58.612Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.707071,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0038913585858585857,
      "avg_duration_per_task_ms": 62517.207070707074,
      "avg_output_tokens_per_task": 3515.439393939394,
      "last_run_timestamp": "2026-06-13T08:06:00.319Z"
    },
    {
      "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": 0.7067875794392524,
      "accuracy_std_dev": 0.019523788595938254,
      "run_count": 18,
      "total_questions": 3531,
      "avg_cost_per_task_usd": 0.007194955536675164,
      "avg_duration_per_task_ms": 73519.27867459643,
      "avg_output_tokens_per_task": 6882.512036250354,
      "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-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": 0.7052954681504223,
      "accuracy_std_dev": 0.015732334162799875,
      "run_count": 20,
      "total_questions": 3909,
      "avg_cost_per_task_usd": 0.006693433870555131,
      "avg_duration_per_task_ms": 93563.2755180353,
      "avg_output_tokens_per_task": 6043.323612177028,
      "last_run_timestamp": "2026-08-29T18:09:55.969Z"
    },
    {
      "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": 0.7052560216881594,
      "accuracy_std_dev": 0.14220511739766134,
      "run_count": 11,
      "total_questions": 1706,
      "avg_cost_per_task_usd": 0.04422351465416178,
      "avg_duration_per_task_ms": 329497.4015240328,
      "avg_output_tokens_per_task": 12749.240914419695,
      "last_run_timestamp": "2026-09-19T08:53:38.031Z"
    },
    {
      "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": 0.7049911764705882,
      "accuracy_std_dev": 0.01828311467941453,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.006265460487225194,
      "avg_duration_per_task_ms": 163546.93166963756,
      "avg_output_tokens_per_task": 4319.140819964349,
      "last_run_timestamp": "2026-08-29T18:14:24.357Z"
    },
    {
      "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": 0.7047773804347826,
      "accuracy_std_dev": 0.10309159308121915,
      "run_count": 13,
      "total_questions": 2484,
      "avg_cost_per_task_usd": 0.10197293719806763,
      "avg_duration_per_task_ms": 616168.0583735909,
      "avg_output_tokens_per_task": 39185.06280193237,
      "last_run_timestamp": "2026-08-29T16:33:36.680Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Phala",
      "accuracy": 0.7040404,
      "accuracy_std_dev": 0.045813599669093895,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.009978467676767678,
      "avg_duration_per_task_ms": 82731.23535353535,
      "avg_output_tokens_per_task": 3750.608080808081,
      "last_run_timestamp": "2026-09-19T09:43:03.159Z"
    },
    {
      "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": 0.7037366126080326,
      "accuracy_std_dev": 0.08290095218562599,
      "run_count": 21,
      "total_questions": 3934,
      "avg_cost_per_task_usd": 0.006630117946110829,
      "avg_duration_per_task_ms": 234021.50737163192,
      "avg_output_tokens_per_task": 12083.909506863243,
      "last_run_timestamp": "2026-08-29T19:16:51.883Z"
    },
    {
      "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": 0.7036724033898304,
      "accuracy_std_dev": 0.03265004974302829,
      "run_count": 3,
      "total_questions": 590,
      "avg_cost_per_task_usd": 0.011227418644067795,
      "avg_duration_per_task_ms": 599401.1288135593,
      "avg_output_tokens_per_task": 55874.53050847458,
      "last_run_timestamp": "2026-09-19T08:50:20.272Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "OpenInference",
      "accuracy": 0.70328275,
      "accuracy_std_dev": 0.02153875190644758,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0019388005050505052,
      "avg_duration_per_task_ms": 636092.8409090909,
      "avg_output_tokens_per_task": 14000.35101010101,
      "last_run_timestamp": "2026-09-05T08:28:07.290Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Venice",
      "accuracy": 0.7030109514661275,
      "accuracy_std_dev": 0.03752877227029357,
      "run_count": 15,
      "total_questions": 2967,
      "avg_cost_per_task_usd": 0.009539730704415232,
      "avg_duration_per_task_ms": 166708.53791708796,
      "avg_output_tokens_per_task": 5363.80316818335,
      "last_run_timestamp": "2026-08-29T19:24:48.378Z"
    },
    {
      "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": 0.6998774823529411,
      "accuracy_std_dev": 0.036475041139420926,
      "run_count": 7,
      "total_questions": 1360,
      "avg_cost_per_task_usd": 0.010993920588235295,
      "avg_duration_per_task_ms": 289576.1580882353,
      "avg_output_tokens_per_task": 46600.06911764706,
      "last_run_timestamp": "2026-09-19T08:58:43.694Z"
    },
    {
      "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": 0.6989537142857144,
      "accuracy_std_dev": 0.013756650596444363,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.0019099855699855697,
      "avg_duration_per_task_ms": 141976.2987012987,
      "avg_output_tokens_per_task": 4153.015151515152,
      "last_run_timestamp": "2026-08-29T16:33:36.891Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Nebius",
      "accuracy": 0.698653,
      "accuracy_std_dev": 0.25478202686610374,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.07541588215488215,
      "avg_duration_per_task_ms": 218316.69360269362,
      "avg_output_tokens_per_task": 18749.255892255893,
      "last_run_timestamp": "2026-07-11T08:07:17.820Z"
    },
    {
      "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": 0.6980896447963801,
      "accuracy_std_dev": 0.028298860510700467,
      "run_count": 7,
      "total_questions": 1326,
      "avg_cost_per_task_usd": 0.014896262443438914,
      "avg_duration_per_task_ms": 719785.2081447964,
      "avg_output_tokens_per_task": 74296.09426847662,
      "last_run_timestamp": "2026-09-19T16:29:00.211Z"
    },
    {
      "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": "Io Net",
      "accuracy": 0.6969694999999999,
      "accuracy_std_dev": 0.02525250000000001,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0008116111111111111,
      "avg_duration_per_task_ms": 29037.578282828283,
      "avg_output_tokens_per_task": 1510.050505050505,
      "last_run_timestamp": "2026-05-09T16:14:50.126Z"
    },
    {
      "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": 0.6961280000000001,
      "accuracy_std_dev": 0.01094299999999998,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.004348992424242424,
      "avg_duration_per_task_ms": 356365.1767676768,
      "avg_output_tokens_per_task": 12056.626262626263,
      "last_run_timestamp": "2026-07-18T08:44:31.047Z"
    },
    {
      "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": 0.6948655000000001,
      "accuracy_std_dev": 0.03859079866820586,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.0038991833333333332,
      "avg_duration_per_task_ms": 161549.07575757575,
      "avg_output_tokens_per_task": 4745.70202020202,
      "last_run_timestamp": "2026-08-01T17:52:16.799Z"
    },
    {
      "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": 0.6925813109756097,
      "accuracy_std_dev": 0.24072562924993626,
      "run_count": 9,
      "total_questions": 1312,
      "avg_cost_per_task_usd": 0.009318611280487804,
      "avg_duration_per_task_ms": 54896.76600609756,
      "avg_output_tokens_per_task": 9351.999237804877,
      "last_run_timestamp": "2026-08-29T18:12:40.868Z"
    },
    {
      "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": 0.6905726,
      "accuracy_std_dev": 0.11069382289468553,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.012376834343434341,
      "avg_duration_per_task_ms": 236253.0919191919,
      "avg_output_tokens_per_task": 10174.568686868686,
      "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-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Z.AI",
      "accuracy": 0.6897547142857142,
      "accuracy_std_dev": 0.016389379799250536,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0065042720057720054,
      "avg_duration_per_task_ms": 263620.07575757575,
      "avg_output_tokens_per_task": 16221.229437229436,
      "last_run_timestamp": "2026-05-23T17:26:25.014Z"
    },
    {
      "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": "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": 0.6876781730769231,
      "accuracy_std_dev": 0.02499090104818154,
      "run_count": 5,
      "total_questions": 936,
      "avg_cost_per_task_usd": 0.010550641025641025,
      "avg_duration_per_task_ms": 352806.86645299144,
      "avg_output_tokens_per_task": 45914.32692307692,
      "last_run_timestamp": "2026-09-19T08:48:39.213Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403:free",
      "model_slug": null,
      "model_name": "Gemma 4 26B A4B IT (free)",
      "provider_name": "Darkbloom",
      "accuracy": 0.684722,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 120,
      "avg_cost_per_task_usd": null,
      "avg_duration_per_task_ms": 382535.98333333334,
      "avg_output_tokens_per_task": 6994.291666666667,
      "last_run_timestamp": "2026-08-04T05:36:23.551Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Ambient",
      "accuracy": 0.6843435,
      "accuracy_std_dev": 0.017676499999999984,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0009698787878787878,
      "avg_duration_per_task_ms": 30045.00505050505,
      "avg_output_tokens_per_task": 1560.3434343434344,
      "last_run_timestamp": "2026-05-23T16:14:51.028Z"
    },
    {
      "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": 0.6819954736842105,
      "accuracy_std_dev": 0.056917293716749924,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.0068329181286549715,
      "avg_duration_per_task_ms": 33886.56246677299,
      "avg_output_tokens_per_task": 5649.480595427964,
      "last_run_timestamp": "2026-08-29T18:10:09.843Z"
    },
    {
      "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": 0.6808069808587688,
      "accuracy_std_dev": 0.21007685166978776,
      "run_count": 10,
      "total_questions": 1933,
      "avg_cost_per_task_usd": 0.015859241076047595,
      "avg_duration_per_task_ms": 373979.4381789964,
      "avg_output_tokens_per_task": 19354.842214174856,
      "last_run_timestamp": "2026-09-19T08:52:30.221Z"
    },
    {
      "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": 0.6804601535703615,
      "accuracy_std_dev": 0.12307264388535231,
      "run_count": 19,
      "total_questions": 3347,
      "avg_cost_per_task_usd": 0.025374532118314908,
      "avg_duration_per_task_ms": 194166.67403645054,
      "avg_output_tokens_per_task": 9087.962354347177,
      "last_run_timestamp": "2026-08-29T19:20:11.540Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Parasail",
      "accuracy": 0.6803968704149127,
      "accuracy_std_dev": 0.16973745683387653,
      "run_count": 18,
      "total_questions": 3326,
      "avg_cost_per_task_usd": 0.0034803749248346364,
      "avg_duration_per_task_ms": 122097.64552014432,
      "avg_output_tokens_per_task": 4482.349368610944,
      "last_run_timestamp": "2026-08-29T19:47:54.565Z"
    },
    {
      "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": 0.6790924761904762,
      "accuracy_std_dev": 0.13080806951758023,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.04236039634439633,
      "avg_duration_per_task_ms": 53557.522607022605,
      "avg_output_tokens_per_task": 15382.483164983165,
      "last_run_timestamp": "2026-08-15T14:46:10.940Z"
    },
    {
      "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": "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": 0.6767873430232558,
      "accuracy_std_dev": 0.041497327246575476,
      "run_count": 16,
      "total_questions": 3096,
      "avg_cost_per_task_usd": 0.014108918604651159,
      "avg_duration_per_task_ms": 334124.46253229975,
      "avg_output_tokens_per_task": 16872.68766149871,
      "last_run_timestamp": "2026-08-01T18:07:49.346Z"
    },
    {
      "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": 0.676548715835141,
      "accuracy_std_dev": 0.020437457580933318,
      "run_count": 14,
      "total_questions": 2766,
      "avg_cost_per_task_usd": 0.005609137382501808,
      "avg_duration_per_task_ms": 246410.7856109906,
      "avg_output_tokens_per_task": 35921.823571945046,
      "last_run_timestamp": "2026-07-18T17:04:25.343Z"
    },
    {
      "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": 0.6752246559691913,
      "accuracy_std_dev": 0.03602352977366553,
      "run_count": 12,
      "total_questions": 2337,
      "avg_cost_per_task_usd": 0.013390630723149338,
      "avg_duration_per_task_ms": 424291.90629011556,
      "avg_output_tokens_per_task": 15722.762088147198,
      "last_run_timestamp": "2026-08-01T17:59:38.872Z"
    },
    {
      "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "Novita",
      "accuracy": 0.6738817142857142,
      "accuracy_std_dev": 0.012906576382541092,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0017926522366522367,
      "avg_duration_per_task_ms": 45642.906204906205,
      "avg_output_tokens_per_task": 5814.865800865801,
      "last_run_timestamp": "2026-05-30T08:06:07.683Z"
    },
    {
      "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": 0.6731349473684209,
      "accuracy_std_dev": 0.085633655090479,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.009196745614035088,
      "avg_duration_per_task_ms": 10451.350345560872,
      "avg_output_tokens_per_task": 11969.038277511961,
      "last_run_timestamp": "2026-08-29T18:50:56.950Z"
    },
    {
      "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": 0.6729346307692308,
      "accuracy_std_dev": 0.008729468916008341,
      "run_count": 3,
      "total_questions": 585,
      "avg_cost_per_task_usd": 0.002463391452991453,
      "avg_duration_per_task_ms": 1210800.4,
      "avg_output_tokens_per_task": 14547.558974358974,
      "last_run_timestamp": "2026-07-18T09:58:41.490Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DeepInfra",
      "accuracy": 0.6716106314888761,
      "accuracy_std_dev": 0.014749164820987887,
      "run_count": 18,
      "total_questions": 3506,
      "avg_cost_per_task_usd": 0.005318032515687393,
      "avg_duration_per_task_ms": 413726.3981745579,
      "avg_output_tokens_per_task": 37904.13605248146,
      "last_run_timestamp": "2026-08-29T21:19:16.003Z"
    },
    {
      "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": 0.671246097613883,
      "accuracy_std_dev": 0.027379954626458713,
      "run_count": 14,
      "total_questions": 2766,
      "avg_cost_per_task_usd": 0.0018295003615328993,
      "avg_duration_per_task_ms": 160145.0853217643,
      "avg_output_tokens_per_task": 3977.9674620390456,
      "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-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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Cloudflare",
      "accuracy": 0.6707074,
      "accuracy_std_dev": 0.056996856052242015,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.028509575757575758,
      "avg_duration_per_task_ms": 178293.2676767677,
      "avg_output_tokens_per_task": 9460.838383838383,
      "last_run_timestamp": "2026-05-23T08:06:11.911Z"
    },
    {
      "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": 0.6703578219761501,
      "accuracy_std_dev": 0.0697645260135041,
      "run_count": 18,
      "total_questions": 3522,
      "avg_cost_per_task_usd": 0.09066784469051674,
      "avg_duration_per_task_ms": 715646.2507098239,
      "avg_output_tokens_per_task": 22289.18341851221,
      "last_run_timestamp": "2026-08-22T17:24:19.901Z"
    },
    {
      "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": 0.6696696261261261,
      "accuracy_std_dev": 0.08367726914995827,
      "run_count": 21,
      "total_questions": 3774,
      "avg_cost_per_task_usd": 0.006798153153153154,
      "avg_duration_per_task_ms": 35321.55060943296,
      "avg_output_tokens_per_task": 4497.988076311606,
      "last_run_timestamp": "2026-08-22T17:34:08.243Z"
    },
    {
      "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": 0.6658203026315791,
      "accuracy_std_dev": 0.02735758829038327,
      "run_count": 22,
      "total_questions": 4332,
      "avg_cost_per_task_usd": 0.0065892525392428425,
      "avg_duration_per_task_ms": 404554.2984764543,
      "avg_output_tokens_per_task": 35346.29085872576,
      "last_run_timestamp": "2026-08-29T21:14:10.017Z"
    },
    {
      "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": 0.6643096000000001,
      "accuracy_std_dev": 0.049762880827379743,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.00878879595959596,
      "avg_duration_per_task_ms": 266170.2818181818,
      "avg_output_tokens_per_task": 43649.65555555555,
      "last_run_timestamp": "2026-09-19T09:51:50.936Z"
    },
    {
      "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": 0.6640851398698884,
      "accuracy_std_dev": 0.02270387404538052,
      "run_count": 33,
      "total_questions": 6456,
      "avg_cost_per_task_usd": 0.006331274473358116,
      "avg_duration_per_task_ms": 262668.35517348204,
      "avg_output_tokens_per_task": 41968.7614622057,
      "last_run_timestamp": "2026-08-29T22:02:09.071Z"
    },
    {
      "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": 0.6634949143621766,
      "accuracy_std_dev": 0.09211341535949208,
      "run_count": 17,
      "total_questions": 3363,
      "avg_cost_per_task_usd": 0.0035216934284864705,
      "avg_duration_per_task_ms": 207363.49955396968,
      "avg_output_tokens_per_task": 3883.495391019923,
      "last_run_timestamp": "2026-08-29T19:11:19.710Z"
    },
    {
      "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": "Sail Research",
      "accuracy": 0.662458,
      "accuracy_std_dev": 0.15968229836929324,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.012940323232323234,
      "avg_duration_per_task_ms": 748366.1805555555,
      "avg_output_tokens_per_task": 27034.502525252527,
      "last_run_timestamp": "2026-09-19T09:48:41.123Z"
    },
    {
      "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": "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": 0.6615472218430034,
      "accuracy_std_dev": 0.011144726307502387,
      "run_count": 9,
      "total_questions": 1758,
      "avg_cost_per_task_usd": 0.002312742320819112,
      "avg_duration_per_task_ms": 194982.48521046643,
      "avg_output_tokens_per_task": 3098.169510807736,
      "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-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Z.AI",
      "accuracy": 0.6588440238663485,
      "accuracy_std_dev": 0.045738584476405955,
      "run_count": 13,
      "total_questions": 2514,
      "avg_cost_per_task_usd": 0.0178249661893397,
      "avg_duration_per_task_ms": 380429.29435163084,
      "avg_output_tokens_per_task": 16824.745425616547,
      "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": "NextBit",
      "accuracy": 0.6578283333333335,
      "accuracy_std_dev": 0.02342969883265444,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.0015064617003367007,
      "avg_duration_per_task_ms": 42020.857323232325,
      "avg_output_tokens_per_task": 3276.22095959596,
      "last_run_timestamp": "2026-06-20T16:19:52.738Z"
    },
    {
      "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": 0.6575277142857143,
      "accuracy_std_dev": 0.021158523643569998,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.005534338383838384,
      "avg_duration_per_task_ms": 453292.816017316,
      "avg_output_tokens_per_task": 34128.61327561327,
      "last_run_timestamp": "2026-08-29T22:11:02.681Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.6573137777777778,
      "accuracy_std_dev": 0.15283980178661852,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0030762968574635243,
      "avg_duration_per_task_ms": 54932.89898989899,
      "avg_output_tokens_per_task": 1996.0347923681256,
      "last_run_timestamp": "2026-07-01T02:34:47.176Z"
    },
    {
      "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": 0.6566732189349114,
      "accuracy_std_dev": 0.013404939721861693,
      "run_count": 13,
      "total_questions": 2535,
      "avg_cost_per_task_usd": 0.002109493885601578,
      "avg_duration_per_task_ms": 61133.307297830375,
      "avg_output_tokens_per_task": 2302.72899408284,
      "last_run_timestamp": "2026-08-29T19:49:56.943Z"
    },
    {
      "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": 0.6566194539007094,
      "accuracy_std_dev": 0.021003345172555248,
      "run_count": 16,
      "total_questions": 3102,
      "avg_cost_per_task_usd": 0.007993988072211475,
      "avg_duration_per_task_ms": 399361.4635718891,
      "avg_output_tokens_per_task": 27495.45293359123,
      "last_run_timestamp": "2026-08-22T17:46:40.285Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Phala",
      "accuracy": 0.656566,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0025412272727272727,
      "avg_duration_per_task_ms": 58356.79797979798,
      "avg_output_tokens_per_task": 3959.469696969697,
      "last_run_timestamp": "2026-06-20T08:04:16.013Z"
    },
    {
      "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": 0.6560645175572519,
      "accuracy_std_dev": 0.017260755028097704,
      "run_count": 20,
      "total_questions": 3930,
      "avg_cost_per_task_usd": 0.0233054806615776,
      "avg_duration_per_task_ms": 505245.40814249363,
      "avg_output_tokens_per_task": 12144.778880407124,
      "last_run_timestamp": "2026-08-29T18:31:43.917Z"
    },
    {
      "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": 0.6551628333333336,
      "accuracy_std_dev": 0.01778329004758867,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.001573050785634119,
      "avg_duration_per_task_ms": 60087.49382716049,
      "avg_output_tokens_per_task": 1458.057519640853,
      "last_run_timestamp": "2026-08-29T20:54:16.286Z"
    },
    {
      "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": 0.6544860441176471,
      "accuracy_std_dev": 0.16388817243977186,
      "run_count": 23,
      "total_questions": 4488,
      "avg_cost_per_task_usd": 0.013230713458110515,
      "avg_duration_per_task_ms": 237553.63123885918,
      "avg_output_tokens_per_task": 8754.879679144386,
      "last_run_timestamp": "2026-08-29T19:03:14.914Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Ambient",
      "accuracy": 0.6523570000000001,
      "accuracy_std_dev": 0.08558150047177253,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.025248665824915827,
      "avg_duration_per_task_ms": 247158.74831649833,
      "avg_output_tokens_per_task": 6019.595959595959,
      "last_run_timestamp": "2026-06-13T09:16:40.692Z"
    },
    {
      "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": 0.6521176887892377,
      "accuracy_std_dev": 0.0771054939636864,
      "run_count": 17,
      "total_questions": 3345,
      "avg_cost_per_task_usd": 0.006340421225710015,
      "avg_duration_per_task_ms": 602311.3940209268,
      "avg_output_tokens_per_task": 18765.147683109117,
      "last_run_timestamp": "2026-08-29T19:15:54.652Z"
    },
    {
      "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": "Io Net",
      "accuracy": 0.6519742727272728,
      "accuracy_std_dev": 0.04715498277361752,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.04022706014692378,
      "avg_duration_per_task_ms": 252622.85904499542,
      "avg_output_tokens_per_task": 12756.500459136823,
      "last_run_timestamp": "2026-06-20T09:54:07.719Z"
    },
    {
      "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": 0.6517467489082969,
      "accuracy_std_dev": 0.01932244128386387,
      "run_count": 14,
      "total_questions": 2748,
      "avg_cost_per_task_usd": 0.004762270378457059,
      "avg_duration_per_task_ms": 99071.56186317322,
      "avg_output_tokens_per_task": 8974.740902474527,
      "last_run_timestamp": "2026-09-19T09:11:59.179Z"
    },
    {
      "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": "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": 0.6498317272727273,
      "accuracy_std_dev": 0.020163709926908996,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.0017987033976124885,
      "avg_duration_per_task_ms": 59537.23829201102,
      "avg_output_tokens_per_task": 2292.5165289256197,
      "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-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": 0.6494412710059173,
      "accuracy_std_dev": 0.019678828707867992,
      "run_count": 13,
      "total_questions": 2535,
      "avg_cost_per_task_usd": 0.005908273767258381,
      "avg_duration_per_task_ms": 187327.3996055227,
      "avg_output_tokens_per_task": 9793.276134122289,
      "last_run_timestamp": "2026-09-19T09:14:07.306Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "AtlasCloud",
      "accuracy": 0.648008,
      "accuracy_std_dev": 0.021443451704424816,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.0008452306397306398,
      "avg_duration_per_task_ms": 27333.61153198653,
      "avg_output_tokens_per_task": 890.7693602693603,
      "last_run_timestamp": "2026-07-01T02:41:54.492Z"
    },
    {
      "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": "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": 0.6475160921501707,
      "accuracy_std_dev": 0.010423132030888464,
      "run_count": 9,
      "total_questions": 1758,
      "avg_cost_per_task_usd": 0.002509189419795222,
      "avg_duration_per_task_ms": 70158.00284414107,
      "avg_output_tokens_per_task": 2711.532423208191,
      "last_run_timestamp": "2026-08-29T17:03:33.297Z"
    },
    {
      "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": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Venice",
      "accuracy": 0.6464646666666666,
      "accuracy_std_dev": 0.02295965891635927,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0018580218855218853,
      "avg_duration_per_task_ms": 38024.51683501683,
      "avg_output_tokens_per_task": 2618.2138047138046,
      "last_run_timestamp": "2026-05-16T08:06:13.049Z"
    },
    {
      "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": 0.6459677395788569,
      "accuracy_std_dev": 0.1480303961068657,
      "run_count": 12,
      "total_questions": 2327,
      "avg_cost_per_task_usd": 0.028217204555221313,
      "avg_duration_per_task_ms": 1621170.644177052,
      "avg_output_tokens_per_task": 20034.05242801891,
      "last_run_timestamp": "2026-08-22T18:19:05.094Z"
    },
    {
      "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": 0.6447810909090911,
      "accuracy_std_dev": 0.013751933189420343,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.00262662580348944,
      "avg_duration_per_task_ms": 31615.61432506887,
      "avg_output_tokens_per_task": 2488.2291092745636,
      "last_run_timestamp": "2026-08-29T19:51:51.198Z"
    },
    {
      "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": 0.6440172000000001,
      "accuracy_std_dev": 0.016796480873153823,
      "run_count": 12,
      "total_questions": 2340,
      "avg_cost_per_task_usd": 0.0037662337606837607,
      "avg_duration_per_task_ms": 61315.614102564105,
      "avg_output_tokens_per_task": 7079.364957264957,
      "last_run_timestamp": "2026-08-29T18:43:21.525Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.6439395,
      "accuracy_std_dev": 0.007575499999999957,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.00040658080808080813,
      "avg_duration_per_task_ms": 74334.95202020202,
      "avg_output_tokens_per_task": 2837.530303030303,
      "last_run_timestamp": "2026-06-13T16:33:02.717Z"
    },
    {
      "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": 0.6437363975720788,
      "accuracy_std_dev": 0.021818161687227465,
      "run_count": 10,
      "total_questions": 1977,
      "avg_cost_per_task_usd": 0.001327898330804249,
      "avg_duration_per_task_ms": 54999.67981790592,
      "avg_output_tokens_per_task": 4282.004046535154,
      "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-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Novita",
      "accuracy": 0.6420172700135687,
      "accuracy_std_dev": 0.04047369989098828,
      "run_count": 23,
      "total_questions": 4422,
      "avg_cost_per_task_usd": 0.012410424694708279,
      "avg_duration_per_task_ms": 541845.6732247851,
      "avg_output_tokens_per_task": 31847.307553143375,
      "last_run_timestamp": "2026-08-30T00:32:50.853Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.6413600199386503,
      "accuracy_std_dev": 0.016520034383968443,
      "run_count": 20,
      "total_questions": 3912,
      "avg_cost_per_task_usd": 0.0015176196319018407,
      "avg_duration_per_task_ms": 81604.62193251534,
      "avg_output_tokens_per_task": 5287.028885480572,
      "last_run_timestamp": "2026-08-29T19:32:59.441Z"
    },
    {
      "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": 0.6411883358705995,
      "accuracy_std_dev": 0.07008215552556424,
      "run_count": 16,
      "total_questions": 3153,
      "avg_cost_per_task_usd": 0.012979633047890899,
      "avg_duration_per_task_ms": 350488.820805582,
      "avg_output_tokens_per_task": 15445.433872502379,
      "last_run_timestamp": "2026-08-01T18:25:56.196Z"
    },
    {
      "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "WandB",
      "accuracy": 0.6408341596638656,
      "accuracy_std_dev": 0.030146989624529465,
      "run_count": 11,
      "total_questions": 2142,
      "avg_cost_per_task_usd": 0.0023272763772175537,
      "avg_duration_per_task_ms": 40174.9827264239,
      "avg_output_tokens_per_task": 1243.1489262371615,
      "last_run_timestamp": "2026-07-18T08:26:39.830Z"
    },
    {
      "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": 0.639886023931624,
      "accuracy_std_dev": 0.03350303592049533,
      "run_count": 10,
      "total_questions": 1755,
      "avg_cost_per_task_usd": 0.016610918518518522,
      "avg_duration_per_task_ms": 866168.4626780627,
      "avg_output_tokens_per_task": 69483.85982905983,
      "last_run_timestamp": "2026-09-19T16:49:53.082Z"
    },
    {
      "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": "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": "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": 0.6374862222222222,
      "accuracy_std_dev": 0.015120312553925432,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.001489671717171717,
      "avg_duration_per_task_ms": 106578.17789001123,
      "avg_output_tokens_per_task": 7528.87822671156,
      "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": "Nebius",
      "accuracy": 0.6374368988195617,
      "accuracy_std_dev": 0.013998977609052932,
      "run_count": 18,
      "total_questions": 3558,
      "avg_cost_per_task_usd": 0.0016592459246767845,
      "avg_duration_per_task_ms": 92130.81225407532,
      "avg_output_tokens_per_task": 5361.156548622822,
      "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": "SiliconFlow",
      "accuracy": 0.6374213227665708,
      "accuracy_std_dev": 0.019782319798481038,
      "run_count": 16,
      "total_questions": 3123,
      "avg_cost_per_task_usd": 0.0016995072046109506,
      "avg_duration_per_task_ms": 146595.44732628882,
      "avg_output_tokens_per_task": 5504.1364073006725,
      "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": "SiliconFlow",
      "accuracy": 0.6370851428571429,
      "accuracy_std_dev": 0.022815823531228588,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.0012857987012987013,
      "avg_duration_per_task_ms": 143754.60606060605,
      "avg_output_tokens_per_task": 7026.642857142857,
      "last_run_timestamp": "2026-08-29T17:04:59.562Z"
    },
    {
      "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": 0.6342237881448958,
      "accuracy_std_dev": 0.029265826681420387,
      "run_count": 14,
      "total_questions": 2733,
      "avg_cost_per_task_usd": 0.006398584705451885,
      "avg_duration_per_task_ms": 104759.4526161727,
      "avg_output_tokens_per_task": 8931.070252469814,
      "last_run_timestamp": "2026-09-19T08:52:51.012Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.6341992857142857,
      "accuracy_std_dev": 0.012328930526139207,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0010588167388167389,
      "avg_duration_per_task_ms": 49239.52308802309,
      "avg_output_tokens_per_task": 3418.099567099567,
      "last_run_timestamp": "2026-06-30T19:57:58.812Z"
    },
    {
      "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": "Crusoe",
      "accuracy": 0.6332902229423202,
      "accuracy_std_dev": 0.011149999831137883,
      "run_count": 8,
      "total_questions": 1543,
      "avg_cost_per_task_usd": 0.016406834737524304,
      "avg_duration_per_task_ms": 444530.66429034347,
      "avg_output_tokens_per_task": 82062.96305897602,
      "last_run_timestamp": "2026-09-19T16:24:34.829Z"
    },
    {
      "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": 0.6327843526405451,
      "accuracy_std_dev": 0.01928203380236522,
      "run_count": 9,
      "total_questions": 1761,
      "avg_cost_per_task_usd": 0.010741952299829643,
      "avg_duration_per_task_ms": 472791.5741056218,
      "avg_output_tokens_per_task": 11222.989210675752,
      "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-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Darkbloom",
      "accuracy": 0.6315639897260275,
      "accuracy_std_dev": 0.025646576254064718,
      "run_count": 3,
      "total_questions": 584,
      "avg_cost_per_task_usd": 0.006708763698630138,
      "avg_duration_per_task_ms": 935572.4657534247,
      "avg_output_tokens_per_task": 37027.727739726026,
      "last_run_timestamp": "2026-09-19T08:08:47.836Z"
    },
    {
      "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": 0.6313515294117649,
      "accuracy_std_dev": 0.02320569910065199,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.0037360388994307405,
      "avg_duration_per_task_ms": 386733.3523086654,
      "avg_output_tokens_per_task": 12360.750158127767,
      "last_run_timestamp": "2026-08-29T19:06:01.882Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6296296666666669,
      "accuracy_std_dev": 0.012371329390525862,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.0035772985409652075,
      "avg_duration_per_task_ms": 14957.81088664422,
      "avg_output_tokens_per_task": 873.4545454545455,
      "last_run_timestamp": "2026-06-30T19:49:03.931Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.6272402163187856,
      "accuracy_std_dev": 0.023564479969529893,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.007974122707147374,
      "avg_duration_per_task_ms": 212675.86464263123,
      "avg_output_tokens_per_task": 6702.41302972802,
      "last_run_timestamp": "2026-07-01T03:52:27.932Z"
    },
    {
      "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": 0.6240829792835564,
      "accuracy_std_dev": 0.024714634180733906,
      "run_count": 12,
      "total_questions": 2317,
      "avg_cost_per_task_usd": 0.007086491152352179,
      "avg_duration_per_task_ms": 448985.8131204143,
      "avg_output_tokens_per_task": 11630.75960293483,
      "last_run_timestamp": "2026-09-19T09:21:13.377Z"
    },
    {
      "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": 0.6235919324055667,
      "accuracy_std_dev": 0.017558878241772012,
      "run_count": 8,
      "total_questions": 1509,
      "avg_cost_per_task_usd": 0.005936661365142478,
      "avg_duration_per_task_ms": 567045.3445990722,
      "avg_output_tokens_per_task": 10721.506295559973,
      "last_run_timestamp": "2026-07-18T08:03:56.103Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Alibaba",
      "accuracy": 0.6233549090909093,
      "accuracy_std_dev": 0.012558972760935413,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.007362255739210285,
      "avg_duration_per_task_ms": 220161.05739210284,
      "avg_output_tokens_per_task": 17663.295224977042,
      "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-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.620681432286024,
      "accuracy_std_dev": 0.013051064268315992,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.004365833875045143,
      "avg_duration_per_task_ms": 90965.72264355363,
      "avg_output_tokens_per_task": 8292.383893102204,
      "last_run_timestamp": "2026-08-01T17:51:29.206Z"
    },
    {
      "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": 0.6202935541022593,
      "accuracy_std_dev": 0.014314030660974362,
      "run_count": 13,
      "total_questions": 2523,
      "avg_cost_per_task_usd": 0.004428355925485533,
      "avg_duration_per_task_ms": 174619.32778438367,
      "avg_output_tokens_per_task": 8993.805390408244,
      "last_run_timestamp": "2026-08-01T18:16:00.374Z"
    },
    {
      "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": 0.6193183750000001,
      "accuracy_std_dev": 0.02645072511736446,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.008760742424242424,
      "avg_duration_per_task_ms": 607268.2133838384,
      "avg_output_tokens_per_task": 15504.423611111111,
      "last_run_timestamp": "2026-07-18T17:46:02.370Z"
    },
    {
      "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": 0.6191422040816327,
      "accuracy_std_dev": 0.01950844855599051,
      "run_count": 18,
      "total_questions": 3528,
      "avg_cost_per_task_usd": 0.002386935657596371,
      "avg_duration_per_task_ms": 23158.938775510203,
      "avg_output_tokens_per_task": 1422.096655328798,
      "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": "Together",
      "accuracy": 0.618687,
      "accuracy_std_dev": 0.01041181185000961,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.001993618686868687,
      "avg_duration_per_task_ms": 11609.738636363636,
      "avg_output_tokens_per_task": 726.0012626262626,
      "last_run_timestamp": "2026-05-23T08:09:18.247Z"
    },
    {
      "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": 0.6176683697888469,
      "accuracy_std_dev": 0.033957228972488325,
      "run_count": 10,
      "total_questions": 1847,
      "avg_cost_per_task_usd": 0.013724128857606929,
      "avg_duration_per_task_ms": 551868.6426637791,
      "avg_output_tokens_per_task": 67292.0590146183,
      "last_run_timestamp": "2026-09-19T16:33:17.940Z"
    },
    {
      "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": 0.6139522500000001,
      "accuracy_std_dev": 0.01942167556977256,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.007408927083333333,
      "avg_duration_per_task_ms": 22613.55808080808,
      "avg_output_tokens_per_task": 1412.095643939394,
      "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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6138280549120549,
      "accuracy_std_dev": 0.022194557320808772,
      "run_count": 12,
      "total_questions": 2331,
      "avg_cost_per_task_usd": 0.0015813187473187472,
      "avg_duration_per_task_ms": 35072.91076791077,
      "avg_output_tokens_per_task": 1478.07722007722,
      "last_run_timestamp": "2026-08-29T18:37:17.171Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DekaLLM",
      "accuracy": 0.6129154305135952,
      "accuracy_std_dev": 0.04617983854978681,
      "run_count": 8,
      "total_questions": 1324,
      "avg_cost_per_task_usd": 0.002775106495468278,
      "avg_duration_per_task_ms": 1032830.075528701,
      "avg_output_tokens_per_task": 25108.58836858006,
      "last_run_timestamp": "2026-07-18T16:00:09.483Z"
    },
    {
      "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": 0.6124774159836067,
      "accuracy_std_dev": 0.020744809856282916,
      "run_count": 15,
      "total_questions": 2928,
      "avg_cost_per_task_usd": 0.0023938254781420767,
      "avg_duration_per_task_ms": 43103.5293715847,
      "avg_output_tokens_per_task": 1524.1318306010928,
      "last_run_timestamp": "2026-08-29T19:00:24.068Z"
    },
    {
      "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": 0.6122354822934233,
      "accuracy_std_dev": 0.01895311219280912,
      "run_count": 18,
      "total_questions": 3558,
      "avg_cost_per_task_usd": 0.00473105171444632,
      "avg_duration_per_task_ms": 207884.38617200675,
      "avg_output_tokens_per_task": 11432.557335581787,
      "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": "auto-routing",
      "accuracy": 0.6118327857142859,
      "accuracy_std_dev": 0.022771416930562546,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.004391077561327561,
      "avg_duration_per_task_ms": 183806.7113997114,
      "avg_output_tokens_per_task": 8779.912337662337,
      "last_run_timestamp": "2026-08-01T18:08:49.019Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6099262895816242,
      "accuracy_std_dev": 0.02527746961067045,
      "run_count": 7,
      "total_questions": 1219,
      "avg_cost_per_task_usd": 0.02817470713699754,
      "avg_duration_per_task_ms": 563426.3125512715,
      "avg_output_tokens_per_task": 29571.532403609515,
      "last_run_timestamp": "2026-09-19T16:13:01.086Z"
    },
    {
      "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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "auto-routing",
      "accuracy": 0.6088711387096775,
      "accuracy_std_dev": 0.030645019138335016,
      "run_count": 19,
      "total_questions": 3720,
      "avg_cost_per_task_usd": 0.002026076612903226,
      "avg_duration_per_task_ms": 34227.742204301074,
      "avg_output_tokens_per_task": 1358.1854838709678,
      "last_run_timestamp": "2026-08-29T18:47:46.000Z"
    },
    {
      "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": 0.6085067890625,
      "accuracy_std_dev": 0.10062656908435016,
      "run_count": 4,
      "total_questions": 768,
      "avg_cost_per_task_usd": 0.010816417968749999,
      "avg_duration_per_task_ms": 300613.0859375,
      "avg_output_tokens_per_task": 26862.471354166668,
      "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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "CoreWeave",
      "accuracy": 0.607744,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.003953611111111112,
      "avg_duration_per_task_ms": 57224.242424242424,
      "avg_output_tokens_per_task": 2094.1161616161617,
      "last_run_timestamp": "2026-08-01T17:44:33.510Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Google",
      "accuracy": 0.606531074232691,
      "accuracy_std_dev": 0.0392900303150812,
      "run_count": 15,
      "total_questions": 2802,
      "avg_cost_per_task_usd": 0.002877521056388294,
      "avg_duration_per_task_ms": 32091.98608137045,
      "avg_output_tokens_per_task": 1530.401142041399,
      "last_run_timestamp": "2026-08-29T18:43:09.218Z"
    },
    {
      "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": 0.6062243505658437,
      "accuracy_std_dev": 0.14728578802372175,
      "run_count": 21,
      "total_questions": 3888,
      "avg_cost_per_task_usd": 0.008590261574074075,
      "avg_duration_per_task_ms": 53306.474022633745,
      "avg_output_tokens_per_task": 28544.59079218107,
      "last_run_timestamp": "2026-08-29T19:48:18.157Z"
    },
    {
      "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": 0.6057348861480076,
      "accuracy_std_dev": 0.07360041928198763,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.004612943074003796,
      "avg_duration_per_task_ms": 545569.0094876661,
      "avg_output_tokens_per_task": 18056.067678684376,
      "last_run_timestamp": "2026-08-22T22:18:10.299Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-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": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Groq",
      "accuracy": 0.6009468851294903,
      "accuracy_std_dev": 0.08092483969818855,
      "run_count": 13,
      "total_questions": 2394,
      "avg_cost_per_task_usd": 0.006364984126984126,
      "avg_duration_per_task_ms": 31444.639097744363,
      "avg_output_tokens_per_task": 10538.471177944863,
      "last_run_timestamp": "2026-08-29T16:23:45.123Z"
    },
    {
      "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": 0.5992354149569821,
      "accuracy_std_dev": 0.019006860146473396,
      "run_count": 23,
      "total_questions": 4533,
      "avg_cost_per_task_usd": 0.04292028435914405,
      "avg_duration_per_task_ms": 513498.9199205824,
      "avg_output_tokens_per_task": 16315.828369733068,
      "last_run_timestamp": "2026-08-30T00:22:07.120Z"
    },
    {
      "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": "Claude Platform on AWS",
      "accuracy": 0.5931538333333333,
      "accuracy_std_dev": 0.09808536433028917,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.5963892592592592,
      "avg_duration_per_task_ms": 620612.2373737374,
      "avg_output_tokens_per_task": 39751.62205387205,
      "last_run_timestamp": "2026-09-19T08:49:57.309Z"
    },
    {
      "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": "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": 0.5870612380952381,
      "accuracy_std_dev": 0.09923890865789474,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.6539967813852815,
      "avg_duration_per_task_ms": 491814.81192881195,
      "avg_output_tokens_per_task": 40702.665704665706,
      "last_run_timestamp": "2026-09-19T09:08:58.129Z"
    },
    {
      "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": 0.5867004444444445,
      "accuracy_std_dev": 0.030454637959544256,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.002968678451178451,
      "avg_duration_per_task_ms": 201841.53367003368,
      "avg_output_tokens_per_task": 10450.568181818182,
      "last_run_timestamp": "2026-08-29T19:16:45.297Z"
    },
    {
      "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": 0.5824918571428572,
      "accuracy_std_dev": 0.09639527555617557,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.6154766688311687,
      "avg_duration_per_task_ms": 632957.1277056277,
      "avg_output_tokens_per_task": 40886.683982683986,
      "last_run_timestamp": "2026-09-19T09:11:28.823Z"
    },
    {
      "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": 0.5820909628506444,
      "accuracy_std_dev": 0.10256844962388531,
      "run_count": 20,
      "total_questions": 3957,
      "avg_cost_per_task_usd": 0.6594307323730098,
      "avg_duration_per_task_ms": 597798.0687389437,
      "avg_output_tokens_per_task": 41132.2928986606,
      "last_run_timestamp": "2026-09-19T09:07:29.588Z"
    },
    {
      "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": 0.5774372016509435,
      "accuracy_std_dev": 0.01609338000361728,
      "run_count": 13,
      "total_questions": 2544,
      "avg_cost_per_task_usd": 0.005421983490566038,
      "avg_duration_per_task_ms": 524101.0585691824,
      "avg_output_tokens_per_task": 35942.22051886792,
      "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": "Phala",
      "accuracy": 0.5772384474734041,
      "accuracy_std_dev": 0.11649000927593507,
      "run_count": 11,
      "total_questions": 1504,
      "avg_cost_per_task_usd": 0.004710451462765957,
      "avg_duration_per_task_ms": 563810.3085106383,
      "avg_output_tokens_per_task": 32534.692819148935,
      "last_run_timestamp": "2026-08-29T17:24:11.557Z"
    },
    {
      "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": 0.5767835180722893,
      "accuracy_std_dev": 0.11488998007127417,
      "run_count": 18,
      "total_questions": 3486,
      "avg_cost_per_task_usd": 0.001997228341939185,
      "avg_duration_per_task_ms": 61275.66838783706,
      "avg_output_tokens_per_task": 1863.6497418244405,
      "last_run_timestamp": "2026-08-29T20:57:58.443Z"
    },
    {
      "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": 0.5764792857142857,
      "accuracy_std_dev": 0.10228809062044081,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.6180922987012987,
      "avg_duration_per_task_ms": 636696.5533910533,
      "avg_output_tokens_per_task": 41167.0367965368,
      "last_run_timestamp": "2026-09-19T08:49:03.801Z"
    },
    {
      "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": 0.5726323012863915,
      "accuracy_std_dev": 0.036706189578649295,
      "run_count": 23,
      "total_questions": 4431,
      "avg_cost_per_task_usd": 0.056247474159331984,
      "avg_duration_per_task_ms": 424509.57323403296,
      "avg_output_tokens_per_task": 17825.642744301513,
      "last_run_timestamp": "2026-08-29T19:11:06.196Z"
    },
    {
      "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": 0.5675273333333334,
      "accuracy_std_dev": 0.09565683495356374,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.6112196195286197,
      "avg_duration_per_task_ms": 627120.2558922559,
      "avg_output_tokens_per_task": 40643.95005611672,
      "last_run_timestamp": "2026-09-19T08:58:01.146Z"
    },
    {
      "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": 0.5565433914539719,
      "accuracy_std_dev": 0.11305693050580784,
      "run_count": 18,
      "total_questions": 3487,
      "avg_cost_per_task_usd": 0.06438679782047606,
      "avg_duration_per_task_ms": 679986.9673071408,
      "avg_output_tokens_per_task": 20386.74533983367,
      "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-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Google",
      "accuracy": 0.5547767757009346,
      "accuracy_std_dev": 0.0771476524984634,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.007002486370716511,
      "avg_duration_per_task_ms": 230184.06152647975,
      "avg_output_tokens_per_task": 21060.570093457944,
      "last_run_timestamp": "2026-08-29T17:01:22.015Z"
    },
    {
      "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-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DigitalOcean",
      "accuracy": 0.5514001070195628,
      "accuracy_std_dev": 0.20533572929527444,
      "run_count": 6,
      "total_questions": 869,
      "avg_cost_per_task_usd": 0.05092492520138089,
      "avg_duration_per_task_ms": 1396586.3498273878,
      "avg_output_tokens_per_task": 19293.715765247412,
      "last_run_timestamp": "2026-08-22T17:09:42.218Z"
    },
    {
      "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": 0.5498265044082287,
      "accuracy_std_dev": 0.03556443036291385,
      "run_count": 21,
      "total_questions": 3743,
      "avg_cost_per_task_usd": 0.0300913881912904,
      "avg_duration_per_task_ms": 386568.3438418381,
      "avg_output_tokens_per_task": 14387.59177130644,
      "last_run_timestamp": "2026-08-29T19:07:49.385Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "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-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Phala",
      "accuracy": 0.5454545714285715,
      "accuracy_std_dev": 0.10688204941343671,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.004287935425685424,
      "avg_duration_per_task_ms": 239211.0404040404,
      "avg_output_tokens_per_task": 11229.227272727272,
      "last_run_timestamp": "2026-06-20T18:20:35.956Z"
    },
    {
      "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": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Phala",
      "accuracy": 0.5427914079754601,
      "accuracy_std_dev": 0.1644485316347865,
      "run_count": 17,
      "total_questions": 3260,
      "avg_cost_per_task_usd": 0.031211190490797545,
      "avg_duration_per_task_ms": 399121.22024539876,
      "avg_output_tokens_per_task": 11046.685889570552,
      "last_run_timestamp": "2026-07-18T16:15:33.587Z"
    },
    {
      "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": "Novita",
      "accuracy": 0.5424244,
      "accuracy_std_dev": 0.020353179334934343,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.004261171717171717,
      "avg_duration_per_task_ms": 201511.99797979798,
      "avg_output_tokens_per_task": 9415.110101010101,
      "last_run_timestamp": "2026-05-23T11:11:04.972Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "DeepInfra",
      "accuracy": 0.5412917072553431,
      "accuracy_std_dev": 0.15868845181824737,
      "run_count": 18,
      "total_questions": 3556,
      "avg_cost_per_task_usd": 0.0018783034308211474,
      "avg_duration_per_task_ms": 76872.3416760405,
      "avg_output_tokens_per_task": 2062.933914510686,
      "last_run_timestamp": "2026-08-29T20:34:10.117Z"
    },
    {
      "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": 0.5402590995515695,
      "accuracy_std_dev": 0.1064112241873893,
      "run_count": 17,
      "total_questions": 3345,
      "avg_cost_per_task_usd": 0.034116275635276534,
      "avg_duration_per_task_ms": 482865.85979073244,
      "avg_output_tokens_per_task": 12392.341405082212,
      "last_run_timestamp": "2026-07-11T16:02:27.787Z"
    },
    {
      "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": 0.5350947692307693,
      "accuracy_std_dev": 0.014328365504260027,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.0032667284382284375,
      "avg_duration_per_task_ms": 72101.14724164724,
      "avg_output_tokens_per_task": 7056.418803418804,
      "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": "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": "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": 0.5312429141104295,
      "accuracy_std_dev": 0.042893663569032354,
      "run_count": 15,
      "total_questions": 2934,
      "avg_cost_per_task_usd": 0.01798347137014315,
      "avg_duration_per_task_ms": 733666.4400136332,
      "avg_output_tokens_per_task": 45997.68916155419,
      "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-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.5303032857142858,
      "accuracy_std_dev": 0.016421324661325395,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0034015209235209233,
      "avg_duration_per_task_ms": 115367.47474747474,
      "avg_output_tokens_per_task": 6747.353535353535,
      "last_run_timestamp": "2026-07-01T03:39:59.085Z"
    },
    {
      "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": "ModelRun",
      "accuracy": 0.5290405,
      "accuracy_std_dev": 0.029959029561219083,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0006086212121212121,
      "avg_duration_per_task_ms": 25756.060606060608,
      "avg_output_tokens_per_task": 699.3156565656566,
      "last_run_timestamp": "2026-06-20T16:15:10.268Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.5282828,
      "accuracy_std_dev": 0.025272813831467202,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.004160109090909091,
      "avg_duration_per_task_ms": 148058.1797979798,
      "avg_output_tokens_per_task": 8276.889898989899,
      "last_run_timestamp": "2026-06-20T17:38:01.923Z"
    },
    {
      "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": 0.5223280439238653,
      "accuracy_std_dev": 0.19977793037975025,
      "run_count": 7,
      "total_questions": 1366,
      "avg_cost_per_task_usd": 0.02168075768667643,
      "avg_duration_per_task_ms": 693100.055636896,
      "avg_output_tokens_per_task": 30376.870424597364,
      "last_run_timestamp": "2026-09-19T16:22:29.649Z"
    },
    {
      "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": "z-ai/glm-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Ambient",
      "accuracy": 0.5176769999999999,
      "accuracy_std_dev": 0.007575999999999972,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.009877323232323232,
      "avg_duration_per_task_ms": 465993.38636363635,
      "avg_output_tokens_per_task": 18118.93181818182,
      "last_run_timestamp": "2026-06-13T11:17:16.955Z"
    },
    {
      "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": "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": 0.5171403267929634,
      "accuracy_std_dev": 0.04098216724042571,
      "run_count": 24,
      "total_questions": 4434,
      "avg_cost_per_task_usd": 0.009192668245376635,
      "avg_duration_per_task_ms": 398119.1844835363,
      "avg_output_tokens_per_task": 23257.760261614796,
      "last_run_timestamp": "2026-08-30T00:11:29.373Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4o-2024-08-06",
      "model_slug": "openai/gpt-4o-2024-08-06",
      "model_name": "OpenAI: GPT-4o (2024-08-06)",
      "provider_name": "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.5154740415003991,
      "accuracy_std_dev": 0.0612012483160448,
      "run_count": 21,
      "total_questions": 3759,
      "avg_cost_per_task_usd": 0.0585535027932961,
      "avg_duration_per_task_ms": 505974.45304602286,
      "avg_output_tokens_per_task": 16676.524607608408,
      "last_run_timestamp": "2026-08-29T18:52:09.401Z"
    },
    {
      "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": 0.5108945297941495,
      "accuracy_std_dev": 0.01301865738320295,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.004122480678945469,
      "avg_duration_per_task_ms": 117292.87612856626,
      "avg_output_tokens_per_task": 8300.09245214879,
      "last_run_timestamp": "2026-08-01T18:08:24.732Z"
    },
    {
      "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": "Friendli",
      "accuracy": 0.510101,
      "accuracy_std_dev": 0.010101000000000027,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0006397070707070707,
      "avg_duration_per_task_ms": 7448.810606060606,
      "avg_output_tokens_per_task": 796.5252525252525,
      "last_run_timestamp": "2026-05-09T08:07:34.813Z"
    },
    {
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.5053764679900744,
      "accuracy_std_dev": 0.20352934331044298,
      "run_count": 11,
      "total_questions": 2015,
      "avg_cost_per_task_usd": 0.01812188188585608,
      "avg_duration_per_task_ms": 1793360.2853598015,
      "avg_output_tokens_per_task": 13961.292803970224,
      "last_run_timestamp": "2026-08-22T17:37:26.215Z"
    },
    {
      "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": "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": 0.5049259153605015,
      "accuracy_std_dev": 0.11693449444168029,
      "run_count": 13,
      "total_questions": 2233,
      "avg_cost_per_task_usd": 0.008191620241827138,
      "avg_duration_per_task_ms": 482531.36945812806,
      "avg_output_tokens_per_task": 40792.52754142409,
      "last_run_timestamp": "2026-09-19T16:17:10.003Z"
    },
    {
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Together",
      "accuracy": 0.5009998431876606,
      "accuracy_std_dev": 0.23720298614056273,
      "run_count": 6,
      "total_questions": 1167,
      "avg_cost_per_task_usd": 0.004498337617823479,
      "avg_duration_per_task_ms": 250896.1996572408,
      "avg_output_tokens_per_task": 23117.198800342758,
      "last_run_timestamp": "2026-07-04T16:00:23.384Z"
    },
    {
      "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": 0.4959782777777778,
      "accuracy_std_dev": 0.02300204246439761,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.0006973439955106623,
      "avg_duration_per_task_ms": 43601.93097643097,
      "avg_output_tokens_per_task": 1512.7345679012346,
      "last_run_timestamp": "2026-08-29T19:21:45.914Z"
    },
    {
      "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": 0.4956976666666667,
      "accuracy_std_dev": 0.01927874840624025,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.0016561282267115599,
      "avg_duration_per_task_ms": 7718.37682379349,
      "avg_output_tokens_per_task": 1641.1672278338945,
      "last_run_timestamp": "2026-08-29T19:15:06.404Z"
    },
    {
      "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": 0.4934997222222222,
      "accuracy_std_dev": 0.02101166846409543,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.0017448033108866444,
      "avg_duration_per_task_ms": 10151.609988776656,
      "avg_output_tokens_per_task": 1588.8221099887767,
      "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-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "SambaNova",
      "accuracy": 0.49248677460142937,
      "accuracy_std_dev": 0.1468475282129393,
      "run_count": 12,
      "total_questions": 1819,
      "avg_cost_per_task_usd": 0.02537227047828477,
      "avg_duration_per_task_ms": 43723.59538207806,
      "avg_output_tokens_per_task": 5430.2094557449145,
      "last_run_timestamp": "2026-08-29T16:54:06.286Z"
    },
    {
      "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": 0.4907005714285715,
      "accuracy_std_dev": 0.024199875069672482,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.0008166632996632996,
      "avg_duration_per_task_ms": 41913.33645983646,
      "avg_output_tokens_per_task": 1437.7967772967772,
      "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-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "Venice",
      "accuracy": 0.4900520065566357,
      "accuracy_std_dev": 0.037634349752895334,
      "run_count": 23,
      "total_questions": 4423,
      "avg_cost_per_task_usd": 0.007361529957042731,
      "avg_duration_per_task_ms": 281581.36536287586,
      "avg_output_tokens_per_task": 15918.30137915442,
      "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": "Together",
      "accuracy": 0.4897118333333333,
      "accuracy_std_dev": 0.027081268771052807,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.0022169312570145903,
      "avg_duration_per_task_ms": 39178.599326599324,
      "avg_output_tokens_per_task": 1582.158810325477,
      "last_run_timestamp": "2026-08-29T18:45:20.686Z"
    },
    {
      "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": "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": 0.4874045119744545,
      "accuracy_std_dev": 0.019802710459293865,
      "run_count": 19,
      "total_questions": 3758,
      "avg_cost_per_task_usd": 0.0007857953698775944,
      "avg_duration_per_task_ms": 54493.003459286854,
      "avg_output_tokens_per_task": 1787.7637040979243,
      "last_run_timestamp": "2026-08-29T19:00:36.673Z"
    },
    {
      "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": 0.48552127324913896,
      "accuracy_std_dev": 0.10753376064866745,
      "run_count": 14,
      "total_questions": 2613,
      "avg_cost_per_task_usd": 0.010886646000765405,
      "avg_duration_per_task_ms": 309006.7183314198,
      "avg_output_tokens_per_task": 14771.444316877152,
      "last_run_timestamp": "2026-08-01T18:13:08.069Z"
    },
    {
      "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": 0.48484835714285707,
      "accuracy_std_dev": 0.019539890950884275,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.0012388686868686867,
      "avg_duration_per_task_ms": 18026.963203463205,
      "avg_output_tokens_per_task": 1322.105339105339,
      "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": "Inceptron",
      "accuracy": 0.4848483333333333,
      "accuracy_std_dev": 0.01797495489717725,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.00031208922558922556,
      "avg_duration_per_task_ms": 18182.459595959597,
      "avg_output_tokens_per_task": 730.1346801346801,
      "last_run_timestamp": "2026-05-16T08:12:45.858Z"
    },
    {
      "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": 0.4848048041942605,
      "accuracy_std_dev": 0.0329764460112243,
      "run_count": 23,
      "total_questions": 4530,
      "avg_cost_per_task_usd": 0.006697250110375276,
      "avg_duration_per_task_ms": 285389.3573951435,
      "avg_output_tokens_per_task": 16704.50375275938,
      "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": "Parasail",
      "accuracy": 0.4829639762922659,
      "accuracy_std_dev": 0.0222392355589962,
      "run_count": 13,
      "total_questions": 2573,
      "avg_cost_per_task_usd": 0.0010490423630003885,
      "avg_duration_per_task_ms": 37775.31752817723,
      "avg_output_tokens_per_task": 1882.54061406918,
      "last_run_timestamp": "2026-08-29T17:00:17.733Z"
    },
    {
      "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": "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": "DeepInfra",
      "accuracy": 0.4761502777777778,
      "accuracy_std_dev": 0.019858171124825148,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.0006140196408529742,
      "avg_duration_per_task_ms": 115870.00168350169,
      "avg_output_tokens_per_task": 1751.4741863075196,
      "last_run_timestamp": "2026-08-29T18:48:55.759Z"
    },
    {
      "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": 0.4751962666666667,
      "accuracy_std_dev": 0.01117488509988159,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.0007880821548821547,
      "avg_duration_per_task_ms": 116358.99057239057,
      "avg_output_tokens_per_task": 1780.7713804713806,
      "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": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "AtlasCloud",
      "accuracy": 0.46464666666666665,
      "accuracy_std_dev": 0.24034588812283755,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0018368636363636363,
      "avg_duration_per_task_ms": 32878.122895622895,
      "avg_output_tokens_per_task": 3002.4276094276092,
      "last_run_timestamp": "2026-06-30T19:37:23.430Z"
    },
    {
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Google",
      "accuracy": 0.4557567443438914,
      "accuracy_std_dev": 0.195794758968546,
      "run_count": 7,
      "total_questions": 1326,
      "avg_cost_per_task_usd": 0.002924125942684766,
      "avg_duration_per_task_ms": 340072.5972850679,
      "avg_output_tokens_per_task": 15733.245852187029,
      "last_run_timestamp": "2026-07-11T15:34:14.755Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.44245509090909096,
      "accuracy_std_dev": 0.40072706253791307,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.0037717125803489442,
      "avg_duration_per_task_ms": 180639.4880624426,
      "avg_output_tokens_per_task": 3668.438934802571,
      "last_run_timestamp": "2026-08-15T14:54:58.071Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "Novita",
      "accuracy": 0.41662231578947373,
      "accuracy_std_dev": 0.25537013940925096,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.0011135746943115364,
      "avg_duration_per_task_ms": 32676.62706007443,
      "avg_output_tokens_per_task": 946.1113769271664,
      "last_run_timestamp": "2026-08-29T18:43:09.114Z"
    },
    {
      "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/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": "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": "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": "auto-routing",
      "accuracy": 0.3156760930232559,
      "accuracy_std_dev": 0.020895107938223238,
      "run_count": 8,
      "total_questions": 1548,
      "avg_cost_per_task_usd": 0.00008417894056847546,
      "avg_duration_per_task_ms": 8191.765503875969,
      "avg_output_tokens_per_task": 362.62467700258395,
      "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-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "Nebius",
      "accuracy": 0.3131315,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.00008114646464646465,
      "avg_duration_per_task_ms": 46871.881313131315,
      "avg_output_tokens_per_task": 1262.8939393939395,
      "last_run_timestamp": "2026-05-09T08:25:54.935Z"
    },
    {
      "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": 0.31154682352941176,
      "accuracy_std_dev": 0.04799129866994309,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.005848113784907903,
      "avg_duration_per_task_ms": 288872.33244206774,
      "avg_output_tokens_per_task": 13122.53505644682,
      "last_run_timestamp": "2026-08-29T19:48:22.407Z"
    },
    {
      "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": 0.30923066923076925,
      "accuracy_std_dev": 0.02667595582842347,
      "run_count": 10,
      "total_questions": 1950,
      "avg_cost_per_task_usd": 0.0000934446153846154,
      "avg_duration_per_task_ms": 5045.05282051282,
      "avg_output_tokens_per_task": 300.8317948717949,
      "last_run_timestamp": "2026-07-18T09:09:45.839Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Cloudflare",
      "accuracy": 0.30896005292479106,
      "accuracy_std_dev": 0.05550927626372292,
      "run_count": 11,
      "total_questions": 2154,
      "avg_cost_per_task_usd": 0.00260005756731662,
      "avg_duration_per_task_ms": 28138.977251624885,
      "avg_output_tokens_per_task": 1086.3834726090993,
      "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": "Novita",
      "accuracy": 0.3056295271111111,
      "accuracy_std_dev": 0.05792376099499717,
      "run_count": 6,
      "total_questions": 1125,
      "avg_cost_per_task_usd": 0.00008087733333333333,
      "avg_duration_per_task_ms": 36138.544,
      "avg_output_tokens_per_task": 344.34222222222223,
      "last_run_timestamp": "2026-07-18T08:00:15.379Z"
    },
    {
      "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": 0.3026373112840467,
      "accuracy_std_dev": 0.026204171217336758,
      "run_count": 8,
      "total_questions": 1542,
      "avg_cost_per_task_usd": 0.00002046757457846952,
      "avg_duration_per_task_ms": 9360.384565499351,
      "avg_output_tokens_per_task": 275.0188067444877,
      "last_run_timestamp": "2026-07-18T08:00:15.193Z"
    },
    {
      "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": 0.29322554528061223,
      "accuracy_std_dev": 0.026492655735409588,
      "run_count": 24,
      "total_questions": 4704,
      "avg_cost_per_task_usd": 0.00015754315476190474,
      "avg_duration_per_task_ms": 151689.59715136053,
      "avg_output_tokens_per_task": 3673.2096088435374,
      "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": "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": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "DekaLLM",
      "accuracy": 0.2874569642004773,
      "accuracy_std_dev": 0.027560829670429898,
      "run_count": 7,
      "total_questions": 1257,
      "avg_cost_per_task_usd": 0.000016271280827366746,
      "avg_duration_per_task_ms": 11521.354017501988,
      "avg_output_tokens_per_task": 197.71917263325378,
      "last_run_timestamp": "2026-07-18T08:04:03.449Z"
    },
    {
      "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": 0.28451170588235297,
      "accuracy_std_dev": 0.025856077905943135,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.0006791901366607251,
      "avg_duration_per_task_ms": 47282.23143196673,
      "avg_output_tokens_per_task": 8804.306892453951,
      "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": "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": "Cerebras",
      "accuracy": 0.2727275,
      "accuracy_std_dev": 0.035353499999999996,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.00008969696969696971,
      "avg_duration_per_task_ms": 633.3914141414141,
      "avg_output_tokens_per_task": 628.320707070707,
      "last_run_timestamp": "2026-05-23T08:04:59.405Z"
    },
    {
      "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": "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": "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": 0.26757776470588235,
      "accuracy_std_dev": 0.021117749736764967,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.000780840760546643,
      "avg_duration_per_task_ms": 19816.014557338087,
      "avg_output_tokens_per_task": 9728.812834224598,
      "last_run_timestamp": "2026-08-29T18:45:59.793Z"
    },
    {
      "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": 0.256419686746988,
      "accuracy_std_dev": 0.07325243553772927,
      "run_count": 14,
      "total_questions": 2739,
      "avg_cost_per_task_usd": 0.001453557867834976,
      "avg_duration_per_task_ms": 454232.56005841546,
      "avg_output_tokens_per_task": 5530.41730558598,
      "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-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "WandB",
      "accuracy": 0.2561763187294633,
      "accuracy_std_dev": 0.022252850067062038,
      "run_count": 14,
      "total_questions": 2739,
      "avg_cost_per_task_usd": 0.0006673464768163562,
      "avg_duration_per_task_ms": 19297.468054034318,
      "avg_output_tokens_per_task": 2614.9587440671776,
      "last_run_timestamp": "2026-07-18T08:26:34.893Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.245614,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 171,
      "avg_cost_per_task_usd": 0.03239533333333334,
      "avg_duration_per_task_ms": 323216.8245614035,
      "avg_output_tokens_per_task": 11659.74269005848,
      "last_run_timestamp": "2026-06-30T21:29:29.599Z"
    },
    {
      "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": "Friendli",
      "accuracy": 0.242424,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00020296464646464647,
      "avg_duration_per_task_ms": 11928.080808080807,
      "avg_output_tokens_per_task": 1760.0050505050506,
      "last_run_timestamp": "2026-05-09T08:06:13.379Z"
    },
    {
      "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": 0.24149673724489795,
      "accuracy_std_dev": 0.024442688065230603,
      "run_count": 12,
      "total_questions": 2352,
      "avg_cost_per_task_usd": 0.00032777253401360546,
      "avg_duration_per_task_ms": 55241.83588435374,
      "avg_output_tokens_per_task": 6268.276785714285,
      "last_run_timestamp": "2026-08-29T16:55:14.828Z"
    },
    {
      "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": 0.2372174911504425,
      "accuracy_std_dev": 0.3611729272686932,
      "run_count": 7,
      "total_questions": 1356,
      "avg_cost_per_task_usd": 0.0035571969026548674,
      "avg_duration_per_task_ms": 170613.9918879056,
      "avg_output_tokens_per_task": 3415.4498525073745,
      "last_run_timestamp": "2026-08-22T17:29:06.480Z"
    },
    {
      "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": "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": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.806667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.20567228,
      "avg_duration_per_task_ms": 69455.26,
      "avg_output_tokens_per_task": 15737.28,
      "last_run_timestamp": "2026-08-15T00:15:11.753Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Azure",
      "accuracy": 0.8057142857142857,
      "accuracy_std_dev": 0.02209041203595731,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.8699426371428572,
      "avg_duration_per_task_ms": 139011.88,
      "avg_output_tokens_per_task": 5785.405714285715,
      "last_run_timestamp": "2026-09-19T08:15:00.954Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.8,
      "accuracy_std_dev": 0.02805880446490904,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.9499947457142858,
      "avg_duration_per_task_ms": 132744.98857142858,
      "avg_output_tokens_per_task": 5787.242857142857,
      "last_run_timestamp": "2026-09-19T08:01:53.305Z"
    },
    {
      "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": "Claude Platform on AWS",
      "accuracy": 0.7973334,
      "accuracy_std_dev": 0.014360451867542312,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.873586176,
      "avg_duration_per_task_ms": 155291.072,
      "avg_output_tokens_per_task": 5726.508,
      "last_run_timestamp": "2026-09-19T08:14:32.378Z"
    },
    {
      "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": "Nebius",
      "accuracy": 0.7966666666666666,
      "accuracy_std_dev": 0.021343747458109477,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08247077333333334,
      "avg_duration_per_task_ms": 126277.50333333333,
      "avg_output_tokens_per_task": 8287.77,
      "last_run_timestamp": "2026-05-23T17:33:06.915Z"
    },
    {
      "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": 0.795,
      "accuracy_std_dev": 0.019649399202011232,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.104553105,
      "avg_duration_per_task_ms": 447536.345,
      "avg_output_tokens_per_task": 18774.07,
      "last_run_timestamp": "2026-09-19T08:12:56.605Z"
    },
    {
      "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": 0.7942857142857143,
      "accuracy_std_dev": 0.014443616234114186,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.8340979885714286,
      "avg_duration_per_task_ms": 141474.59714285715,
      "avg_output_tokens_per_task": 5747.588571428571,
      "last_run_timestamp": "2026-09-19T08:01:53.308Z"
    },
    {
      "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": 0.7933333333333333,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.8155115399999999,
      "avg_duration_per_task_ms": 189088.09333333332,
      "avg_output_tokens_per_task": 7070.38,
      "last_run_timestamp": "2026-09-19T08:02:29.967Z"
    },
    {
      "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": 0.792381,
      "accuracy_std_dev": 0.022375873792994085,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.49301859142857146,
      "avg_duration_per_task_ms": 131851.76571428572,
      "avg_output_tokens_per_task": 7199.7,
      "last_run_timestamp": "2026-09-19T08:01:53.318Z"
    },
    {
      "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": 0.79,
      "accuracy_std_dev": 0.003332999999999975,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.08253803,
      "avg_duration_per_task_ms": 85518.8,
      "avg_output_tokens_per_task": 4440.97,
      "last_run_timestamp": "2026-09-19T08:01:53.363Z"
    },
    {
      "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": 0.7884616923076923,
      "accuracy_std_dev": 0.02158726126280039,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.9622519753846155,
      "avg_duration_per_task_ms": 100478.9923076923,
      "avg_output_tokens_per_task": 5605.658461538462,
      "last_run_timestamp": "2026-09-19T08:15:41.519Z"
    },
    {
      "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": "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Wafer",
      "accuracy": 0.7866666,
      "accuracy_std_dev": 0.005962698872155093,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.118893224,
      "avg_duration_per_task_ms": 130658.76,
      "avg_output_tokens_per_task": 11469.216,
      "last_run_timestamp": "2026-09-20T10:30:07.954Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Morph",
      "accuracy": 0.7858823529411765,
      "accuracy_std_dev": 0.0414686149812562,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.08477611529411765,
      "avg_duration_per_task_ms": 198992.7294117647,
      "avg_output_tokens_per_task": 11058.981176470588,
      "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": "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": "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": 0.78533336,
      "accuracy_std_dev": 0.0372379576957491,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.10169507999999998,
      "avg_duration_per_task_ms": 377017.644,
      "avg_output_tokens_per_task": 16270.1992,
      "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": "Z.AI",
      "accuracy": 0.785,
      "accuracy_std_dev": 0.010928810045014038,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.08068941,
      "avg_duration_per_task_ms": 125123.275,
      "avg_output_tokens_per_task": 4412.62,
      "last_run_timestamp": "2026-09-19T08:07:46.498Z"
    },
    {
      "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": 0.7838095714285714,
      "accuracy_std_dev": 0.023599102563790993,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.46679962,
      "avg_duration_per_task_ms": 112990.85857142857,
      "avg_output_tokens_per_task": 7266.794285714286,
      "last_run_timestamp": "2026-09-19T08:08:46.293Z"
    },
    {
      "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": 0.7833334999999999,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.7261046600000001,
      "avg_duration_per_task_ms": 117322.17,
      "avg_output_tokens_per_task": 6966.33,
      "last_run_timestamp": "2026-09-19T08:08:24.219Z"
    },
    {
      "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": 0.7833334999999999,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.08340285,
      "avg_duration_per_task_ms": 39163.99,
      "avg_output_tokens_per_task": 4681.96,
      "last_run_timestamp": "2026-09-19T08:11:03.394Z"
    },
    {
      "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": 0.7822223333333334,
      "accuracy_std_dev": 0.013698571612479267,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06831790666666666,
      "avg_duration_per_task_ms": 97592.92,
      "avg_output_tokens_per_task": 11907.126666666667,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Fireworks",
      "accuracy": 0.7822223333333334,
      "accuracy_std_dev": 0.031427125187575726,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.10594911333333333,
      "avg_duration_per_task_ms": 72818.32666666666,
      "avg_output_tokens_per_task": 4800.32,
      "last_run_timestamp": "2026-09-12T08:12:03.011Z"
    },
    {
      "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": 0.7816665,
      "accuracy_std_dev": 0.028431350069069865,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.15782728499999998,
      "avg_duration_per_task_ms": 597754.365,
      "avg_output_tokens_per_task": 12437.015,
      "last_run_timestamp": "2026-09-19T08:01:36.454Z"
    },
    {
      "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": 0.7809090909090909,
      "accuracy_std_dev": 0.024968384418025142,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.5751214472727272,
      "avg_duration_per_task_ms": 120619.24545454545,
      "avg_output_tokens_per_task": 7275.691818181818,
      "last_run_timestamp": "2026-09-19T08:15:48.414Z"
    },
    {
      "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": 0.7806349047619048,
      "accuracy_std_dev": 0.023267850671002527,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.48060664571428574,
      "avg_duration_per_task_ms": 128644.07238095238,
      "avg_output_tokens_per_task": 7184.6923809523805,
      "last_run_timestamp": "2026-09-19T08:12:50.307Z"
    },
    {
      "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": 0.78,
      "accuracy_std_dev": 0.021081640250859665,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.016256343333333336,
      "avg_duration_per_task_ms": 172479.31,
      "avg_output_tokens_per_task": 11984.07,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.78,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03263777,
      "avg_duration_per_task_ms": 159319.31,
      "avg_output_tokens_per_task": 15620.74,
      "last_run_timestamp": "2026-09-19T08:12:48.205Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Phala",
      "accuracy": 0.78,
      "accuracy_std_dev": 1.1102230246251565e-16,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.14591014000000002,
      "avg_duration_per_task_ms": 245617.41333333333,
      "avg_output_tokens_per_task": 6015.373333333333,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Fireworks",
      "accuracy": 0.78,
      "accuracy_std_dev": 0.028284271247461888,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.031875935,
      "avg_duration_per_task_ms": 94447.85,
      "avg_output_tokens_per_task": 3704.84,
      "last_run_timestamp": "2026-04-25T14:38:47.693Z"
    },
    {
      "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": 0.77916675,
      "accuracy_std_dev": 0.022715135068660724,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.4318646125000001,
      "avg_duration_per_task_ms": 135663.39,
      "avg_output_tokens_per_task": 7317.8775,
      "last_run_timestamp": "2026-09-19T08:14:21.987Z"
    },
    {
      "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": 0.7789859999999998,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 46,
      "avg_cost_per_task_usd": 0.333418,
      "avg_duration_per_task_ms": 227429.76086956522,
      "avg_output_tokens_per_task": 5856.152173913043,
      "last_run_timestamp": "2026-08-08T08:09:44.831Z"
    },
    {
      "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": 0.777778,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.7786092066666667,
      "avg_duration_per_task_ms": 116605.5,
      "avg_output_tokens_per_task": 6845.246666666667,
      "last_run_timestamp": "2026-09-19T08:02:40.695Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ambient",
      "accuracy": 0.777778,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.012714583333333333,
      "avg_duration_per_task_ms": 516568.7291666667,
      "avg_output_tokens_per_task": 14230.208333333334,
      "last_run_timestamp": "2026-08-03T17:20:47.207Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "BaseTen",
      "accuracy": 0.777777825958702,
      "accuracy_std_dev": 0.019674442666690654,
      "run_count": 7,
      "total_questions": 339,
      "avg_cost_per_task_usd": 0.06404329498525074,
      "avg_duration_per_task_ms": 85698.01474926254,
      "avg_output_tokens_per_task": 4107.47197640118,
      "last_run_timestamp": "2026-09-19T08:08:24.135Z"
    },
    {
      "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": "Together",
      "accuracy": 0.7775,
      "accuracy_std_dev": 0.028173569173961615,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.12033262749999998,
      "avg_duration_per_task_ms": 104051.4975,
      "avg_output_tokens_per_task": 15655.525,
      "last_run_timestamp": "2026-06-20T17:08:55.950Z"
    },
    {
      "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": 0.7773333,
      "accuracy_std_dev": 0.014047595289230106,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.434012314,
      "avg_duration_per_task_ms": 97749.408,
      "avg_output_tokens_per_task": 5681.338,
      "last_run_timestamp": "2026-09-19T08:07:34.969Z"
    },
    {
      "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": 0.7771428571428571,
      "accuracy_std_dev": 0.030207219333078877,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.47215102000000003,
      "avg_duration_per_task_ms": 119873.53142857143,
      "avg_output_tokens_per_task": 7280.5142857142855,
      "last_run_timestamp": "2026-09-19T08:15:47.962Z"
    },
    {
      "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": 0.7770272162162162,
      "accuracy_std_dev": 0.015058546012148703,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.17008879054054055,
      "avg_duration_per_task_ms": 358528.91891891893,
      "avg_output_tokens_per_task": 20610.358108108107,
      "last_run_timestamp": "2026-09-19T08:15:57.539Z"
    },
    {
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Io Net",
      "accuracy": 0.7766665,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06410135,
      "avg_duration_per_task_ms": 123064.78,
      "avg_output_tokens_per_task": 4532.5,
      "last_run_timestamp": "2026-09-19T08:14:24.978Z"
    },
    {
      "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": 0.7760001999999999,
      "accuracy_std_dev": 0.041601285236877006,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.08482736,
      "avg_duration_per_task_ms": 237369.612,
      "avg_output_tokens_per_task": 11416.556,
      "last_run_timestamp": "2026-09-19T08:07:09.158Z"
    },
    {
      "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": 0.775,
      "accuracy_std_dev": 0.025549537941027427,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.007826415,
      "avg_duration_per_task_ms": 211518.38,
      "avg_output_tokens_per_task": 4265.165,
      "last_run_timestamp": "2026-09-19T08:02:29.952Z"
    },
    {
      "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": 0.7749997500000001,
      "accuracy_std_dev": 0.021794360891925678,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.21402777999999997,
      "avg_duration_per_task_ms": 289883.17,
      "avg_output_tokens_per_task": 10604.905,
      "last_run_timestamp": "2026-09-19T08:06:32.644Z"
    },
    {
      "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": 0.7747619285714286,
      "accuracy_std_dev": 0.03189058782655896,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.44453945142857143,
      "avg_duration_per_task_ms": 80162.03857142857,
      "avg_output_tokens_per_task": 5717.024285714286,
      "last_run_timestamp": "2026-09-19T08:12:06.974Z"
    },
    {
      "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": 0.7746666,
      "accuracy_std_dev": 0.025785425686616063,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.473358384,
      "avg_duration_per_task_ms": 128837.64,
      "avg_output_tokens_per_task": 7803.012,
      "last_run_timestamp": "2026-09-19T08:15:49.500Z"
    },
    {
      "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": 0.7744443333333334,
      "accuracy_std_dev": 0.035986907770589154,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06337854666666667,
      "avg_duration_per_task_ms": 170014.64,
      "avg_output_tokens_per_task": 7038.76,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "StreamLake",
      "accuracy": 0.7740476428571429,
      "accuracy_std_dev": 0.027736848735950892,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.030251097857142858,
      "avg_duration_per_task_ms": 120961.04571428572,
      "avg_output_tokens_per_task": 5680.855,
      "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-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Wafer",
      "accuracy": 0.7733335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.08473178,
      "avg_duration_per_task_ms": 234930.66,
      "avg_output_tokens_per_task": 22523.52,
      "last_run_timestamp": "2026-09-20T10:30:07.955Z"
    },
    {
      "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": "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Decart",
      "accuracy": 0.7733333333333333,
      "accuracy_std_dev": 0.02177310607658502,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06427175333333333,
      "avg_duration_per_task_ms": 56109.4,
      "avg_output_tokens_per_task": 4673.253333333333,
      "last_run_timestamp": "2026-09-19T08:02:29.959Z"
    },
    {
      "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": 0.7733331666666667,
      "accuracy_std_dev": 0.00384890557676971,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.006986003333333334,
      "avg_duration_per_task_ms": 162086.69,
      "avg_output_tokens_per_task": 16532.63,
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "StreamLake",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.057549260000000005,
      "avg_duration_per_task_ms": 140710.34,
      "avg_output_tokens_per_task": 4743.94,
      "last_run_timestamp": "2026-09-11T04:09:34.110Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Azure",
      "accuracy": 0.7728203846153846,
      "accuracy_std_dev": 0.029289100594270586,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.49779368153846165,
      "avg_duration_per_task_ms": 130806.25846153845,
      "avg_output_tokens_per_task": 7936.776923076923,
      "last_run_timestamp": "2026-09-19T08:16:40.718Z"
    },
    {
      "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": 0.772727090909091,
      "accuracy_std_dev": 0.03440164644700108,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.14892930181818181,
      "avg_duration_per_task_ms": 546664.8418181818,
      "avg_output_tokens_per_task": 35137.674545454545,
      "last_run_timestamp": "2026-09-19T08:12:02.486Z"
    },
    {
      "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": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Azure",
      "accuracy": 0.7723808571428572,
      "accuracy_std_dev": 0.018662706092468933,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.4297011400000001,
      "avg_duration_per_task_ms": 103448.23714285715,
      "avg_output_tokens_per_task": 5585.251428571429,
      "last_run_timestamp": "2026-09-19T08:14:49.297Z"
    },
    {
      "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": 0.772,
      "accuracy_std_dev": 0.01808611499465822,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.137215892,
      "avg_duration_per_task_ms": 326138.736,
      "avg_output_tokens_per_task": 19302.048,
      "last_run_timestamp": "2026-09-19T08:05:35.724Z"
    },
    {
      "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": 0.772,
      "accuracy_std_dev": 0.01359732000064717,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.099907088,
      "avg_duration_per_task_ms": 347432.108,
      "avg_output_tokens_per_task": 19671.476,
      "last_run_timestamp": "2026-09-19T08:02:29.955Z"
    },
    {
      "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": 0.7714287142857142,
      "accuracy_std_dev": 0.024615140153480954,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.53567294,
      "avg_duration_per_task_ms": 174941.87714285715,
      "avg_output_tokens_per_task": 10495.22,
      "last_run_timestamp": "2026-09-19T08:07:29.514Z"
    },
    {
      "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": 0.7713334,
      "accuracy_std_dev": 0.0331394434343729,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.122264977,
      "avg_duration_per_task_ms": 407781.64,
      "avg_output_tokens_per_task": 17384.362,
      "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": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Anthropic",
      "accuracy": 0.7713333,
      "accuracy_std_dev": 0.0188680293992245,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.4928351919999999,
      "avg_duration_per_task_ms": 131035.412,
      "avg_output_tokens_per_task": 8091.076,
      "last_run_timestamp": "2026-09-19T08:08:44.449Z"
    },
    {
      "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": 0.7712177472324724,
      "accuracy_std_dev": 0.052442826700422794,
      "run_count": 11,
      "total_questions": 542,
      "avg_cost_per_task_usd": 0.09838513653136532,
      "avg_duration_per_task_ms": 1843075.8265682657,
      "avg_output_tokens_per_task": 19082.356088560886,
      "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": "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": 0.7711113333333333,
      "accuracy_std_dev": 0.02199873172606905,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08392440666666666,
      "avg_duration_per_task_ms": 191039.60666666666,
      "avg_output_tokens_per_task": 7574.006666666667,
      "last_run_timestamp": "2026-09-19T08:01:53.305Z"
    },
    {
      "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": "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": 0.7706666,
      "accuracy_std_dev": 0.024073989653565938,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.008760292,
      "avg_duration_per_task_ms": 280566.352,
      "avg_output_tokens_per_task": 6313.424,
      "last_run_timestamp": "2026-09-20T10:30:07.571Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Baidu",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07565745,
      "avg_duration_per_task_ms": 364682.91,
      "avg_output_tokens_per_task": 21233.7,
      "last_run_timestamp": "2026-09-19T08:01:53.308Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Cloudflare",
      "accuracy": 0.7699997500000001,
      "accuracy_std_dev": 0.015275197762631398,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.100920285,
      "avg_duration_per_task_ms": 129822.61,
      "avg_output_tokens_per_task": 4551.79,
      "last_run_timestamp": "2026-09-19T08:15:16.571Z"
    },
    {
      "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": 0.7695238571428571,
      "accuracy_std_dev": 0.017036849485986643,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.10181366857142858,
      "avg_duration_per_task_ms": 251127.43142857144,
      "avg_output_tokens_per_task": 7309.714285714285,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Parasail",
      "accuracy": 0.7694109317269077,
      "accuracy_std_dev": 0.035180887422576496,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.05336053413654618,
      "avg_duration_per_task_ms": 321533.1927710843,
      "avg_output_tokens_per_task": 6818.439759036145,
      "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-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7693334,
      "accuracy_std_dev": 0.029992482022000132,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.554491632,
      "avg_duration_per_task_ms": 184901.904,
      "avg_output_tokens_per_task": 10536.404,
      "last_run_timestamp": "2026-09-19T08:02:29.873Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.7688889999999999,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.10034226,
      "avg_duration_per_task_ms": 157612.40666666668,
      "avg_output_tokens_per_task": 8568.786666666667,
      "last_run_timestamp": "2026-06-30T19:14:03.389Z"
    },
    {
      "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": 0.7688889999999999,
      "accuracy_std_dev": 0.003142382535593046,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.07201002000000001,
      "avg_duration_per_task_ms": 47850.82666666667,
      "avg_output_tokens_per_task": 4678.52,
      "last_run_timestamp": "2026-09-19T08:15:44.326Z"
    },
    {
      "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": "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": 0.7686666,
      "accuracy_std_dev": 0.023438065906554666,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.058330564,
      "avg_duration_per_task_ms": 133631.792,
      "avg_output_tokens_per_task": 4186.012,
      "last_run_timestamp": "2026-09-19T08:16:41.275Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.7684563154362417,
      "accuracy_std_dev": 0.006017520216464235,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.1581116510067114,
      "avg_duration_per_task_ms": 537705.5637583892,
      "avg_output_tokens_per_task": 10979.71812080537,
      "last_run_timestamp": "2026-08-08T08:27:49.545Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.7683858338108882,
      "accuracy_std_dev": 0.03056258771966879,
      "run_count": 14,
      "total_questions": 698,
      "avg_cost_per_task_usd": 0.028291176217765036,
      "avg_duration_per_task_ms": 202040.56017191976,
      "avg_output_tokens_per_task": 8723.312320916906,
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.7683335,
      "accuracy_std_dev": 0.025980986639271425,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.074955525,
      "avg_duration_per_task_ms": 242542.055,
      "avg_output_tokens_per_task": 12922.96,
      "last_run_timestamp": "2026-09-19T08:01:36.428Z"
    },
    {
      "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": 0.767573537414966,
      "accuracy_std_dev": 0.02493140611535758,
      "run_count": 3,
      "total_questions": 147,
      "avg_cost_per_task_usd": 0.062283115646258505,
      "avg_duration_per_task_ms": 166794.86394557822,
      "avg_output_tokens_per_task": 4126.721088435374,
      "last_run_timestamp": "2026-09-19T08:15:43.811Z"
    },
    {
      "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": 0.7673684210526316,
      "accuracy_std_dev": 0.027435852827077042,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.06853673368421052,
      "avg_duration_per_task_ms": 181102.96,
      "avg_output_tokens_per_task": 7826.743157894737,
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7671691306532663,
      "accuracy_std_dev": 0.03712240837960737,
      "run_count": 16,
      "total_questions": 796,
      "avg_cost_per_task_usd": 0.11847029773869346,
      "avg_duration_per_task_ms": 600977.8140703518,
      "avg_output_tokens_per_task": 7982.221105527638,
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Chutes",
      "accuracy": 0.7670723286004056,
      "accuracy_std_dev": 0.034808554659020254,
      "run_count": 10,
      "total_questions": 493,
      "avg_cost_per_task_usd": 0.0794524705882353,
      "avg_duration_per_task_ms": 974078.9594320487,
      "avg_output_tokens_per_task": 25171.47261663286,
      "last_run_timestamp": "2026-09-19T16:00:24.105Z"
    },
    {
      "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": 0.7669025283018868,
      "accuracy_std_dev": 0.01965390363525289,
      "run_count": 17,
      "total_questions": 848,
      "avg_cost_per_task_usd": 0.06553669221698114,
      "avg_duration_per_task_ms": 138580.51533018867,
      "avg_output_tokens_per_task": 8122.042452830188,
      "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": "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Wafer",
      "accuracy": 0.7666669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.045774379999999996,
      "avg_duration_per_task_ms": 165566.44,
      "avg_output_tokens_per_task": 12426.8,
      "last_run_timestamp": "2026-09-20T21:00:04.179Z"
    },
    {
      "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": 0.76666675,
      "accuracy_std_dev": 0.028284389098714877,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.080202585,
      "avg_duration_per_task_ms": 46082.84,
      "avg_output_tokens_per_task": 4823.05,
      "last_run_timestamp": "2026-09-19T08:16:29.607Z"
    },
    {
      "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": 0.7666666666666667,
      "accuracy_std_dev": 0.019625871060640555,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.8067599599999999,
      "avg_duration_per_task_ms": 188419.76,
      "avg_output_tokens_per_task": 6949.133333333333,
      "last_run_timestamp": "2026-09-19T08:01:53.306Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Parasail",
      "accuracy": 0.7666666666666667,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08184152,
      "avg_duration_per_task_ms": 115682.96666666666,
      "avg_output_tokens_per_task": 4471.64,
      "last_run_timestamp": "2026-09-19T08:13:25.267Z"
    },
    {
      "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": 0.7666666,
      "accuracy_std_dev": 0.03910112246266086,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.07291049200000001,
      "avg_duration_per_task_ms": 332506.232,
      "avg_output_tokens_per_task": 12583.292,
      "last_run_timestamp": "2026-09-19T08:01:53.355Z"
    },
    {
      "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": 0.7666665454545455,
      "accuracy_std_dev": 0.014494959069612862,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.017875983636363633,
      "avg_duration_per_task_ms": 241228.09454545454,
      "avg_output_tokens_per_task": 17448.958181818183,
      "last_run_timestamp": "2026-09-19T08:08:46.406Z"
    },
    {
      "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": "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "AkashML",
      "accuracy": 0.766,
      "accuracy_std_dev": 0.0200997512422418,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.073923902,
      "avg_duration_per_task_ms": 82060.868,
      "avg_output_tokens_per_task": 4072.032,
      "last_run_timestamp": "2026-06-13T11:18:01.084Z"
    },
    {
      "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": 0.76599995,
      "accuracy_std_dev": 0.022201089564422297,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.5305423589999999,
      "avg_duration_per_task_ms": 122447.576,
      "avg_output_tokens_per_task": 8112.119,
      "last_run_timestamp": "2026-09-19T08:14:26.593Z"
    },
    {
      "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": 0.7658824117647058,
      "accuracy_std_dev": 0.041720086264163854,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.010050117647058824,
      "avg_duration_per_task_ms": 111496.44235294117,
      "avg_output_tokens_per_task": 10075.014117647059,
      "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/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7655556666666666,
      "accuracy_std_dev": 0.0355206634353708,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.23852161333333333,
      "avg_duration_per_task_ms": 442810.72,
      "avg_output_tokens_per_task": 14679.013333333334,
      "last_run_timestamp": "2026-09-19T08:02:40.690Z"
    },
    {
      "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": 0.7653968095238095,
      "accuracy_std_dev": 0.03808201053443309,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.051203074285714294,
      "avg_duration_per_task_ms": 167458.42666666667,
      "avg_output_tokens_per_task": 7624.500952380952,
      "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": "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": 0.7653333999999999,
      "accuracy_std_dev": 0.01995535461574161,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.233724332,
      "avg_duration_per_task_ms": 181237.788,
      "avg_output_tokens_per_task": 6561.384,
      "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": "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": 0.7652380714285715,
      "accuracy_std_dev": 0.03138900937194484,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.5665733142857142,
      "avg_duration_per_task_ms": 182477.46285714285,
      "avg_output_tokens_per_task": 10583.794285714286,
      "last_run_timestamp": "2026-09-19T08:14:01.322Z"
    },
    {
      "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": 0.7651851481481481,
      "accuracy_std_dev": 0.03854701154165998,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.08648212370370369,
      "avg_duration_per_task_ms": 139440.87037037036,
      "avg_output_tokens_per_task": 5742.8592592592595,
      "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": "Novita",
      "accuracy": 0.7650793809523809,
      "accuracy_std_dev": 0.034969675378037926,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.010378357142857146,
      "avg_duration_per_task_ms": 113551.27047619048,
      "avg_output_tokens_per_task": 9004.36761904762,
      "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": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7644444761904762,
      "accuracy_std_dev": 0.03577317233509074,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.12052889047619045,
      "avg_duration_per_task_ms": 163976.96857142856,
      "avg_output_tokens_per_task": 5957.86,
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Together",
      "accuracy": 0.7644444444444445,
      "accuracy_std_dev": 0.020608041101101583,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.027644917777777778,
      "avg_duration_per_task_ms": 219174.32,
      "avg_output_tokens_per_task": 5612.551111111111,
      "last_run_timestamp": "2026-06-20T10:37:11.919Z"
    },
    {
      "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": 0.7642424545454546,
      "accuracy_std_dev": 0.014290618654231216,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.016105059999999997,
      "avg_duration_per_task_ms": 208696.1781818182,
      "avg_output_tokens_per_task": 19948.01090909091,
      "last_run_timestamp": "2026-09-19T08:21:12.865Z"
    },
    {
      "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": 0.7641560048076923,
      "accuracy_std_dev": 0.054154465237873047,
      "run_count": 25,
      "total_questions": 1248,
      "avg_cost_per_task_usd": 0.1535889431089744,
      "avg_duration_per_task_ms": 302284.3413461539,
      "avg_output_tokens_per_task": 17197.23798076923,
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Baidu",
      "accuracy": 0.7641026153846153,
      "accuracy_std_dev": 0.022423789261676347,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.042344887692307695,
      "avg_duration_per_task_ms": 152484.15384615384,
      "avg_output_tokens_per_task": 9627.38153846154,
      "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": "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": "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7633335,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.09941562,
      "avg_duration_per_task_ms": 141031.62,
      "avg_output_tokens_per_task": 6151.46,
      "last_run_timestamp": "2026-09-19T08:12:15.594Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.7631372941176471,
      "accuracy_std_dev": 0.035211220731239695,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.05713043764705882,
      "avg_duration_per_task_ms": 206801.39411764705,
      "avg_output_tokens_per_task": 9987.18,
      "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": "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": 0.7628574285714285,
      "accuracy_std_dev": 0.018807973414463997,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.4976220542857142,
      "avg_duration_per_task_ms": 133220.5942857143,
      "avg_output_tokens_per_task": 8119.648571428572,
      "last_run_timestamp": "2026-09-19T08:12:01.797Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7626666,
      "accuracy_std_dev": 0.03227647784749755,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.146569104,
      "avg_duration_per_task_ms": 373419.58,
      "avg_output_tokens_per_task": 24831.152,
      "last_run_timestamp": "2026-08-29T08:26:04.467Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Io Net",
      "accuracy": 0.7626261818181818,
      "accuracy_std_dev": 0.028670609588697295,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.021812868686868686,
      "avg_duration_per_task_ms": 396305.23232323234,
      "avg_output_tokens_per_task": 9161.318181818182,
      "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": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "xAI",
      "accuracy": 0.7622223333333333,
      "accuracy_std_dev": 0.021998563367234276,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.3364002133333333,
      "avg_duration_per_task_ms": 240112.26,
      "avg_output_tokens_per_task": 11849.62,
      "last_run_timestamp": "2026-09-01T13:04:16.543Z"
    },
    {
      "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": 0.7622223333333333,
      "accuracy_std_dev": 0.016629618082071397,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.10776526666666666,
      "avg_duration_per_task_ms": 82608.36666666667,
      "avg_output_tokens_per_task": 4630.54,
      "last_run_timestamp": "2026-09-19T08:12:48.493Z"
    },
    {
      "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": 0.7622221666666666,
      "accuracy_std_dev": 0.026851245460975486,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03588941,
      "avg_duration_per_task_ms": 516654.7166666667,
      "avg_output_tokens_per_task": 13637.093333333334,
      "last_run_timestamp": "2026-09-19T08:06:14.154Z"
    },
    {
      "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": 0.762,
      "accuracy_std_dev": 0.029522112241504674,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.517524358,
      "avg_duration_per_task_ms": 222700.632,
      "avg_output_tokens_per_task": 10257.826,
      "last_run_timestamp": "2026-09-19T08:15:44.074Z"
    },
    {
      "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": "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": 0.7619047142857143,
      "accuracy_std_dev": 0.022172283571511325,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.41295487714285717,
      "avg_duration_per_task_ms": 99216.78285714285,
      "avg_output_tokens_per_task": 5645.82,
      "last_run_timestamp": "2026-09-19T08:02:29.843Z"
    },
    {
      "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": 0.76166675,
      "accuracy_std_dev": 0.002886895683515417,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.070178685,
      "avg_duration_per_task_ms": 33757.38,
      "avg_output_tokens_per_task": 4761.855,
      "last_run_timestamp": "2026-09-19T08:08:46.255Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.7616666666666667,
      "accuracy_std_dev": 0.03782268572636741,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.09442261499999999,
      "avg_duration_per_task_ms": 188047.78333333333,
      "avg_output_tokens_per_task": 3614.1383333333333,
      "last_run_timestamp": "2026-06-20T09:40:45.528Z"
    },
    {
      "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": 0.7615604913294798,
      "accuracy_std_dev": 0.02313264800127777,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.20780732369942198,
      "avg_duration_per_task_ms": 649367.0115606936,
      "avg_output_tokens_per_task": 52988.69942196532,
      "last_run_timestamp": "2026-09-19T08:07:24.262Z"
    },
    {
      "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-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Google",
      "accuracy": 0.7613333999999999,
      "accuracy_std_dev": 0.02166412395505529,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.6074076380000001,
      "avg_duration_per_task_ms": 125643.917,
      "avg_output_tokens_per_task": 8207.207,
      "last_run_timestamp": "2026-09-19T08:12:07.822Z"
    },
    {
      "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": 0.7612698571428571,
      "accuracy_std_dev": 0.019584924351059803,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.712182400952381,
      "avg_duration_per_task_ms": 75822.46,
      "avg_output_tokens_per_task": 5652.990476190476,
      "last_run_timestamp": "2026-09-19T08:15:48.243Z"
    },
    {
      "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": 0.7610441927710843,
      "accuracy_std_dev": 0.04002571797891325,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.016136098393574293,
      "avg_duration_per_task_ms": 328656.0351405622,
      "avg_output_tokens_per_task": 8275.835341365462,
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Alibaba",
      "accuracy": 0.7610256538461538,
      "accuracy_std_dev": 0.03457639874910488,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.11041888999999995,
      "avg_duration_per_task_ms": 290001.2453846154,
      "avg_output_tokens_per_task": 16905.416153846152,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "StreamLake",
      "accuracy": 0.7610153534482759,
      "accuracy_std_dev": 0.044412661393182805,
      "run_count": 14,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.06370122844827586,
      "avg_duration_per_task_ms": 278513.2284482759,
      "avg_output_tokens_per_task": 9221.025862068966,
      "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-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "WandB",
      "accuracy": 0.7609524285714285,
      "accuracy_std_dev": 0.018662793573288653,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.021895571428571427,
      "avg_duration_per_task_ms": 241457.5,
      "avg_output_tokens_per_task": 8715.474285714286,
      "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": "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Chutes",
      "accuracy": 0.7609489251824817,
      "accuracy_std_dev": 0.033505551928248105,
      "run_count": 22,
      "total_questions": 1096,
      "avg_cost_per_task_usd": 0.08300357664233575,
      "avg_duration_per_task_ms": 369852.16697080294,
      "avg_output_tokens_per_task": 15814.203467153284,
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.7601780858778625,
      "accuracy_std_dev": 0.02811995722322186,
      "run_count": 21,
      "total_questions": 1048,
      "avg_cost_per_task_usd": 0.0422277356870229,
      "avg_duration_per_task_ms": 162724.19942748093,
      "avg_output_tokens_per_task": 7589.8158396946565,
      "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": "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": 0.7600003333333333,
      "accuracy_std_dev": 0.009428090415820642,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.2124113333333333,
      "avg_duration_per_task_ms": 390658.8466666667,
      "avg_output_tokens_per_task": 13665.773333333333,
      "last_run_timestamp": "2026-09-19T08:01:34.086Z"
    },
    {
      "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": 0.7600002499999999,
      "accuracy_std_dev": 0.018257509870940786,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.14997882,
      "avg_duration_per_task_ms": 322280.56,
      "avg_output_tokens_per_task": 20351.64,
      "last_run_timestamp": "2026-09-19T08:16:41.563Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01747962,
      "avg_duration_per_task_ms": 589736.28,
      "avg_output_tokens_per_task": 13675.16,
      "last_run_timestamp": "2026-08-01T08:06:01.204Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "NextBit",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.015634861336129593,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.038081735,
      "avg_duration_per_task_ms": 203007.56,
      "avg_output_tokens_per_task": 17943.12,
      "last_run_timestamp": "2026-09-19T08:18:11.626Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "StreamLake",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.024688728874297516,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.007911357142857142,
      "avg_duration_per_task_ms": 267145.0057142857,
      "avg_output_tokens_per_task": 15598.677142857143,
      "last_run_timestamp": "2026-09-19T08:07:37.278Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0.01247214674785381,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.021322305,
      "avg_duration_per_task_ms": 109483.545,
      "avg_output_tokens_per_task": 16713.61,
      "last_run_timestamp": "2026-09-12T08:11:31.906Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.01333300000000004,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.05555669,
      "avg_duration_per_task_ms": 95287.39,
      "avg_output_tokens_per_task": 15474.24,
      "last_run_timestamp": "2026-09-19T08:01:53.394Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0.016996895142937136,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.10744996000000001,
      "avg_duration_per_task_ms": 335547.935,
      "avg_output_tokens_per_task": 20056.545,
      "last_run_timestamp": "2026-09-19T08:08:42.555Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.15061989,
      "avg_duration_per_task_ms": 306259.77,
      "avg_output_tokens_per_task": 19605.64,
      "last_run_timestamp": "2026-09-19T08:01:34.094Z"
    },
    {
      "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": "DeepInfra",
      "accuracy": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03723274,
      "avg_duration_per_task_ms": 352117.32,
      "avg_output_tokens_per_task": 12561.96,
      "last_run_timestamp": "2026-09-19T08:12:48.363Z"
    },
    {
      "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": 0.022705750619611758,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.077266932,
      "avg_duration_per_task_ms": 260289.312,
      "avg_output_tokens_per_task": 11860.184,
      "last_run_timestamp": "2026-09-19T08:01:53.315Z"
    },
    {
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Inceptron",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.03265986323710907,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05974882666666666,
      "avg_duration_per_task_ms": 170255.57333333333,
      "avg_output_tokens_per_task": 4552.706666666667,
      "last_run_timestamp": "2026-09-19T08:16:20.619Z"
    },
    {
      "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": 0.76,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04466577,
      "avg_duration_per_task_ms": 116479.54,
      "avg_output_tokens_per_task": 4371.72,
      "last_run_timestamp": "2026-09-19T08:15:43.993Z"
    },
    {
      "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": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05904302,
      "avg_duration_per_task_ms": 70821.54,
      "avg_output_tokens_per_task": 4624.02,
      "last_run_timestamp": "2026-09-19T08:01:53.310Z"
    },
    {
      "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-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Parasail",
      "accuracy": 0.7593333999999999,
      "accuracy_std_dev": 0.014360396158880871,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.1251978,
      "avg_duration_per_task_ms": 443535.78,
      "avg_output_tokens_per_task": 22633.804,
      "last_run_timestamp": "2026-09-19T08:17:30.607Z"
    },
    {
      "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": 0.7590475714285715,
      "accuracy_std_dev": 0.014443569136046514,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.008607337142857142,
      "avg_duration_per_task_ms": 374540.4914285714,
      "avg_output_tokens_per_task": 13803.397142857142,
      "last_run_timestamp": "2026-09-19T08:14:25.256Z"
    },
    {
      "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": 0.759040841981132,
      "accuracy_std_dev": 0.036093876622690174,
      "run_count": 17,
      "total_questions": 848,
      "avg_cost_per_task_usd": 0.08171993514150944,
      "avg_duration_per_task_ms": 269188.51061320753,
      "avg_output_tokens_per_task": 4527.392688679245,
      "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Novita",
      "accuracy": 0.7588095,
      "accuracy_std_dev": 0.034551746496585184,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.052020888571428586,
      "avg_duration_per_task_ms": 138271.82571428572,
      "avg_output_tokens_per_task": 5773.107857142857,
      "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": "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Venice",
      "accuracy": 0.7585878072519083,
      "accuracy_std_dev": 0.03764932431268607,
      "run_count": 21,
      "total_questions": 1048,
      "avg_cost_per_task_usd": 0.0847812709923664,
      "avg_duration_per_task_ms": 181956.90935114503,
      "avg_output_tokens_per_task": 7087.411259541985,
      "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-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": "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": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.758,
      "accuracy_std_dev": 0.03148898925815182,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.19792223,
      "avg_duration_per_task_ms": 121774.793,
      "avg_output_tokens_per_task": 7801.686,
      "last_run_timestamp": "2026-09-19T08:15:55.793Z"
    },
    {
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7578826216216217,
      "accuracy_std_dev": 0.006612862617656559,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.020859918918918918,
      "avg_duration_per_task_ms": 101574.95945945945,
      "avg_output_tokens_per_task": 13789.418918918918,
      "last_run_timestamp": "2026-09-19T08:10:29.894Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Relace",
      "accuracy": 0.757778,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.006585946666666667,
      "avg_duration_per_task_ms": 132381.80666666667,
      "avg_output_tokens_per_task": 4303.826666666667,
      "last_run_timestamp": "2026-09-19T08:02:29.956Z"
    },
    {
      "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-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7577776666666667,
      "accuracy_std_dev": 0.017497908167041637,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.006062186666666667,
      "avg_duration_per_task_ms": 148585.37333333332,
      "avg_output_tokens_per_task": 4189.8,
      "last_run_timestamp": "2026-09-19T08:06:51.600Z"
    },
    {
      "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": 0.7574602857142857,
      "accuracy_std_dev": 0.030548852532643923,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.010414794285714285,
      "avg_duration_per_task_ms": 137613.95523809525,
      "avg_output_tokens_per_task": 9270.931428571428,
      "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": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7571428571428571,
      "accuracy_std_dev": 0.029783673564212878,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.19815623142857142,
      "avg_duration_per_task_ms": 133956.76285714286,
      "avg_output_tokens_per_task": 7857.111428571428,
      "last_run_timestamp": "2026-09-19T08:06:50.238Z"
    },
    {
      "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-v4.1-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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Reka",
      "accuracy": 0.7566667499999999,
      "accuracy_std_dev": 0.013743584364622665,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.085799,
      "avg_duration_per_task_ms": 225293.88,
      "avg_output_tokens_per_task": 12078.17,
      "last_run_timestamp": "2026-09-19T08:13:58.337Z"
    },
    {
      "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": 0.7566667499999999,
      "accuracy_std_dev": 0.013743948165920146,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07678554,
      "avg_duration_per_task_ms": 131792.745,
      "avg_output_tokens_per_task": 4961.875,
      "last_run_timestamp": "2026-09-19T08:14:27.258Z"
    },
    {
      "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": 0.7566667,
      "accuracy_std_dev": 0.024630617804066543,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.02296196,
      "avg_duration_per_task_ms": 168920.314,
      "avg_output_tokens_per_task": 18473.208,
      "last_run_timestamp": "2026-09-19T08:15:24.878Z"
    },
    {
      "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": 0.7566666666666667,
      "accuracy_std_dev": 0.009999888893827235,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.035569739999999996,
      "avg_duration_per_task_ms": 285830.0233333333,
      "avg_output_tokens_per_task": 13698.04,
      "last_run_timestamp": "2026-09-19T08:01:53.309Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "GMICloud",
      "accuracy": 0.7564705294117647,
      "accuracy_std_dev": 0.03521125545336429,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.008393454117647059,
      "avg_duration_per_task_ms": 146412.67529411765,
      "avg_output_tokens_per_task": 9095.61294117647,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Parasail",
      "accuracy": 0.7564102600732601,
      "accuracy_std_dev": 0.0234723881283154,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.02063667582417582,
      "avg_duration_per_task_ms": 550307.2802197802,
      "avg_output_tokens_per_task": 17523.981684981685,
      "last_run_timestamp": "2026-09-19T08:08:46.172Z"
    },
    {
      "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": 0.7561938175675675,
      "accuracy_std_dev": 0.02331332973798942,
      "run_count": 6,
      "total_questions": 296,
      "avg_cost_per_task_usd": 0.20687053378378378,
      "avg_duration_per_task_ms": 486380.6959459459,
      "avg_output_tokens_per_task": 20517.347972972973,
      "last_run_timestamp": "2026-09-19T08:13:57.893Z"
    },
    {
      "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": 0.7561904285714286,
      "accuracy_std_dev": 0.02465178812324715,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.3425374,
      "avg_duration_per_task_ms": 222364.59142857144,
      "avg_output_tokens_per_task": 11866.32,
      "last_run_timestamp": "2026-09-19T08:08:02.482Z"
    },
    {
      "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": 0.7559998,
      "accuracy_std_dev": 0.029393940419072766,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.550821472,
      "avg_duration_per_task_ms": 179743.46,
      "avg_output_tokens_per_task": 10831.904,
      "last_run_timestamp": "2026-09-19T08:02:29.757Z"
    },
    {
      "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": 0.7559259444444444,
      "accuracy_std_dev": 0.03443646283531493,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.05180537888888888,
      "avg_duration_per_task_ms": 157203.02777777778,
      "avg_output_tokens_per_task": 6346.41,
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "GMICloud",
      "accuracy": 0.7556390701754386,
      "accuracy_std_dev": 0.0748655823238389,
      "run_count": 16,
      "total_questions": 798,
      "avg_cost_per_task_usd": 0.1314042468671679,
      "avg_duration_per_task_ms": 282979.45112781954,
      "avg_output_tokens_per_task": 13309.800751879699,
      "last_run_timestamp": "2026-08-29T16:58:30.455Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7555927852348994,
      "accuracy_std_dev": 0.017391785347782254,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.04238428187919464,
      "avg_duration_per_task_ms": 512275.43624161073,
      "avg_output_tokens_per_task": 21918.96979865772,
      "last_run_timestamp": "2026-09-19T16:00:24.104Z"
    },
    {
      "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": 0.7554257228714524,
      "accuracy_std_dev": 0.04133335435850861,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.05269022621035059,
      "avg_duration_per_task_ms": 158796.83138564273,
      "avg_output_tokens_per_task": 6105.924040066778,
      "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": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Azure",
      "accuracy": 0.7553846923076923,
      "accuracy_std_dev": 0.03916955640129917,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.19397882,
      "avg_duration_per_task_ms": 104918.87538461538,
      "avg_output_tokens_per_task": 7649.404615384616,
      "last_run_timestamp": "2026-09-19T08:12:02.264Z"
    },
    {
      "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": 0.7552381428571429,
      "accuracy_std_dev": 0.025625984192548305,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.5360551200000001,
      "avg_duration_per_task_ms": 191556.04571428572,
      "avg_output_tokens_per_task": 10244.057142857142,
      "last_run_timestamp": "2026-09-19T08:01:55.555Z"
    },
    {
      "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": 0.7552380000000001,
      "accuracy_std_dev": 0.043456084643826665,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.19647111142857143,
      "avg_duration_per_task_ms": 107624.06285714285,
      "avg_output_tokens_per_task": 7682.417142857143,
      "last_run_timestamp": "2026-09-19T08:14:23.378Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.755,
      "accuracy_std_dev": 0.021794494717703387,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.049016185,
      "avg_duration_per_task_ms": 156512.84,
      "avg_output_tokens_per_task": 3394.74,
      "last_run_timestamp": "2026-04-25T11:20:08.052Z"
    },
    {
      "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": 0.7547709770992367,
      "accuracy_std_dev": 0.037708741308906626,
      "run_count": 21,
      "total_questions": 1048,
      "avg_cost_per_task_usd": 0.06855190171755726,
      "avg_duration_per_task_ms": 232773.2967557252,
      "avg_output_tokens_per_task": 7463.7099236641225,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Baidu",
      "accuracy": 0.7543859845201238,
      "accuracy_std_dev": 0.03746648424278414,
      "run_count": 13,
      "total_questions": 646,
      "avg_cost_per_task_usd": 0.06285631269349846,
      "avg_duration_per_task_ms": 163970.43962848297,
      "avg_output_tokens_per_task": 7998.698142414861,
      "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": "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": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Google",
      "accuracy": 0.7542857142857143,
      "accuracy_std_dev": 0.02676848927758313,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.5495588485714286,
      "avg_duration_per_task_ms": 192291.6257142857,
      "avg_output_tokens_per_task": 10585.511428571428,
      "last_run_timestamp": "2026-09-19T08:12:49.049Z"
    },
    {
      "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": 0.7542857142857143,
      "accuracy_std_dev": 0.023752365678957953,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.023086720000000005,
      "avg_duration_per_task_ms": 119604,
      "avg_output_tokens_per_task": 15643.062857142857,
      "last_run_timestamp": "2026-09-20T10:30:07.540Z"
    },
    {
      "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": 0.7542856785714285,
      "accuracy_std_dev": 0.04289943601032333,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.050694632142857146,
      "avg_duration_per_task_ms": 138996.95214285713,
      "avg_output_tokens_per_task": 5646.897857142857,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Baidu",
      "accuracy": 0.7538889166666667,
      "accuracy_std_dev": 0.03686387162127335,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.06654354,
      "avg_duration_per_task_ms": 244757.80833333332,
      "avg_output_tokens_per_task": 10635.376666666667,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7535714285714286,
      "accuracy_std_dev": 0.04815591057583436,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.04571141285714286,
      "avg_duration_per_task_ms": 126701.03428571428,
      "avg_output_tokens_per_task": 5683.6,
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Novita",
      "accuracy": 0.7533335,
      "accuracy_std_dev": 0.026666500000000037,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.0840499,
      "avg_duration_per_task_ms": 277299.54,
      "avg_output_tokens_per_task": 19530.63,
      "last_run_timestamp": "2026-09-19T08:08:43.307Z"
    },
    {
      "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": "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": 0.7533335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01056969,
      "avg_duration_per_task_ms": 192129.64,
      "avg_output_tokens_per_task": 4500.45,
      "last_run_timestamp": "2026-09-19T08:15:09.696Z"
    },
    {
      "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": 0.7533333999999999,
      "accuracy_std_dev": 0.028284365530094557,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.194302048,
      "avg_duration_per_task_ms": 129735.288,
      "avg_output_tokens_per_task": 7732.768,
      "last_run_timestamp": "2026-09-19T08:02:29.788Z"
    },
    {
      "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": 0.7533333333333333,
      "accuracy_std_dev": 0.02372709031653247,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.027049573333333334,
      "avg_duration_per_task_ms": 129531.18666666666,
      "avg_output_tokens_per_task": 16746.393333333333,
      "last_run_timestamp": "2026-09-19T08:12:02.467Z"
    },
    {
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Together",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.044721359549995794,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.08804555777777777,
      "avg_duration_per_task_ms": 73317.22,
      "avg_output_tokens_per_task": 4526.791111111111,
      "last_run_timestamp": "2026-06-20T18:29:02.205Z"
    },
    {
      "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": 0.7533333333333333,
      "accuracy_std_dev": 0.019626210745383905,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.15976216666666668,
      "avg_duration_per_task_ms": 259508.71333333335,
      "avg_output_tokens_per_task": 8887.92,
      "last_run_timestamp": "2026-09-19T08:11:59.701Z"
    },
    {
      "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": 0.7533333333333333,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08241664,
      "avg_duration_per_task_ms": 157302.79333333333,
      "avg_output_tokens_per_task": 4631.613333333334,
      "last_run_timestamp": "2026-09-19T08:08:45.555Z"
    },
    {
      "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": 0.7533333214285716,
      "accuracy_std_dev": 0.043861247675444484,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.036317195,
      "avg_duration_per_task_ms": 122335.48285714285,
      "avg_output_tokens_per_task": 5681.47,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cloudflare",
      "accuracy": 0.7533333,
      "accuracy_std_dev": 0.020870052975735364,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.02329514,
      "avg_duration_per_task_ms": 307869.234,
      "avg_output_tokens_per_task": 17664.796,
      "last_run_timestamp": "2026-09-19T08:06:51.688Z"
    },
    {
      "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": 0.7533332857142857,
      "accuracy_std_dev": 0.03804763096829185,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.22995042095238097,
      "avg_duration_per_task_ms": 102003.29238095238,
      "avg_output_tokens_per_task": 7520.403809523809,
      "last_run_timestamp": "2026-09-19T08:15:48.049Z"
    },
    {
      "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": 0.7533332500000001,
      "accuracy_std_dev": 0.016996797086142445,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.036730425000000004,
      "avg_duration_per_task_ms": 213947.705,
      "avg_output_tokens_per_task": 18813.24,
      "last_run_timestamp": "2026-09-19T08:13:20.914Z"
    },
    {
      "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": 0.7533332,
      "accuracy_std_dev": 0.03675737870088128,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.42735367599999996,
      "avg_duration_per_task_ms": 94631.4,
      "avg_output_tokens_per_task": 5528.516,
      "last_run_timestamp": "2026-09-19T08:15:14.773Z"
    },
    {
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "GMICloud",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02738974,
      "avg_duration_per_task_ms": 82133.22,
      "avg_output_tokens_per_task": 9075.3,
      "last_run_timestamp": "2026-09-19T16:00:30.815Z"
    },
    {
      "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": "Relace",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.2831028,
      "avg_duration_per_task_ms": 544501.1,
      "avg_output_tokens_per_task": 8096.24,
      "last_run_timestamp": "2026-08-23T18:00:01.881Z"
    },
    {
      "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Friendli",
      "accuracy": 0.7529824736842106,
      "accuracy_std_dev": 0.1257765564827976,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.04461744315789474,
      "avg_duration_per_task_ms": 41775.2,
      "avg_output_tokens_per_task": 3596.8168421052633,
      "last_run_timestamp": "2026-07-01T02:07:35.472Z"
    },
    {
      "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": 0.7527272727272727,
      "accuracy_std_dev": 0.016441999795926792,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.05201301999999999,
      "avg_duration_per_task_ms": 275857.43454545457,
      "avg_output_tokens_per_task": 30292.38909090909,
      "last_run_timestamp": "2026-09-19T16:00:28.636Z"
    },
    {
      "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": 0.7521489914040115,
      "accuracy_std_dev": 0.01514003431640894,
      "run_count": 7,
      "total_questions": 349,
      "avg_cost_per_task_usd": 0.013516624641833811,
      "avg_duration_per_task_ms": 212155.76217765044,
      "avg_output_tokens_per_task": 16555.63323782235,
      "last_run_timestamp": "2026-09-19T08:18:14.972Z"
    },
    {
      "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": 0.7521211818181818,
      "accuracy_std_dev": 0.03491049713147814,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.18250516363636365,
      "avg_duration_per_task_ms": 550090.9563636364,
      "avg_output_tokens_per_task": 26501.607272727273,
      "last_run_timestamp": "2026-09-19T08:14:01.746Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.752,
      "accuracy_std_dev": 0.02400000000000002,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.031335448,
      "avg_duration_per_task_ms": 314672.816,
      "avg_output_tokens_per_task": 20249.848,
      "last_run_timestamp": "2026-06-13T19:57:55.549Z"
    },
    {
      "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": 0.752,
      "accuracy_std_dev": 0.025785560129653974,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.12171823600000001,
      "avg_duration_per_task_ms": 271294.616,
      "avg_output_tokens_per_task": 12082.784,
      "last_run_timestamp": "2026-09-19T08:02:39.236Z"
    },
    {
      "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": 0.7519999333333334,
      "accuracy_std_dev": 0.042457799179054144,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.07472457066666666,
      "avg_duration_per_task_ms": 97051.48533333333,
      "avg_output_tokens_per_task": 11508.92,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7517460476190476,
      "accuracy_std_dev": 0.03886781145721899,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.0710823885714286,
      "avg_duration_per_task_ms": 177196.37523809524,
      "avg_output_tokens_per_task": 5993.54,
      "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": "Io Net",
      "accuracy": 0.7517008979591836,
      "accuracy_std_dev": 0.00847199999999998,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.008818877551020407,
      "avg_duration_per_task_ms": 183019.20408163266,
      "avg_output_tokens_per_task": 3602.8367346938776,
      "last_run_timestamp": "2026-09-19T08:15:22.322Z"
    },
    {
      "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": 0.7516667499999999,
      "accuracy_std_dev": 0.031046636068783707,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.08865978499999999,
      "avg_duration_per_task_ms": 293099.965,
      "avg_output_tokens_per_task": 12000.945,
      "last_run_timestamp": "2026-09-19T08:02:29.880Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.7513043478260869,
      "accuracy_std_dev": 0.033096829546845286,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.008512693043478257,
      "avg_duration_per_task_ms": 141943.57217391304,
      "avg_output_tokens_per_task": 8289.525217391305,
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "StreamLake",
      "accuracy": 0.7510040481927711,
      "accuracy_std_dev": 0.022883894021997232,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.007396576305220883,
      "avg_duration_per_task_ms": 255494.20682730924,
      "avg_output_tokens_per_task": 11511.991967871487,
      "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": "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": 0.75100005,
      "accuracy_std_dev": 0.022634698291505915,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.370373307,
      "avg_duration_per_task_ms": 170883.772,
      "avg_output_tokens_per_task": 11914.488,
      "last_run_timestamp": "2026-09-19T08:15:51.168Z"
    },
    {
      "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": 0.7508740961538461,
      "accuracy_std_dev": 0.08177386372104274,
      "run_count": 23,
      "total_questions": 1144,
      "avg_cost_per_task_usd": 0.05178484615384617,
      "avg_duration_per_task_ms": 150234.52097902098,
      "avg_output_tokens_per_task": 20045.546328671328,
      "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": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.7506666,
      "accuracy_std_dev": 0.029242276854581632,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.336778826,
      "avg_duration_per_task_ms": 211305.436,
      "avg_output_tokens_per_task": 11826.724,
      "last_run_timestamp": "2026-09-19T08:14:36.128Z"
    },
    {
      "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": 0.7505555833333333,
      "accuracy_std_dev": 0.037757344555857254,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.016059435,
      "avg_duration_per_task_ms": 293963.405,
      "avg_output_tokens_per_task": 5996.211666666667,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Fireworks",
      "accuracy": 0.7505555,
      "accuracy_std_dev": 0.03274012621004996,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.02124632666666667,
      "avg_duration_per_task_ms": 354797.86666666664,
      "avg_output_tokens_per_task": 18494.39,
      "last_run_timestamp": "2026-09-19T08:13:24.062Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Venice",
      "accuracy": 0.7504762857142857,
      "accuracy_std_dev": 0.034523051552401846,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.022417262857142857,
      "avg_duration_per_task_ms": 166619.53714285715,
      "avg_output_tokens_per_task": 14078.568571428572,
      "last_run_timestamp": "2026-09-19T08:12:29.747Z"
    },
    {
      "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": 0.7504761428571428,
      "accuracy_std_dev": 0.029354239613134367,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.013216268571428573,
      "avg_duration_per_task_ms": 156715.7,
      "avg_output_tokens_per_task": 16682.777142857143,
      "last_run_timestamp": "2026-09-19T08:07:36.192Z"
    },
    {
      "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": 0.7504761428571428,
      "accuracy_std_dev": 0.023328467908964635,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.02283790857142857,
      "avg_duration_per_task_ms": 124586.03428571428,
      "avg_output_tokens_per_task": 7423.028571428571,
      "last_run_timestamp": "2026-09-19T16:00:32.489Z"
    },
    {
      "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": "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": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "Nebius",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.030000000000000027,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.11403106000000002,
      "avg_duration_per_task_ms": 126507.38,
      "avg_output_tokens_per_task": 4905.84,
      "last_run_timestamp": "2026-06-20T10:57:18.552Z"
    },
    {
      "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": 0.7497436153846154,
      "accuracy_std_dev": 0.0478327757706888,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.08224532461538461,
      "avg_duration_per_task_ms": 182548.32615384617,
      "avg_output_tokens_per_task": 7228.066153846154,
      "last_run_timestamp": "2026-09-19T08:13:49.205Z"
    },
    {
      "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": 0.7495238571428572,
      "accuracy_std_dev": 0.028918598713475383,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.48306242571428565,
      "avg_duration_per_task_ms": 195308.45714285714,
      "avg_output_tokens_per_task": 9967.34857142857,
      "last_run_timestamp": "2026-09-19T08:17:29.725Z"
    },
    {
      "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": 0.7495237857142857,
      "accuracy_std_dev": 0.031138569678084756,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.5694101414285715,
      "avg_duration_per_task_ms": 203607.95571428572,
      "avg_output_tokens_per_task": 9913.828571428572,
      "last_run_timestamp": "2026-09-19T08:14:26.104Z"
    },
    {
      "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": 0.74916675,
      "accuracy_std_dev": 0.018837731788554076,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.021088265000000002,
      "avg_duration_per_task_ms": 122926.245,
      "avg_output_tokens_per_task": 16099.0975,
      "last_run_timestamp": "2026-09-19T08:12:00.893Z"
    },
    {
      "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": 0.7491665000000001,
      "accuracy_std_dev": 0.04238543054222759,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.019819255,
      "avg_duration_per_task_ms": 82254.675,
      "avg_output_tokens_per_task": 4110.175,
      "last_run_timestamp": "2026-09-19T08:15:43.803Z"
    },
    {
      "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": 0.7490876715328467,
      "accuracy_std_dev": 0.019546598328901252,
      "run_count": 11,
      "total_questions": 548,
      "avg_cost_per_task_usd": 0.0068139452554744516,
      "avg_duration_per_task_ms": 184530.0200729927,
      "avg_output_tokens_per_task": 12199.35583941606,
      "last_run_timestamp": "2026-09-19T08:10:30.975Z"
    },
    {
      "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": 0.749,
      "accuracy_std_dev": 0.027830806272905576,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.033878175999999996,
      "avg_duration_per_task_ms": 238766.5,
      "avg_output_tokens_per_task": 5451.338,
      "last_run_timestamp": "2026-08-29T16:10:33.029Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "DeepInfra",
      "accuracy": 0.748889,
      "accuracy_std_dev": 0.03281068781154497,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03277637333333334,
      "avg_duration_per_task_ms": 282626.5333333333,
      "avg_output_tokens_per_task": 18948.02,
      "last_run_timestamp": "2026-09-19T08:01:36.428Z"
    },
    {
      "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": 0.7488585479452055,
      "accuracy_std_dev": 0.02333557629219577,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.13015020547945205,
      "avg_duration_per_task_ms": 94301.3698630137,
      "avg_output_tokens_per_task": 9432.931506849314,
      "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-5.1-20260406",
      "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": "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": 0.7483332500000001,
      "accuracy_std_dev": 0.01723988431479461,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.010718860000000002,
      "avg_duration_per_task_ms": 133490.525,
      "avg_output_tokens_per_task": 4191.97,
      "last_run_timestamp": "2026-09-19T08:01:53.307Z"
    },
    {
      "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": 0.7483330000000001,
      "accuracy_std_dev": 0.02598076211353318,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.024764005,
      "avg_duration_per_task_ms": 370315.33,
      "avg_output_tokens_per_task": 4324.135,
      "last_run_timestamp": "2026-09-19T08:15:51.681Z"
    },
    {
      "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": 0.7480001,
      "accuracy_std_dev": 0.02612788548830541,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.35369032,
      "avg_duration_per_task_ms": 233260.038,
      "avg_output_tokens_per_task": 10816.624,
      "last_run_timestamp": "2026-09-19T08:19:00.240Z"
    },
    {
      "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": 0.7477778333333334,
      "accuracy_std_dev": 0.02759945204417816,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.10948868333333332,
      "avg_duration_per_task_ms": 328943.99666666664,
      "avg_output_tokens_per_task": 18817.083333333332,
      "last_run_timestamp": "2026-09-19T08:02:29.971Z"
    },
    {
      "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": 0.7476191428571428,
      "accuracy_std_dev": 0.035710930637265534,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.5503071828571428,
      "avg_duration_per_task_ms": 184401.16,
      "avg_output_tokens_per_task": 10642.84857142857,
      "last_run_timestamp": "2026-09-19T08:12:56.382Z"
    },
    {
      "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": 0.7476191428571428,
      "accuracy_std_dev": 0.02090910249101076,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.3218781142857143,
      "avg_duration_per_task_ms": 199636.20857142858,
      "avg_output_tokens_per_task": 11432.102857142858,
      "last_run_timestamp": "2026-09-19T08:12:55.142Z"
    },
    {
      "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": 0.7474746666666666,
      "accuracy_std_dev": 0.013351080637442798,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.14427683333333333,
      "avg_duration_per_task_ms": 274328.10101010103,
      "avg_output_tokens_per_task": 7734.318181818182,
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Morph",
      "accuracy": 0.7473334,
      "accuracy_std_dev": 0.02318052631067726,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.019461324000000002,
      "avg_duration_per_task_ms": 260422.68,
      "avg_output_tokens_per_task": 9034.204,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7472726363636364,
      "accuracy_std_dev": 0.0339286130072398,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.09873066363636362,
      "avg_duration_per_task_ms": 157555.76,
      "avg_output_tokens_per_task": 8576.863636363636,
      "last_run_timestamp": "2026-09-19T08:15:47.690Z"
    },
    {
      "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": "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DekaLLM",
      "accuracy": 0.746667,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.05560944,
      "avg_duration_per_task_ms": 195101.15,
      "avg_output_tokens_per_task": 14684.06,
      "last_run_timestamp": "2026-09-19T08:08:01.712Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Phala",
      "accuracy": 0.7466666666666667,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0582748,
      "avg_duration_per_task_ms": 126998.29333333333,
      "avg_output_tokens_per_task": 4129.593333333333,
      "last_run_timestamp": "2026-09-19T08:17:43.474Z"
    },
    {
      "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": 0.7466666666666667,
      "accuracy_std_dev": 0.028803034820811656,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.009823306666666667,
      "avg_duration_per_task_ms": 205830.71333333335,
      "avg_output_tokens_per_task": 4522.053333333333,
      "last_run_timestamp": "2026-09-19T08:08:18.540Z"
    },
    {
      "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": 0.7466665,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.31129542,
      "avg_duration_per_task_ms": 181218.77,
      "avg_output_tokens_per_task": 16905.63,
      "last_run_timestamp": "2026-09-19T08:02:29.954Z"
    },
    {
      "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": "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": 0.7466665,
      "accuracy_std_dev": 0.011546716708657956,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.1707169,
      "avg_duration_per_task_ms": 302727.215,
      "avg_output_tokens_per_task": 10539.265,
      "last_run_timestamp": "2026-09-19T08:02:39.146Z"
    },
    {
      "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": "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": 0.7466665,
      "accuracy_std_dev": 0.03126950946609175,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.007685050000000001,
      "avg_duration_per_task_ms": 185392.305,
      "avg_output_tokens_per_task": 4351.84,
      "last_run_timestamp": "2026-09-19T08:09:43.611Z"
    },
    {
      "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": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03271783,
      "avg_duration_per_task_ms": 346196.84,
      "avg_output_tokens_per_task": 4294.49,
      "last_run_timestamp": "2026-08-29T08:54:52.477Z"
    },
    {
      "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": 0.7461684252873563,
      "accuracy_std_dev": 0.022865583595063955,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.14753891091954022,
      "avg_duration_per_task_ms": 476020.4597701149,
      "avg_output_tokens_per_task": 10344.031609195403,
      "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": "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": 0.745819397993311,
      "accuracy_std_dev": 0.027782945492193997,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.03796495317725753,
      "avg_duration_per_task_ms": 162257.28929765886,
      "avg_output_tokens_per_task": 7729.632107023412,
      "last_run_timestamp": "2026-09-19T08:15:37.736Z"
    },
    {
      "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": 0.7458193712374581,
      "accuracy_std_dev": 0.020725900684590307,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.03691470903010033,
      "avg_duration_per_task_ms": 270360.6237458194,
      "avg_output_tokens_per_task": 23763.14381270903,
      "last_run_timestamp": "2026-09-19T16:00:32.493Z"
    },
    {
      "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": 0.7456140464396285,
      "accuracy_std_dev": 0.047057992260714526,
      "run_count": 13,
      "total_questions": 646,
      "avg_cost_per_task_usd": 0.011993541795665636,
      "avg_duration_per_task_ms": 632093.0835913313,
      "avg_output_tokens_per_task": 8861.049535603715,
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.7454545454545455,
      "accuracy_std_dev": 0.045199256436691,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.10578343272727275,
      "avg_duration_per_task_ms": 664793.96,
      "avg_output_tokens_per_task": 24353.294545454544,
      "last_run_timestamp": "2026-09-19T16:00:22.544Z"
    },
    {
      "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": 0.7453334,
      "accuracy_std_dev": 0.028094905827213558,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.348214824,
      "avg_duration_per_task_ms": 202851.62,
      "avg_output_tokens_per_task": 11895.116,
      "last_run_timestamp": "2026-09-19T08:06:16.236Z"
    },
    {
      "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": 0.7452631578947368,
      "accuracy_std_dev": 0.03464626575042041,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.37969141473684204,
      "avg_duration_per_task_ms": 207895.33052631578,
      "avg_output_tokens_per_task": 11954.22,
      "last_run_timestamp": "2026-09-19T08:16:29.435Z"
    },
    {
      "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": 0.7451159230769231,
      "accuracy_std_dev": 0.042379107322140314,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.11416542307692307,
      "avg_duration_per_task_ms": 465658.6978021978,
      "avg_output_tokens_per_task": 24052.309523809523,
      "last_run_timestamp": "2026-09-19T08:13:25.177Z"
    },
    {
      "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": 0.7450981176470588,
      "accuracy_std_dev": 0.03403410890628669,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.08884740352941176,
      "avg_duration_per_task_ms": 247677.12588235294,
      "avg_output_tokens_per_task": 7114.02,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Io Net",
      "accuracy": 0.745,
      "accuracy_std_dev": 0.051071844820148544,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.05017629,
      "avg_duration_per_task_ms": 100886.12833333333,
      "avg_output_tokens_per_task": 4164.488333333334,
      "last_run_timestamp": "2026-06-20T12:09:06.906Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "NextBit",
      "accuracy": 0.7444446666666666,
      "accuracy_std_dev": 0.016629618082071397,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12615662666666666,
      "avg_duration_per_task_ms": 491171.2133333333,
      "avg_output_tokens_per_task": 19592.646666666667,
      "last_run_timestamp": "2026-09-19T08:16:40.730Z"
    },
    {
      "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": 0.7444446666666666,
      "accuracy_std_dev": 0.003142853940113798,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.022935546666666664,
      "avg_duration_per_task_ms": 247026.00666666665,
      "avg_output_tokens_per_task": 4198.3133333333335,
      "last_run_timestamp": "2026-09-19T08:15:43.987Z"
    },
    {
      "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": 0.7444443333333334,
      "accuracy_std_dev": 0.03823267621928775,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08758441666666665,
      "avg_duration_per_task_ms": 265114.18333333335,
      "avg_output_tokens_per_task": 12207.423333333334,
      "last_run_timestamp": "2026-09-19T08:12:01.887Z"
    },
    {
      "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": 0.7442919863013698,
      "accuracy_std_dev": 0.006500008529908962,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.012089356164383563,
      "avg_duration_per_task_ms": 63776.74657534246,
      "avg_output_tokens_per_task": 3609.541095890411,
      "last_run_timestamp": "2026-09-19T08:07:54.122Z"
    },
    {
      "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": 0.744221408759124,
      "accuracy_std_dev": 0.0694836949766747,
      "run_count": 22,
      "total_questions": 1096,
      "avg_cost_per_task_usd": 0.010320726277372263,
      "avg_duration_per_task_ms": 190073.67335766423,
      "avg_output_tokens_per_task": 8391.975364963504,
      "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-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Novita",
      "accuracy": 0.7441045812619503,
      "accuracy_std_dev": 0.03256003617627679,
      "run_count": 21,
      "total_questions": 1046,
      "avg_cost_per_task_usd": 0.061009545889101346,
      "avg_duration_per_task_ms": 226154.17782026768,
      "avg_output_tokens_per_task": 8148.751434034417,
      "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": "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": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Azure",
      "accuracy": 0.7438095714285714,
      "accuracy_std_dev": 0.03184425677241906,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.51650314,
      "avg_duration_per_task_ms": 230071.5342857143,
      "avg_output_tokens_per_task": 10020.925714285715,
      "last_run_timestamp": "2026-09-19T08:10:31.322Z"
    },
    {
      "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": 0.743589829431438,
      "accuracy_std_dev": 0.021188684958014764,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.028942936454849497,
      "avg_duration_per_task_ms": 685874.1371237459,
      "avg_output_tokens_per_task": 21249.709030100334,
      "last_run_timestamp": "2026-09-19T16:00:32.531Z"
    },
    {
      "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": 0.7433335,
      "accuracy_std_dev": 0.012018411802314,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.14946610500000002,
      "avg_duration_per_task_ms": 148382.875,
      "avg_output_tokens_per_task": 20254.93,
      "last_run_timestamp": "2026-09-19T08:01:53.309Z"
    },
    {
      "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-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "InferenceNet",
      "accuracy": 0.7433335,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.4675528,
      "avg_duration_per_task_ms": 415045.58,
      "avg_output_tokens_per_task": 10466.34,
      "last_run_timestamp": "2026-09-19T08:02:29.875Z"
    },
    {
      "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": 0.7433332500000001,
      "accuracy_std_dev": 0.028867484592314266,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.00987351,
      "avg_duration_per_task_ms": 304668.76,
      "avg_output_tokens_per_task": 4196.865,
      "last_run_timestamp": "2026-09-19T08:16:20.350Z"
    },
    {
      "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": 0.7433330000000001,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01160009,
      "avg_duration_per_task_ms": 224850.4,
      "avg_output_tokens_per_task": 4363.45,
      "last_run_timestamp": "2026-09-19T08:12:02.555Z"
    },
    {
      "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": 0.7432844835164835,
      "accuracy_std_dev": 0.04305427570650295,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.16399129853479852,
      "avg_duration_per_task_ms": 413747.94505494507,
      "avg_output_tokens_per_task": 24261.919413919415,
      "last_run_timestamp": "2026-09-12T16:00:20.988Z"
    },
    {
      "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": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "auto-routing",
      "accuracy": 0.7426666000000001,
      "accuracy_std_dev": 0.03641722913457311,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.038778624,
      "avg_duration_per_task_ms": 203795.688,
      "avg_output_tokens_per_task": 13449.364,
      "last_run_timestamp": "2026-09-19T08:01:36.459Z"
    },
    {
      "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": 0.7422839058641976,
      "accuracy_std_dev": 0.037889381261823826,
      "run_count": 26,
      "total_questions": 1296,
      "avg_cost_per_task_usd": 0.05804707330246913,
      "avg_duration_per_task_ms": 156059.7939814815,
      "avg_output_tokens_per_task": 6063.099537037037,
      "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": "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": 0.7422223333333333,
      "accuracy_std_dev": 0.003142853940113798,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.029752413333333328,
      "avg_duration_per_task_ms": 184850.22,
      "avg_output_tokens_per_task": 18168.666666666668,
      "last_run_timestamp": "2026-09-19T08:01:34.087Z"
    },
    {
      "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": "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": 0.742222,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.17656242666666666,
      "avg_duration_per_task_ms": 161433.59333333332,
      "avg_output_tokens_per_task": 12527.553333333333,
      "last_run_timestamp": "2026-09-19T08:01:53.306Z"
    },
    {
      "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": 0.7418405617977527,
      "accuracy_std_dev": 0.038139869614963295,
      "run_count": 25,
      "total_questions": 1246,
      "avg_cost_per_task_usd": 0.07584754895666132,
      "avg_duration_per_task_ms": 258067.6573033708,
      "avg_output_tokens_per_task": 5980.402889245586,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.741791919866444,
      "accuracy_std_dev": 0.04646084840389512,
      "run_count": 12,
      "total_questions": 599,
      "avg_cost_per_task_usd": 0.06980493155258766,
      "avg_duration_per_task_ms": 1045800.3572621035,
      "avg_output_tokens_per_task": 6776.651085141903,
      "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-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.7416665,
      "accuracy_std_dev": 0.030322223240554103,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.008989355000000001,
      "avg_duration_per_task_ms": 277870.45,
      "avg_output_tokens_per_task": 5790.995,
      "last_run_timestamp": "2026-09-19T08:18:29.751Z"
    },
    {
      "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": 0.7413501746835442,
      "accuracy_std_dev": 0.03375155758853722,
      "run_count": 16,
      "total_questions": 790,
      "avg_cost_per_task_usd": 0.08565632531645569,
      "avg_duration_per_task_ms": 345925.9759493671,
      "avg_output_tokens_per_task": 10900.69746835443,
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Parasail",
      "accuracy": 0.7412597005604483,
      "accuracy_std_dev": 0.13410921080552224,
      "run_count": 25,
      "total_questions": 1249,
      "avg_cost_per_task_usd": 0.09422745396317052,
      "avg_duration_per_task_ms": 253939.67013610888,
      "avg_output_tokens_per_task": 13385.784627702162,
      "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": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Together",
      "accuracy": 0.7411111666666667,
      "accuracy_std_dev": 0.016063189268393202,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04227563666666667,
      "avg_duration_per_task_ms": 199528.75333333333,
      "avg_output_tokens_per_task": 12876.693333333333,
      "last_run_timestamp": "2026-09-19T08:08:37.614Z"
    },
    {
      "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": 0.7411110000000001,
      "accuracy_std_dev": 0.01901924301858516,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.009757083333333335,
      "avg_duration_per_task_ms": 401716.95666666667,
      "avg_output_tokens_per_task": 13276.11,
      "last_run_timestamp": "2026-09-19T08:02:29.962Z"
    },
    {
      "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": 0.7410255384615385,
      "accuracy_std_dev": 0.022470704136772563,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.02259469692307692,
      "avg_duration_per_task_ms": 344926.68615384615,
      "avg_output_tokens_per_task": 11307.772307692308,
      "last_run_timestamp": "2026-09-19T08:14:32.203Z"
    },
    {
      "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": 0.7409522857142857,
      "accuracy_std_dev": 0.03499274949273386,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.028752597142857144,
      "avg_duration_per_task_ms": 301953.97142857144,
      "avg_output_tokens_per_task": 17336.282857142858,
      "last_run_timestamp": "2026-09-19T08:14:26.428Z"
    },
    {
      "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": 0.7408834144838213,
      "accuracy_std_dev": 0.03308486202761159,
      "run_count": 26,
      "total_questions": 1298,
      "avg_cost_per_task_usd": 0.23451982434514637,
      "avg_duration_per_task_ms": 25362.9437596302,
      "avg_output_tokens_per_task": 5252.202619414484,
      "last_run_timestamp": "2026-08-15T14:16:11.865Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Phala",
      "accuracy": 0.7402778333333333,
      "accuracy_std_dev": 0.039085846085536155,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.11041305000000001,
      "avg_duration_per_task_ms": 227804.8675,
      "avg_output_tokens_per_task": 5546.4675,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Wafer",
      "accuracy": 0.7400002499999999,
      "accuracy_std_dev": 0.023093866430017766,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.22648705,
      "avg_duration_per_task_ms": 279140.11,
      "avg_output_tokens_per_task": 9247.74,
      "last_run_timestamp": "2026-09-20T10:30:07.954Z"
    },
    {
      "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": 0.74,
      "accuracy_std_dev": 0.03211111264967318,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.352840368,
      "avg_duration_per_task_ms": 224909.196,
      "avg_output_tokens_per_task": 10939.108,
      "last_run_timestamp": "2026-09-19T08:02:39.143Z"
    },
    {
      "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": 0.74,
      "accuracy_std_dev": 0.013333400001499999,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.11328389999999999,
      "avg_duration_per_task_ms": 818117.716,
      "avg_output_tokens_per_task": 22197.328,
      "last_run_timestamp": "2026-09-19T08:17:23.601Z"
    },
    {
      "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": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.020059399999999998,
      "avg_duration_per_task_ms": 171159.34,
      "avg_output_tokens_per_task": 17767.96,
      "last_run_timestamp": "2026-09-19T06:00:02.503Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "AkashML",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06375521,
      "avg_duration_per_task_ms": 66151.97,
      "avg_output_tokens_per_task": 3290.09,
      "last_run_timestamp": "2026-06-20T10:18:25.780Z"
    },
    {
      "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": "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "GMICloud",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.01247214674785381,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07827993,
      "avg_duration_per_task_ms": 152857.22,
      "avg_output_tokens_per_task": 4638.005,
      "last_run_timestamp": "2026-09-19T08:18:06.927Z"
    },
    {
      "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": 0.74,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.060206680000000005,
      "avg_duration_per_task_ms": 137710.6,
      "avg_output_tokens_per_task": 4712.36,
      "last_run_timestamp": "2026-09-12T08:11:07.479Z"
    },
    {
      "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": 0.01247214674785381,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.01460845,
      "avg_duration_per_task_ms": 194359.135,
      "avg_output_tokens_per_task": 4223.43,
      "last_run_timestamp": "2026-09-19T08:19:08.255Z"
    },
    {
      "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": 0.7399997500000001,
      "accuracy_std_dev": 0.031622882011092875,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.17434206,
      "avg_duration_per_task_ms": 333592.31,
      "avg_output_tokens_per_task": 11320.165,
      "last_run_timestamp": "2026-09-19T08:12:48.095Z"
    },
    {
      "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": 0.7398561463790447,
      "accuracy_std_dev": 0.12410568998523018,
      "run_count": 26,
      "total_questions": 1298,
      "avg_cost_per_task_usd": 0.03732514175654854,
      "avg_duration_per_task_ms": 125480.83204930663,
      "avg_output_tokens_per_task": 5421.03312788906,
      "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": "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": 0.7394636465517241,
      "accuracy_std_dev": 0.03855261066843646,
      "run_count": 14,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.027936172413793103,
      "avg_duration_per_task_ms": 390627.1882183908,
      "avg_output_tokens_per_task": 12578.12356321839,
      "last_run_timestamp": "2026-09-19T08:12:27.057Z"
    },
    {
      "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": 0.7393940909090909,
      "accuracy_std_dev": 0.022822002176745397,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.5216491690909092,
      "avg_duration_per_task_ms": 720472.1036363636,
      "avg_output_tokens_per_task": 8824.263636363636,
      "last_run_timestamp": "2026-09-19T08:12:49.147Z"
    },
    {
      "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": 0.7393334,
      "accuracy_std_dev": 0.04024381912592295,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.161637104,
      "avg_duration_per_task_ms": 338949.096,
      "avg_output_tokens_per_task": 23077.328,
      "last_run_timestamp": "2026-09-19T08:16:41.340Z"
    },
    {
      "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": 0.7389743076923077,
      "accuracy_std_dev": 0.037834173804165296,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.06794495307692308,
      "avg_duration_per_task_ms": 166675.83615384615,
      "avg_output_tokens_per_task": 5442.511538461538,
      "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": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Relace",
      "accuracy": 0.7386666000000001,
      "accuracy_std_dev": 0.04470162612075761,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.005570916,
      "avg_duration_per_task_ms": 333893.98,
      "avg_output_tokens_per_task": 16678.144,
      "last_run_timestamp": "2026-09-19T08:07:54.230Z"
    },
    {
      "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": 0.7380952857142857,
      "accuracy_std_dev": 0.02487169616086462,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.0953370257142857,
      "avg_duration_per_task_ms": 895388.7342857143,
      "avg_output_tokens_per_task": 40354.45428571429,
      "last_run_timestamp": "2026-09-19T16:00:28.635Z"
    },
    {
      "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": 0.7377779999999999,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.030773986666666673,
      "avg_duration_per_task_ms": 124426.42666666667,
      "avg_output_tokens_per_task": 16143.133333333333,
      "last_run_timestamp": "2026-09-19T08:12:32.887Z"
    },
    {
      "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": 0.7377779999999999,
      "accuracy_std_dev": 0.01257094435593444,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08044259333333334,
      "avg_duration_per_task_ms": 104455.07333333333,
      "avg_output_tokens_per_task": 4500.253333333333,
      "last_run_timestamp": "2026-09-19T08:08:42.732Z"
    },
    {
      "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": 0.7377778333333334,
      "accuracy_std_dev": 0.023700698016561092,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06858023666666667,
      "avg_duration_per_task_ms": 76768.37666666666,
      "avg_output_tokens_per_task": 8623.293333333333,
      "last_run_timestamp": "2026-09-19T08:02:29.843Z"
    },
    {
      "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": 0.7375757272727272,
      "accuracy_std_dev": 0.015640496284214565,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.008148763636363637,
      "avg_duration_per_task_ms": 115345.35272727272,
      "avg_output_tokens_per_task": 10256.063636363637,
      "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": "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": 0.7372451530612244,
      "accuracy_std_dev": 0.017853512726841714,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.022030132653061228,
      "avg_duration_per_task_ms": 127977.05612244898,
      "avg_output_tokens_per_task": 3810.637755102041,
      "last_run_timestamp": "2026-09-19T08:02:39.237Z"
    },
    {
      "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": 0.7368888666666666,
      "accuracy_std_dev": 0.02741899242706698,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.06124634666666667,
      "avg_duration_per_task_ms": 229793.576,
      "avg_output_tokens_per_task": 9405.036,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.7367347959183673,
      "accuracy_std_dev": 0.023800704305545237,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.01745185306122449,
      "avg_duration_per_task_ms": 351822.13469387754,
      "avg_output_tokens_per_task": 13937.518367346938,
      "last_run_timestamp": "2026-09-20T10:30:07.536Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Fireworks",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.17216302,
      "avg_duration_per_task_ms": 234612.69,
      "avg_output_tokens_per_task": 17080.8,
      "last_run_timestamp": "2026-09-19T08:12:02.180Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Phala",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.020005850000000002,
      "avg_duration_per_task_ms": 329585.74,
      "avg_output_tokens_per_task": 11689.47,
      "last_run_timestamp": "2026-08-29T16:43:42.382Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7365892425940752,
      "accuracy_std_dev": 0.023972801823733503,
      "run_count": 25,
      "total_questions": 1249,
      "avg_cost_per_task_usd": 0.10787875340272218,
      "avg_duration_per_task_ms": 162741.05764611688,
      "avg_output_tokens_per_task": 8093.5548438751,
      "last_run_timestamp": "2026-09-19T08:15:13.229Z"
    },
    {
      "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": 0.7365404329896906,
      "accuracy_std_dev": 0.04686504205487875,
      "run_count": 6,
      "total_questions": 291,
      "avg_cost_per_task_usd": 0.08916013058419243,
      "avg_duration_per_task_ms": 682362.1065292096,
      "avg_output_tokens_per_task": 7674.209621993127,
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Venice",
      "accuracy": 0.7362636996336996,
      "accuracy_std_dev": 0.025314314324162108,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.1250874249084249,
      "avg_duration_per_task_ms": 496350.12637362635,
      "avg_output_tokens_per_task": 24530.12454212454,
      "last_run_timestamp": "2026-09-19T16:00:28.613Z"
    },
    {
      "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": 0.7360844606525911,
      "accuracy_std_dev": 0.10956776688246402,
      "run_count": 21,
      "total_questions": 1042,
      "avg_cost_per_task_usd": 0.0232232245681382,
      "avg_duration_per_task_ms": 366903.96257197694,
      "avg_output_tokens_per_task": 8722.57773512476,
      "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "Xiaomi",
      "accuracy": 0.7358309365994236,
      "accuracy_std_dev": 0.0391618221121395,
      "run_count": 14,
      "total_questions": 694,
      "avg_cost_per_task_usd": 0.021028703170028824,
      "avg_duration_per_task_ms": 301715.76657060516,
      "avg_output_tokens_per_task": 12016.485590778098,
      "last_run_timestamp": "2026-09-19T08:12:02.071Z"
    },
    {
      "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": 0.7357541832402235,
      "accuracy_std_dev": 0.09788376088289748,
      "run_count": 18,
      "total_questions": 895,
      "avg_cost_per_task_usd": 0.09288570949720674,
      "avg_duration_per_task_ms": 339472.96424581006,
      "avg_output_tokens_per_task": 14305.882681564246,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.7357256822966507,
      "accuracy_std_dev": 0.0866496604884399,
      "run_count": 21,
      "total_questions": 1045,
      "avg_cost_per_task_usd": 0.04395789377990429,
      "avg_duration_per_task_ms": 207924.5062200957,
      "avg_output_tokens_per_task": 6683.989473684211,
      "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": "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": "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": 0.7355556666666666,
      "accuracy_std_dev": 0.011331263428624758,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.25940026000000005,
      "avg_duration_per_task_ms": 198685.81333333332,
      "avg_output_tokens_per_task": 13136.706666666667,
      "last_run_timestamp": "2026-09-19T08:07:33.248Z"
    },
    {
      "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": 0.7355555925925926,
      "accuracy_std_dev": 0.028688777003300966,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.03225971111111112,
      "avg_duration_per_task_ms": 115799.6525925926,
      "avg_output_tokens_per_task": 5434.608148148148,
      "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": "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.7351751940700809,
      "accuracy_std_dev": 0.08409205356369888,
      "run_count": 15,
      "total_questions": 742,
      "avg_cost_per_task_usd": 0.05205572641509434,
      "avg_duration_per_task_ms": 112166.10916442049,
      "avg_output_tokens_per_task": 4389.22371967655,
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Parasail",
      "accuracy": 0.7351129425051335,
      "accuracy_std_dev": 0.026203090894778056,
      "run_count": 10,
      "total_questions": 487,
      "avg_cost_per_task_usd": 0.016946958932238188,
      "avg_duration_per_task_ms": 95053.74127310062,
      "avg_output_tokens_per_task": 5135.201232032854,
      "last_run_timestamp": "2026-09-19T16:00:28.637Z"
    },
    {
      "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": 0.735,
      "accuracy_std_dev": 0.15570928990590127,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.1830062116666667,
      "avg_duration_per_task_ms": 302480.2633333333,
      "avg_output_tokens_per_task": 14547.835833333333,
      "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": "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": 0.7349019999999999,
      "accuracy_std_dev": 0.03131376846391639,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.3548720011764706,
      "avg_duration_per_task_ms": 207107.23529411765,
      "avg_output_tokens_per_task": 10311.083529411764,
      "last_run_timestamp": "2026-09-19T08:12:57.394Z"
    },
    {
      "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": 0.7346031428571429,
      "accuracy_std_dev": 0.06504760378682801,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.05404573238095239,
      "avg_duration_per_task_ms": 148223.61523809523,
      "avg_output_tokens_per_task": 6767.127619047619,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7339999,
      "accuracy_std_dev": 0.02772076878966382,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.07026960600000001,
      "avg_duration_per_task_ms": 306542.06,
      "avg_output_tokens_per_task": 10129.164,
      "last_run_timestamp": "2026-09-19T08:15:44.224Z"
    },
    {
      "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": 0.7338094642857144,
      "accuracy_std_dev": 0.04249849817302972,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.037900790000000004,
      "avg_duration_per_task_ms": 160187.04071428571,
      "avg_output_tokens_per_task": 5735.9064285714285,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Modal",
      "accuracy": 0.7333335,
      "accuracy_std_dev": 0.009428090417293783,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.38862442999999997,
      "avg_duration_per_task_ms": 146345.67333333334,
      "avg_output_tokens_per_task": 10987.79,
      "last_run_timestamp": "2026-09-19T08:08:42.374Z"
    },
    {
      "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-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "SiliconFlow",
      "accuracy": 0.733333347826087,
      "accuracy_std_dev": 0.1334925998058413,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.07098599043478261,
      "avg_duration_per_task_ms": 169848.0147826087,
      "avg_output_tokens_per_task": 6912.339130434782,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Decart",
      "accuracy": 0.7333333333333333,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.008480786666666665,
      "avg_duration_per_task_ms": 397846.0266666667,
      "avg_output_tokens_per_task": 19127.326666666668,
      "last_run_timestamp": "2026-08-22T08:38:00.859Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Fireworks",
      "accuracy": 0.7333333333333333,
      "accuracy_std_dev": 0.024944382578492966,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.035478496666666665,
      "avg_duration_per_task_ms": 63670.043333333335,
      "avg_output_tokens_per_task": 4095.423333333333,
      "last_run_timestamp": "2026-05-23T14:30:04.906Z"
    },
    {
      "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": 0.7333333333333333,
      "accuracy_std_dev": 0.03250838448428001,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.06780395555555555,
      "avg_duration_per_task_ms": 133879.79333333333,
      "avg_output_tokens_per_task": 7446.308888888889,
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Mancer 2",
      "accuracy": 0.7333333333333333,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09733843333333332,
      "avg_duration_per_task_ms": 435402.0933333333,
      "avg_output_tokens_per_task": 12527.286666666667,
      "last_run_timestamp": "2026-09-19T08:13:35.856Z"
    },
    {
      "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": 0.73333325,
      "accuracy_std_dev": 0.018257509870940786,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.018071915,
      "avg_duration_per_task_ms": 95128.1,
      "avg_output_tokens_per_task": 4179.29,
      "last_run_timestamp": "2026-09-19T08:08:46.469Z"
    },
    {
      "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": "openai/gpt-5.2-20251211",
      "model_slug": "openai/gpt-5.2",
      "model_name": "OpenAI: GPT-5.2",
      "provider_name": "auto-routing",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.2412942,
      "avg_duration_per_task_ms": 207134.84,
      "avg_output_tokens_per_task": 12538.34,
      "last_run_timestamp": "2026-08-01T08:06:12.346Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Inceptron",
      "accuracy": 0.7329792361275088,
      "accuracy_std_dev": 0.03497211277654473,
      "run_count": 17,
      "total_questions": 847,
      "avg_cost_per_task_usd": 0.052468580873671775,
      "avg_duration_per_task_ms": 188219.26328217238,
      "avg_output_tokens_per_task": 7981.263282172373,
      "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": "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": 0.7328770821917807,
      "accuracy_std_dev": 0.020632433065981837,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.011465595890410957,
      "avg_duration_per_task_ms": 215673.5684931507,
      "avg_output_tokens_per_task": 4199.7054794520545,
      "last_run_timestamp": "2026-09-19T08:07:58.188Z"
    },
    {
      "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": 0.7328571428571429,
      "accuracy_std_dev": 0.02487604007260557,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.3672934914285714,
      "avg_duration_per_task_ms": 232269.72285714286,
      "avg_output_tokens_per_task": 10834.185714285713,
      "last_run_timestamp": "2026-09-19T08:15:37.640Z"
    },
    {
      "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": "SiliconFlow",
      "accuracy": 0.7328571428571429,
      "accuracy_std_dev": 0.02986363565766697,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.01948525,
      "avg_duration_per_task_ms": 74288.42,
      "avg_output_tokens_per_task": 3454.5,
      "last_run_timestamp": "2026-06-06T18:57:18.338Z"
    },
    {
      "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": 0.7328309698492462,
      "accuracy_std_dev": 0.02818147501795285,
      "run_count": 4,
      "total_questions": 199,
      "avg_cost_per_task_usd": 0.07178455778894473,
      "avg_duration_per_task_ms": 101856.46733668342,
      "avg_output_tokens_per_task": 4297.814070351758,
      "last_run_timestamp": "2026-09-19T08:01:36.430Z"
    },
    {
      "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": 0.7327989038076153,
      "accuracy_std_dev": 0.1632218869480668,
      "run_count": 20,
      "total_questions": 998,
      "avg_cost_per_task_usd": 0.015171186372745492,
      "avg_duration_per_task_ms": 223130.1102204409,
      "avg_output_tokens_per_task": 8631.145290581162,
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Z.AI",
      "accuracy": 0.7327778333333334,
      "accuracy_std_dev": 0.04525304505193238,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.09581732333333334,
      "avg_duration_per_task_ms": 160631.045,
      "avg_output_tokens_per_task": 8478.698333333334,
      "last_run_timestamp": "2026-09-19T08:07:22.421Z"
    },
    {
      "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": 0.7326667,
      "accuracy_std_dev": 0.04049139881271083,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.06109215,
      "avg_duration_per_task_ms": 223095.216,
      "avg_output_tokens_per_task": 7046.56,
      "last_run_timestamp": "2026-08-29T08:12:20.454Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.7326390000000002,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.014196437499999999,
      "avg_duration_per_task_ms": 145457.3125,
      "avg_output_tokens_per_task": 11907.520833333334,
      "last_run_timestamp": "2026-09-19T08:13:54.058Z"
    },
    {
      "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": 0.7326315789473684,
      "accuracy_std_dev": 0.099281507423305,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.07054750842105263,
      "avg_duration_per_task_ms": 389640.21789473685,
      "avg_output_tokens_per_task": 37304.677894736844,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "GMICloud",
      "accuracy": 0.7326006153846155,
      "accuracy_std_dev": 0.02879146589925737,
      "run_count": 11,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.09310996886446886,
      "avg_duration_per_task_ms": 226920.88644688643,
      "avg_output_tokens_per_task": 8628.521978021978,
      "last_run_timestamp": "2026-09-19T08:02:39.232Z"
    },
    {
      "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": 0.7325926666666667,
      "accuracy_std_dev": 0.02318878093197464,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.00917222,
      "avg_duration_per_task_ms": 347978.0577777778,
      "avg_output_tokens_per_task": 13519.62,
      "last_run_timestamp": "2026-09-19T08:14:36.215Z"
    },
    {
      "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": 0.7322223333333333,
      "accuracy_std_dev": 0.024770556995935476,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07378156,
      "avg_duration_per_task_ms": 94238.19333333333,
      "avg_output_tokens_per_task": 11168.733333333334,
      "last_run_timestamp": "2026-09-19T16:00:24.103Z"
    },
    {
      "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": 0.7322221666666667,
      "accuracy_std_dev": 0.036599313561034026,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03977107333333332,
      "avg_duration_per_task_ms": 224271.51666666666,
      "avg_output_tokens_per_task": 20806.72333333333,
      "last_run_timestamp": "2026-09-19T08:01:53.320Z"
    },
    {
      "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": 0.7321211818181819,
      "accuracy_std_dev": 0.03937046246944339,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.04438583818181819,
      "avg_duration_per_task_ms": 119798.46909090909,
      "avg_output_tokens_per_task": 8507.941818181818,
      "last_run_timestamp": "2026-09-19T08:15:44.608Z"
    },
    {
      "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": 0.7318549798387096,
      "accuracy_std_dev": 0.04649888648139865,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.08042021975806452,
      "avg_duration_per_task_ms": 208719.5060483871,
      "avg_output_tokens_per_task": 8252.131048387097,
      "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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "NextBit",
      "accuracy": 0.7315789473684211,
      "accuracy_std_dev": 0.061133249354843225,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.0719111105263158,
      "avg_duration_per_task_ms": 470454.90105263155,
      "avg_output_tokens_per_task": 22870.695789473684,
      "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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7314969432387313,
      "accuracy_std_dev": 0.09382820702418469,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.07530729799666108,
      "avg_duration_per_task_ms": 205955.36894824708,
      "avg_output_tokens_per_task": 22577.263772954924,
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.7311556834170854,
      "accuracy_std_dev": 0.027058284160943663,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.02928007286432161,
      "avg_duration_per_task_ms": 481200.60552763817,
      "avg_output_tokens_per_task": 8314.90703517588,
      "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": "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-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Cloudflare",
      "accuracy": 0.730909090909091,
      "accuracy_std_dev": 0.04541816726107927,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.020825716363636365,
      "avg_duration_per_task_ms": 64202.28,
      "avg_output_tokens_per_task": 2951.6781818181817,
      "last_run_timestamp": "2026-05-23T09:28:32.982Z"
    },
    {
      "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": 0.7302799198473282,
      "accuracy_std_dev": 0.08316610177772252,
      "run_count": 21,
      "total_questions": 1048,
      "avg_cost_per_task_usd": 0.009317546755725192,
      "avg_duration_per_task_ms": 128886.75190839695,
      "avg_output_tokens_per_task": 7970.051526717557,
      "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": "Novita",
      "accuracy": 0.7300784646464646,
      "accuracy_std_dev": 0.0316177947350811,
      "run_count": 12,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04173444949494949,
      "avg_duration_per_task_ms": 219437.9090909091,
      "avg_output_tokens_per_task": 8350.580808080807,
      "last_run_timestamp": "2026-09-19T08:02:40.716Z"
    },
    {
      "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": 0.7300089191374662,
      "accuracy_std_dev": 0.028871736576028047,
      "run_count": 15,
      "total_questions": 742,
      "avg_cost_per_task_usd": 0.05800620754716981,
      "avg_duration_per_task_ms": 288668.6940700809,
      "avg_output_tokens_per_task": 8401.919137466308,
      "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": "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Phala",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.016666999999999987,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.029231550000000005,
      "avg_duration_per_task_ms": 115126.35,
      "avg_output_tokens_per_task": 15983.89,
      "last_run_timestamp": "2026-09-19T08:14:06.082Z"
    },
    {
      "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": 0.73,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.013875090000000001,
      "avg_duration_per_task_ms": 398843.69,
      "avg_output_tokens_per_task": 15479.71,
      "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": "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": "Relace",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.17840627,
      "avg_duration_per_task_ms": 413224.86,
      "avg_output_tokens_per_task": 8594.17,
      "last_run_timestamp": "2026-09-19T08:19:01.901Z"
    },
    {
      "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": 0.73,
      "accuracy_std_dev": 0.011385630973585375,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.26713958666666665,
      "avg_duration_per_task_ms": 328724.2033333333,
      "avg_output_tokens_per_task": 9769.553333333333,
      "last_run_timestamp": "2026-09-19T08:02:39.150Z"
    },
    {
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Novita",
      "accuracy": 0.7298765555555555,
      "accuracy_std_dev": 0.03755363117160363,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.035774098518518524,
      "avg_duration_per_task_ms": 170692.90592592594,
      "avg_output_tokens_per_task": 5967.2385185185185,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "auto-routing",
      "accuracy": 0.7295238571428572,
      "accuracy_std_dev": 0.054494033042758974,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.007525531428571428,
      "avg_duration_per_task_ms": 323281.29714285716,
      "avg_output_tokens_per_task": 15952.262857142858,
      "last_run_timestamp": "2026-09-19T08:01:53.413Z"
    },
    {
      "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": 0.7293334,
      "accuracy_std_dev": 0.03336017053073918,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.5038457519999999,
      "avg_duration_per_task_ms": 191881.072,
      "avg_output_tokens_per_task": 9815.496,
      "last_run_timestamp": "2026-09-19T08:08:36.271Z"
    },
    {
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Together",
      "accuracy": 0.7288888888888889,
      "accuracy_std_dev": 0.04121608220220313,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.013316022222222224,
      "avg_duration_per_task_ms": 85589.12666666666,
      "avg_output_tokens_per_task": 3636.0377777777776,
      "last_run_timestamp": "2026-06-20T09:31:46.744Z"
    },
    {
      "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": 0.7287470402684564,
      "accuracy_std_dev": 0.02324185312940621,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.007624875838926174,
      "avg_duration_per_task_ms": 319985.19295302016,
      "avg_output_tokens_per_task": 14093.862416107382,
      "last_run_timestamp": "2026-09-19T08:08:01.540Z"
    },
    {
      "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": 0.728581156626506,
      "accuracy_std_dev": 0.046862578058083824,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.017320277108433733,
      "avg_duration_per_task_ms": 435636.6847389558,
      "avg_output_tokens_per_task": 8247.0140562249,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Together",
      "accuracy": 0.7285714285714285,
      "accuracy_std_dev": 0.04323641700120444,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.09354192,
      "avg_duration_per_task_ms": 71979.10142857143,
      "avg_output_tokens_per_task": 3596.4542857142856,
      "last_run_timestamp": "2026-06-20T10:33:11.231Z"
    },
    {
      "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": 0.7285395183946488,
      "accuracy_std_dev": 0.03368260924011414,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.07459216722408027,
      "avg_duration_per_task_ms": 211118.72575250836,
      "avg_output_tokens_per_task": 8202.498327759198,
      "last_run_timestamp": "2026-09-19T08:15:24.969Z"
    },
    {
      "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": 0.727891030612245,
      "accuracy_std_dev": 0.023240919929888738,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.32119672959183676,
      "avg_duration_per_task_ms": 168213.4693877551,
      "avg_output_tokens_per_task": 61497.88775510204,
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "SiliconFlow",
      "accuracy": 0.72777775,
      "accuracy_std_dev": 0.028523387632271777,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.07063925333333335,
      "avg_duration_per_task_ms": 211161.295,
      "avg_output_tokens_per_task": 7779.595,
      "last_run_timestamp": "2026-09-19T08:02:29.968Z"
    },
    {
      "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": "CoreWeave",
      "accuracy": 0.7277776666666667,
      "accuracy_std_dev": 0.021914694321897864,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06174203666666666,
      "avg_duration_per_task_ms": 69022.76333333334,
      "avg_output_tokens_per_task": 9250.976666666667,
      "last_run_timestamp": "2026-09-19T08:02:29.873Z"
    },
    {
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Google",
      "accuracy": 0.7276743037249284,
      "accuracy_std_dev": 0.10515567507327442,
      "run_count": 28,
      "total_questions": 1396,
      "avg_cost_per_task_usd": 0.06674029011461317,
      "avg_duration_per_task_ms": 57192.48782234957,
      "avg_output_tokens_per_task": 5553.101002865329,
      "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": "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": 0.727619,
      "accuracy_std_dev": 0.031153175806924456,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.3295257971428571,
      "avg_duration_per_task_ms": 203387.6257142857,
      "avg_output_tokens_per_task": 10294.07142857143,
      "last_run_timestamp": "2026-09-19T08:12:07.194Z"
    },
    {
      "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": 0.7274073888888889,
      "accuracy_std_dev": 0.03589348401581003,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.018933201111111114,
      "avg_duration_per_task_ms": 205864.55777777778,
      "avg_output_tokens_per_task": 6492.102222222222,
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Novita",
      "accuracy": 0.7271793846153847,
      "accuracy_std_dev": 0.020500031911939387,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.06991619076923077,
      "avg_duration_per_task_ms": 246359.73846153845,
      "avg_output_tokens_per_task": 8152.42923076923,
      "last_run_timestamp": "2026-09-19T08:14:58.451Z"
    },
    {
      "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": 0.7270382569659443,
      "accuracy_std_dev": 0.0355308519613672,
      "run_count": 13,
      "total_questions": 646,
      "avg_cost_per_task_usd": 0.03975771362229102,
      "avg_duration_per_task_ms": 168092.38235294117,
      "avg_output_tokens_per_task": 8093.3250773993805,
      "last_run_timestamp": "2026-09-19T08:12:07.843Z"
    },
    {
      "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": "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": 0.726667,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02366282,
      "avg_duration_per_task_ms": 103172.5,
      "avg_output_tokens_per_task": 4199.07,
      "last_run_timestamp": "2026-09-19T08:01:53.400Z"
    },
    {
      "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": 0.7266666818181818,
      "accuracy_std_dev": 0.0799747182275785,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.0664802909090909,
      "avg_duration_per_task_ms": 285594.60545454547,
      "avg_output_tokens_per_task": 27787.627272727274,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Parasail",
      "accuracy": 0.7266666666666667,
      "accuracy_std_dev": 0.030306899350184633,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.436704,
      "avg_duration_per_task_ms": 230592.73333333334,
      "avg_output_tokens_per_task": 8499.046666666667,
      "last_run_timestamp": "2026-09-19T08:21:11.819Z"
    },
    {
      "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": 0.7266665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06289768999999999,
      "avg_duration_per_task_ms": 327232.67,
      "avg_output_tokens_per_task": 8904.97,
      "last_run_timestamp": "2026-09-19T08:08:46.957Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Chutes",
      "accuracy": 0.726324590738423,
      "accuracy_std_dev": 0.05193905756264829,
      "run_count": 16,
      "total_questions": 799,
      "avg_cost_per_task_usd": 0.026944361702127665,
      "avg_duration_per_task_ms": 165610.5068836045,
      "avg_output_tokens_per_task": 4209.88986232791,
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Venice",
      "accuracy": 0.726,
      "accuracy_std_dev": 0.02707188783960217,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.029576488,
      "avg_duration_per_task_ms": 174487.376,
      "avg_output_tokens_per_task": 9637.224,
      "last_run_timestamp": "2026-09-19T16:00:32.511Z"
    },
    {
      "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": 0.7257141428571429,
      "accuracy_std_dev": 0.03693465701721885,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.2823705371428571,
      "avg_duration_per_task_ms": 385575.6,
      "avg_output_tokens_per_task": 8735.288571428571,
      "last_run_timestamp": "2026-09-19T08:02:40.804Z"
    },
    {
      "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": 0.7256411282051283,
      "accuracy_std_dev": 0.021901115569475008,
      "run_count": 4,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.02016126153846154,
      "avg_duration_per_task_ms": 253617.73333333334,
      "avg_output_tokens_per_task": 9900.05641025641,
      "last_run_timestamp": "2026-09-19T08:02:29.876Z"
    },
    {
      "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": 0.7255556666666666,
      "accuracy_std_dev": 0.02114024890003792,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.058183759999999994,
      "avg_duration_per_task_ms": 881187.9266666666,
      "avg_output_tokens_per_task": 30140.16,
      "last_run_timestamp": "2026-09-19T08:15:00.862Z"
    },
    {
      "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": 0.7253915503355705,
      "accuracy_std_dev": 0.1089886276971389,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.13323116275167785,
      "avg_duration_per_task_ms": 557640.1711409396,
      "avg_output_tokens_per_task": 25872.763422818793,
      "last_run_timestamp": "2026-09-19T08:15:43.906Z"
    },
    {
      "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": "DekaLLM",
      "accuracy": 0.725,
      "accuracy_std_dev": 0.05545268253204706,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.06306274499999999,
      "avg_duration_per_task_ms": 307709.445,
      "avg_output_tokens_per_task": 29336.17,
      "last_run_timestamp": "2026-05-16T11:09:22.970Z"
    },
    {
      "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": 0.7249544826958105,
      "accuracy_std_dev": 0.08103658076897909,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.056935027322404366,
      "avg_duration_per_task_ms": 478336.01821493625,
      "avg_output_tokens_per_task": 36495.10746812386,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Decart",
      "accuracy": 0.7247474713804714,
      "accuracy_std_dev": 0.08284079836897833,
      "run_count": 12,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.06815684343434343,
      "avg_duration_per_task_ms": 133143.7861952862,
      "avg_output_tokens_per_task": 12155.594276094276,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Parasail",
      "accuracy": 0.7244444999999999,
      "accuracy_std_dev": 0.032640910103988625,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.10185021333333334,
      "avg_duration_per_task_ms": 107547.30666666667,
      "avg_output_tokens_per_task": 8542.055,
      "last_run_timestamp": "2026-09-19T08:02:40.834Z"
    },
    {
      "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": 0.7244444666666666,
      "accuracy_std_dev": 0.030663471653563444,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.09078783866666666,
      "avg_duration_per_task_ms": 211616.812,
      "avg_output_tokens_per_task": 7390.202666666667,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DeepInfra",
      "accuracy": 0.7244443333333334,
      "accuracy_std_dev": 0.022662526857249517,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.47615288666666666,
      "avg_duration_per_task_ms": 770487.4266666666,
      "avg_output_tokens_per_task": 9661.306666666667,
      "last_run_timestamp": "2026-09-19T08:14:36.033Z"
    },
    {
      "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": 0.7243589615384616,
      "accuracy_std_dev": 0.033839378789852756,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.03414883846153846,
      "avg_duration_per_task_ms": 357708.1330769231,
      "avg_output_tokens_per_task": 8693.559230769231,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Venice",
      "accuracy": 0.7241015454545454,
      "accuracy_std_dev": 0.06989016607046616,
      "run_count": 19,
      "total_questions": 946,
      "avg_cost_per_task_usd": 0.05174934883720931,
      "avg_duration_per_task_ms": 224979.1617336152,
      "avg_output_tokens_per_task": 6867.136363636364,
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Ionstream",
      "accuracy": 0.7237445479452055,
      "accuracy_std_dev": 0.01534481785127761,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.060365219178082186,
      "avg_duration_per_task_ms": 536636.493150685,
      "avg_output_tokens_per_task": 12764.582191780823,
      "last_run_timestamp": "2026-09-19T08:12:49.799Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Ionstream",
      "accuracy": 0.7233335,
      "accuracy_std_dev": 0.016666499999999973,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.16905831,
      "avg_duration_per_task_ms": 501852.61,
      "avg_output_tokens_per_task": 40516.58,
      "last_run_timestamp": "2026-09-19T08:15:01.262Z"
    },
    {
      "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": 0.7233335,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.020558879999999998,
      "avg_duration_per_task_ms": 168582.3,
      "avg_output_tokens_per_task": 7751,
      "last_run_timestamp": "2026-09-19T16:00:28.636Z"
    },
    {
      "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": 0.7233335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.013037239999999999,
      "avg_duration_per_task_ms": 45998.83,
      "avg_output_tokens_per_task": 3876.1,
      "last_run_timestamp": "2026-09-19T08:01:36.457Z"
    },
    {
      "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": 0.7233333076923077,
      "accuracy_std_dev": 0.03726778815831465,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.031507099230769234,
      "avg_duration_per_task_ms": 241507.73384615386,
      "avg_output_tokens_per_task": 8188.238461538462,
      "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": "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Alibaba",
      "accuracy": 0.7225326592178771,
      "accuracy_std_dev": 0.03504286667804778,
      "run_count": 18,
      "total_questions": 895,
      "avg_cost_per_task_usd": 0.11005581787709497,
      "avg_duration_per_task_ms": 148535.98547486032,
      "avg_output_tokens_per_task": 8251.711731843576,
      "last_run_timestamp": "2026-09-19T08:12:39.649Z"
    },
    {
      "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": 0.7224880535885168,
      "accuracy_std_dev": 0.07672207222462067,
      "run_count": 21,
      "total_questions": 1045,
      "avg_cost_per_task_usd": 0.10305108325358853,
      "avg_duration_per_task_ms": 101482.16076555024,
      "avg_output_tokens_per_task": 7390.176076555024,
      "last_run_timestamp": "2026-09-19T08:18:15.795Z"
    },
    {
      "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": 0.722222316091954,
      "accuracy_std_dev": 0.031278207950127646,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.1106247183908046,
      "avg_duration_per_task_ms": 557346.2183908046,
      "avg_output_tokens_per_task": 24361.238505747126,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7222222356902357,
      "accuracy_std_dev": 0.024462876516718586,
      "run_count": 6,
      "total_questions": 297,
      "avg_cost_per_task_usd": 0.3977955387205387,
      "avg_duration_per_task_ms": 396867.7037037037,
      "avg_output_tokens_per_task": 8974.410774410775,
      "last_run_timestamp": "2026-09-19T08:08:39.796Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Novita",
      "accuracy": 0.7221293377926422,
      "accuracy_std_dev": 0.03777626478101768,
      "run_count": 24,
      "total_questions": 1196,
      "avg_cost_per_task_usd": 0.041799248327759206,
      "avg_duration_per_task_ms": 295964.41471571906,
      "avg_output_tokens_per_task": 9079.295986622074,
      "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": "Novita",
      "accuracy": 0.7220366861435726,
      "accuracy_std_dev": 0.042939457372476286,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.038994771285475786,
      "avg_duration_per_task_ms": 293176.03756260435,
      "avg_output_tokens_per_task": 8310.446577629382,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Relace",
      "accuracy": 0.7216665,
      "accuracy_std_dev": 0.0016665000000000152,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.010394970000000002,
      "avg_duration_per_task_ms": 238439.51,
      "avg_output_tokens_per_task": 14864.89,
      "last_run_timestamp": "2026-09-19T08:12:48.672Z"
    },
    {
      "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": 0.721603583518931,
      "accuracy_std_dev": 0.028097817578797066,
      "run_count": 9,
      "total_questions": 449,
      "avg_cost_per_task_usd": 0.0643464008908686,
      "avg_duration_per_task_ms": 278378.1247216036,
      "avg_output_tokens_per_task": 8733.859688195991,
      "last_run_timestamp": "2026-09-19T08:18:14.730Z"
    },
    {
      "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": 0.7214111715328467,
      "accuracy_std_dev": 0.029999031700351234,
      "run_count": 11,
      "total_questions": 548,
      "avg_cost_per_task_usd": 0.10158374817518248,
      "avg_duration_per_task_ms": 91062.5401459854,
      "avg_output_tokens_per_task": 10938.527372262773,
      "last_run_timestamp": "2026-09-19T08:15:48.329Z"
    },
    {
      "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": 0.7210941505922165,
      "accuracy_std_dev": 0.03106900471084039,
      "run_count": 12,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.1026584399323181,
      "avg_duration_per_task_ms": 291565.6379018613,
      "avg_output_tokens_per_task": 13900.038917089678,
      "last_run_timestamp": "2026-09-19T08:12:26.857Z"
    },
    {
      "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": 0.7205555555555555,
      "accuracy_std_dev": 0.06057596801741523,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.0473878211111111,
      "avg_duration_per_task_ms": 163562.02,
      "avg_output_tokens_per_task": 7193.1466666666665,
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Mara",
      "accuracy": 0.7205390303030302,
      "accuracy_std_dev": 0.006190499999999988,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.1266920202020202,
      "avg_duration_per_task_ms": 184926.9090909091,
      "avg_output_tokens_per_task": 9286.343434343435,
      "last_run_timestamp": "2026-09-19T16:00:24.088Z"
    },
    {
      "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": 0.7203921176470588,
      "accuracy_std_dev": 0.042378295969520495,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.08408886352941176,
      "avg_duration_per_task_ms": 167842.00235294117,
      "avg_output_tokens_per_task": 6586.608235294118,
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "auto-routing",
      "accuracy": 0.7201584739970283,
      "accuracy_std_dev": 0.027151418343285973,
      "run_count": 27,
      "total_questions": 1346,
      "avg_cost_per_task_usd": 0.02274276225854383,
      "avg_duration_per_task_ms": 279206.04011887073,
      "avg_output_tokens_per_task": 8300.97028231798,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "ModelRun",
      "accuracy": 0.720112529535865,
      "accuracy_std_dev": 0.054372151849016614,
      "run_count": 19,
      "total_questions": 948,
      "avg_cost_per_task_usd": 0.01960626371308017,
      "avg_duration_per_task_ms": 52898.649789029536,
      "avg_output_tokens_per_task": 3908.25,
      "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-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Cloudflare",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.11915537433269066,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.013665298181818183,
      "avg_duration_per_task_ms": 146966.81818181818,
      "avg_output_tokens_per_task": 8880.298181818182,
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03307506,
      "avg_duration_per_task_ms": 411711.74,
      "avg_output_tokens_per_task": 12561.32,
      "last_run_timestamp": "2026-09-19T08:09:58.333Z"
    },
    {
      "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": 0.72,
      "accuracy_std_dev": 0.030306862688176752,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3160981133333334,
      "avg_duration_per_task_ms": 461033.8333333333,
      "avg_output_tokens_per_task": 8961.443333333333,
      "last_run_timestamp": "2026-09-19T08:01:53.394Z"
    },
    {
      "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": "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": "DigitalOcean",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.013332999999999984,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.013628439999999999,
      "avg_duration_per_task_ms": 407344.14,
      "avg_output_tokens_per_task": 6225.73,
      "last_run_timestamp": "2026-09-19T08:19:00.695Z"
    },
    {
      "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": "Makora",
      "accuracy": 0.71999975,
      "accuracy_std_dev": 0.008164965810553043,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.010504485,
      "avg_duration_per_task_ms": 163005.685,
      "avg_output_tokens_per_task": 18389.035,
      "last_run_timestamp": "2026-09-19T08:08:45.471Z"
    },
    {
      "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": 0.7193335,
      "accuracy_std_dev": 0.02882120747036806,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.021860742000000002,
      "avg_duration_per_task_ms": 113473.506,
      "avg_output_tokens_per_task": 6563.78,
      "last_run_timestamp": "2026-09-19T16:00:28.611Z"
    },
    {
      "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": "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-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.7184294759615385,
      "accuracy_std_dev": 0.1596697904675889,
      "run_count": 21,
      "total_questions": 1040,
      "avg_cost_per_task_usd": 0.00781735096153846,
      "avg_duration_per_task_ms": 183262.19615384616,
      "avg_output_tokens_per_task": 7099.238461538462,
      "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": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Azure",
      "accuracy": 0.7180951428571429,
      "accuracy_std_dev": 0.029323408959048256,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.35473523142857144,
      "avg_duration_per_task_ms": 239903.80285714284,
      "avg_output_tokens_per_task": 10770.68857142857,
      "last_run_timestamp": "2026-09-19T08:12:57.483Z"
    },
    {
      "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": 0.7175246360294117,
      "accuracy_std_dev": 0.06146597028417599,
      "run_count": 11,
      "total_questions": 544,
      "avg_cost_per_task_usd": 0.12441467830882355,
      "avg_duration_per_task_ms": 999918.8308823529,
      "avg_output_tokens_per_task": 23722.5,
      "last_run_timestamp": "2026-09-19T16:00:30.867Z"
    },
    {
      "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": 0.7174358846153847,
      "accuracy_std_dev": 0.03935710293759606,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.03685178846153846,
      "avg_duration_per_task_ms": 327901.3692307692,
      "avg_output_tokens_per_task": 11511.836153846154,
      "last_run_timestamp": "2026-08-29T16:11:10.137Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "auto-routing",
      "accuracy": 0.7174358461538461,
      "accuracy_std_dev": 0.03186920095989991,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.06687680461538463,
      "avg_duration_per_task_ms": 239262.95230769232,
      "avg_output_tokens_per_task": 7639.2738461538465,
      "last_run_timestamp": "2026-09-19T08:13:05.171Z"
    },
    {
      "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": 0.71733335,
      "accuracy_std_dev": 0.14592836629294353,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.013306622999999998,
      "avg_duration_per_task_ms": 168498.887,
      "avg_output_tokens_per_task": 7340.109,
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "GMICloud",
      "accuracy": 0.7166666,
      "accuracy_std_dev": 0.033993437276039026,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.006462468,
      "avg_duration_per_task_ms": 374148.128,
      "avg_output_tokens_per_task": 12453.44,
      "last_run_timestamp": "2026-09-19T08:17:29.225Z"
    },
    {
      "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": "Modal",
      "accuracy": 0.7166665000000001,
      "accuracy_std_dev": 0.023333500000000007,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.041574799999999995,
      "avg_duration_per_task_ms": 130749.55,
      "avg_output_tokens_per_task": 23910.86,
      "last_run_timestamp": "2026-09-19T08:12:53.527Z"
    },
    {
      "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": "DeepInfra",
      "accuracy": 0.7162163243243244,
      "accuracy_std_dev": 0.04755751813214069,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.07101010135135136,
      "avg_duration_per_task_ms": 195853.2972972973,
      "avg_output_tokens_per_task": 10371.256756756757,
      "last_run_timestamp": "2026-06-30T19:48:55.167Z"
    },
    {
      "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": 0.7161505810564663,
      "accuracy_std_dev": 0.134318387636952,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.07765216575591986,
      "avg_duration_per_task_ms": 227644.69489981784,
      "avg_output_tokens_per_task": 5419.0373406193075,
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7161505118397086,
      "accuracy_std_dev": 0.03840795876268506,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.04527011657559199,
      "avg_duration_per_task_ms": 358551.0282331512,
      "avg_output_tokens_per_task": 7699.158469945355,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Together",
      "accuracy": 0.716,
      "accuracy_std_dev": 0.024979991993593582,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.035929062000000005,
      "avg_duration_per_task_ms": 69854.32,
      "avg_output_tokens_per_task": 3091.818,
      "last_run_timestamp": "2026-05-16T13:21:53.025Z"
    },
    {
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.715,
      "accuracy_std_dev": 0.0016670000000000296,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04256534,
      "avg_duration_per_task_ms": 221321.22,
      "avg_output_tokens_per_task": 26641.22,
      "last_run_timestamp": "2026-09-19T08:12:49.670Z"
    },
    {
      "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": 0.715,
      "accuracy_std_dev": 0.05624832837071695,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.17734545000000002,
      "avg_duration_per_task_ms": 304190.625,
      "avg_output_tokens_per_task": 10162.455,
      "last_run_timestamp": "2026-09-19T08:01:53.330Z"
    },
    {
      "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": "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": 0.71466672,
      "accuracy_std_dev": 0.03461532403144019,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.04134587520000001,
      "avg_duration_per_task_ms": 175611.5344,
      "avg_output_tokens_per_task": 5977.7496,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Novita",
      "accuracy": 0.7145237857142858,
      "accuracy_std_dev": 0.03996240254052832,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.038241464999999995,
      "avg_duration_per_task_ms": 157853.17285714287,
      "avg_output_tokens_per_task": 5122.270714285714,
      "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": "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Mara",
      "accuracy": 0.7141900923694778,
      "accuracy_std_dev": 0.03172465567661847,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.03655370080321286,
      "avg_duration_per_task_ms": 44500.680722891564,
      "avg_output_tokens_per_task": 5750.947791164659,
      "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-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Friendli",
      "accuracy": 0.7141667083333333,
      "accuracy_std_dev": 0.07459301504524356,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.05860941499999999,
      "avg_duration_per_task_ms": 55370.06916666667,
      "avg_output_tokens_per_task": 5220.491666666667,
      "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": "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": 0.7138048585858586,
      "accuracy_std_dev": 0.07263833137825143,
      "run_count": 12,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04328885521885522,
      "avg_duration_per_task_ms": 234338.40235690237,
      "avg_output_tokens_per_task": 24374.796296296296,
      "last_run_timestamp": "2026-09-19T16:00:22.536Z"
    },
    {
      "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": 0.7137096552419355,
      "accuracy_std_dev": 0.08299828694453579,
      "run_count": 20,
      "total_questions": 992,
      "avg_cost_per_task_usd": 0.05201826612903226,
      "avg_duration_per_task_ms": 463441.87298387097,
      "avg_output_tokens_per_task": 37741.06451612903,
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Novita",
      "accuracy": 0.7135207630522089,
      "accuracy_std_dev": 0.04226729284636053,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.0358368875502008,
      "avg_duration_per_task_ms": 206817.1746987952,
      "avg_output_tokens_per_task": 6480.377510040161,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Venice",
      "accuracy": 0.7133335,
      "accuracy_std_dev": 0.006666499999999964,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.041616520000000004,
      "avg_duration_per_task_ms": 141684.91,
      "avg_output_tokens_per_task": 13886.08,
      "last_run_timestamp": "2026-09-19T08:06:54.323Z"
    },
    {
      "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": 0.7133335,
      "accuracy_std_dev": 0.02708037419885478,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.31487034,
      "avg_duration_per_task_ms": 416220.92,
      "avg_output_tokens_per_task": 8817.575,
      "last_run_timestamp": "2026-09-19T08:08:01.627Z"
    },
    {
      "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": "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7133332857142858,
      "accuracy_std_dev": 0.03703300972960005,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.32220135714285714,
      "avg_duration_per_task_ms": 407576.0685714286,
      "avg_output_tokens_per_task": 9593.411428571428,
      "last_run_timestamp": "2026-09-19T08:14:26.342Z"
    },
    {
      "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": "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Venice",
      "accuracy": 0.713333,
      "accuracy_std_dev": 0.037416573867739444,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.23679781500000005,
      "avg_duration_per_task_ms": 241514.25,
      "avg_output_tokens_per_task": 20554.75,
      "last_run_timestamp": "2026-09-19T08:07:26.392Z"
    },
    {
      "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": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.03806284,
      "avg_duration_per_task_ms": 229552.52,
      "avg_output_tokens_per_task": 12003.52,
      "last_run_timestamp": "2026-09-19T08:01:53.307Z"
    },
    {
      "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": 0.7131498513761468,
      "accuracy_std_dev": 0.15547389772642844,
      "run_count": 11,
      "total_questions": 545,
      "avg_cost_per_task_usd": 0.07655383853211008,
      "avg_duration_per_task_ms": 42631.811009174315,
      "avg_output_tokens_per_task": 4027.154128440367,
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "WandB",
      "accuracy": 0.7123519847715736,
      "accuracy_std_dev": 0.02073271203919062,
      "run_count": 4,
      "total_questions": 197,
      "avg_cost_per_task_usd": 0.08227161928934011,
      "avg_duration_per_task_ms": 178521.6345177665,
      "avg_output_tokens_per_task": 15413.27918781726,
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Friendli",
      "accuracy": 0.7121428571428572,
      "accuracy_std_dev": 0.03285629950656808,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.06669165142857143,
      "avg_duration_per_task_ms": 190237.55285714284,
      "avg_output_tokens_per_task": 9631.787857142857,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "BaseTen",
      "accuracy": 0.7121040825791854,
      "accuracy_std_dev": 0.036688548792126165,
      "run_count": 18,
      "total_questions": 884,
      "avg_cost_per_task_usd": 0.09261668552036198,
      "avg_duration_per_task_ms": 73568.46040723982,
      "avg_output_tokens_per_task": 7541.7307692307695,
      "last_run_timestamp": "2026-09-19T08:14:35.848Z"
    },
    {
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "auto-routing",
      "accuracy": 0.7116858275862069,
      "accuracy_std_dev": 0.05977603864750285,
      "run_count": 14,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.024618107758620687,
      "avg_duration_per_task_ms": 136410.2212643678,
      "avg_output_tokens_per_task": 6874.7801724137935,
      "last_run_timestamp": "2026-09-19T16:00:24.105Z"
    },
    {
      "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": 0.7116787627737226,
      "accuracy_std_dev": 0.14935343511392893,
      "run_count": 11,
      "total_questions": 548,
      "avg_cost_per_task_usd": 0.10796777189781023,
      "avg_duration_per_task_ms": 148941.3010948905,
      "avg_output_tokens_per_task": 7939.4324817518245,
      "last_run_timestamp": "2026-09-19T08:12:25.116Z"
    },
    {
      "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": 0.7115308465171193,
      "accuracy_std_dev": 0.1512936932248644,
      "run_count": 17,
      "total_questions": 847,
      "avg_cost_per_task_usd": 0.0987406765053129,
      "avg_duration_per_task_ms": 230873.95631641085,
      "avg_output_tokens_per_task": 6065.546635182999,
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Fireworks",
      "accuracy": 0.7114285714285714,
      "accuracy_std_dev": 0.04323641700120444,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.013884184285714283,
      "avg_duration_per_task_ms": 81133.41285714286,
      "avg_output_tokens_per_task": 3767.2885714285712,
      "last_run_timestamp": "2026-06-20T08:38:34.608Z"
    },
    {
      "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": 0.7110792363896848,
      "accuracy_std_dev": 0.04266601421056426,
      "run_count": 28,
      "total_questions": 1396,
      "avg_cost_per_task_usd": 0.01819265616045845,
      "avg_duration_per_task_ms": 135492.12392550142,
      "avg_output_tokens_per_task": 5865.690544412608,
      "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": "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": 0.711,
      "accuracy_std_dev": 0.034786648141204975,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.014063111,
      "avg_duration_per_task_ms": 146044.761,
      "avg_output_tokens_per_task": 4579.124,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7109804117647058,
      "accuracy_std_dev": 0.05357489628334922,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.07005771294117646,
      "avg_duration_per_task_ms": 197809.54470588235,
      "avg_output_tokens_per_task": 7675.065882352941,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Phala",
      "accuracy": 0.7108696086956522,
      "accuracy_std_dev": 0.03557966787700785,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.04292889913043479,
      "avg_duration_per_task_ms": 192476.08521739132,
      "avg_output_tokens_per_task": 4315.182608695653,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.71033335,
      "accuracy_std_dev": 0.028555927386927915,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.031039679999999997,
      "avg_duration_per_task_ms": 206099.06,
      "avg_output_tokens_per_task": 6448.078,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7102565384615385,
      "accuracy_std_dev": 0.0380387521707805,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.04511441076923077,
      "avg_duration_per_task_ms": 639644.943076923,
      "avg_output_tokens_per_task": 6877.484615384616,
      "last_run_timestamp": "2026-08-29T16:47:05.512Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.71,
      "accuracy_std_dev": 0.023332999999999993,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.022729809999999996,
      "avg_duration_per_task_ms": 177235.62,
      "avg_output_tokens_per_task": 14095.11,
      "last_run_timestamp": "2026-09-19T08:19:02.035Z"
    },
    {
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Parasail",
      "accuracy": 0.71,
      "accuracy_std_dev": 0.03958114029012637,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.019811326666666667,
      "avg_duration_per_task_ms": 130532.76,
      "avg_output_tokens_per_task": 3719.2966666666666,
      "last_run_timestamp": "2026-05-16T23:02:52.223Z"
    },
    {
      "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": "DekaLLM",
      "accuracy": 0.71,
      "accuracy_std_dev": 0.043588989435406726,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.02106588,
      "avg_duration_per_task_ms": 66495.715,
      "avg_output_tokens_per_task": 1737.11,
      "last_run_timestamp": "2026-05-02T22:49:37.642Z"
    },
    {
      "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": 0.7098539343065693,
      "accuracy_std_dev": 0.053780596538925134,
      "run_count": 11,
      "total_questions": 548,
      "avg_cost_per_task_usd": 0.09852927554744526,
      "avg_duration_per_task_ms": 122619.91970802919,
      "avg_output_tokens_per_task": 8950.552919708029,
      "last_run_timestamp": "2026-09-19T08:06:45.320Z"
    },
    {
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Mancer 2",
      "accuracy": 0.7093334,
      "accuracy_std_dev": 0.04572378520463939,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.041164836,
      "avg_duration_per_task_ms": 324796.104,
      "avg_output_tokens_per_task": 14246.772,
      "last_run_timestamp": "2026-09-19T08:07:37.463Z"
    },
    {
      "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": 0.7091666875,
      "accuracy_std_dev": 0.07271997399676271,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.047436745,
      "avg_duration_per_task_ms": 166812.7925,
      "avg_output_tokens_per_task": 6450.30625,
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "CoreWeave",
      "accuracy": 0.7088888888888889,
      "accuracy_std_dev": 0.032810672760363424,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.08025308222222222,
      "avg_duration_per_task_ms": 72592.39333333333,
      "avg_output_tokens_per_task": 8513.65111111111,
      "last_run_timestamp": "2026-09-19T08:07:49.572Z"
    },
    {
      "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-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Phala",
      "accuracy": 0.7087543737373738,
      "accuracy_std_dev": 0.06057975459466635,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.2684698484848485,
      "avg_duration_per_task_ms": 369025.3080808081,
      "avg_output_tokens_per_task": 8958.444444444445,
      "last_run_timestamp": "2026-09-19T08:20:02.696Z"
    },
    {
      "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": 0.7087053928571428,
      "accuracy_std_dev": 0.05722548330707566,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.04908466741071428,
      "avg_duration_per_task_ms": 738553.59375,
      "avg_output_tokens_per_task": 50678.734375,
      "last_run_timestamp": "2026-09-12T16:00:26.253Z"
    },
    {
      "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": 0.7083335,
      "accuracy_std_dev": 0.00833350000000005,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.040671439999999996,
      "avg_duration_per_task_ms": 147596.48,
      "avg_output_tokens_per_task": 22423.42,
      "last_run_timestamp": "2026-09-19T08:09:43.896Z"
    },
    {
      "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": 0.7083334549549549,
      "accuracy_std_dev": 0.032652328452317504,
      "run_count": 9,
      "total_questions": 444,
      "avg_cost_per_task_usd": 0.020894054054054055,
      "avg_duration_per_task_ms": 99127.93243243243,
      "avg_output_tokens_per_task": 6316.898648648648,
      "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/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Venice",
      "accuracy": 0.70833325,
      "accuracy_std_dev": 0.02511102568967462,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.37473283499999993,
      "avg_duration_per_task_ms": 183894.045,
      "avg_output_tokens_per_task": 11774.475,
      "last_run_timestamp": "2026-09-19T08:06:51.776Z"
    },
    {
      "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": 0.7078693118040089,
      "accuracy_std_dev": 0.037495695230779585,
      "run_count": 9,
      "total_questions": 449,
      "avg_cost_per_task_usd": 0.13154344097995546,
      "avg_duration_per_task_ms": 74759.52783964365,
      "avg_output_tokens_per_task": 10228.11135857461,
      "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-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7073808571428571,
      "accuracy_std_dev": 0.0460964802078317,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.02892358,
      "avg_duration_per_task_ms": 123301.34714285714,
      "avg_output_tokens_per_task": 4939.147142857143,
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Novita",
      "accuracy": 0.7073252731854838,
      "accuracy_std_dev": 0.20674647626668274,
      "run_count": 20,
      "total_questions": 992,
      "avg_cost_per_task_usd": 0.10164331249999999,
      "avg_duration_per_task_ms": 197665.9133064516,
      "avg_output_tokens_per_task": 10714.620967741936,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Fireworks",
      "accuracy": 0.7069688236775818,
      "accuracy_std_dev": 0.03585183782858977,
      "run_count": 8,
      "total_questions": 397,
      "avg_cost_per_task_usd": 0.3883721914357683,
      "avg_duration_per_task_ms": 201494.02770780856,
      "avg_output_tokens_per_task": 8021.390428211587,
      "last_run_timestamp": "2026-09-19T08:15:14.726Z"
    },
    {
      "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": 0.706867824120603,
      "accuracy_std_dev": 0.024112945099248537,
      "run_count": 8,
      "total_questions": 398,
      "avg_cost_per_task_usd": 0.12147276381909548,
      "avg_duration_per_task_ms": 366567.0226130653,
      "avg_output_tokens_per_task": 8862.291457286432,
      "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": "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": "Morph",
      "accuracy": 0.706667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02753204,
      "avg_duration_per_task_ms": 825683.52,
      "avg_output_tokens_per_task": 15373.7,
      "last_run_timestamp": "2026-09-12T08:11:33.505Z"
    },
    {
      "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": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Io Net",
      "accuracy": 0.7066666666666667,
      "accuracy_std_dev": 0.033993463423951896,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.017786823333333337,
      "avg_duration_per_task_ms": 128082.29,
      "avg_output_tokens_per_task": 8061.84,
      "last_run_timestamp": "2026-05-23T11:46:07.128Z"
    },
    {
      "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": 0.7064618061465722,
      "accuracy_std_dev": 0.09810038286783905,
      "run_count": 17,
      "total_questions": 846,
      "avg_cost_per_task_usd": 0.11687022458628843,
      "avg_duration_per_task_ms": 102570.99054373523,
      "avg_output_tokens_per_task": 7925.6973995271865,
      "last_run_timestamp": "2026-09-20T10:30:07.563Z"
    },
    {
      "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": 0.7061538717948719,
      "accuracy_std_dev": 0.03305471857947331,
      "run_count": 39,
      "total_questions": 1950,
      "avg_cost_per_task_usd": 0.024778430256410244,
      "avg_duration_per_task_ms": 140799.78256410256,
      "avg_output_tokens_per_task": 6649.691794871795,
      "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": "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7048717692307692,
      "accuracy_std_dev": 0.05145929153321672,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.03895267076923078,
      "avg_duration_per_task_ms": 181504.9023076923,
      "avg_output_tokens_per_task": 5031.964615384615,
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7046486190476191,
      "accuracy_std_dev": 0.037923417525396554,
      "run_count": 6,
      "total_questions": 294,
      "avg_cost_per_task_usd": 0.11717974149659866,
      "avg_duration_per_task_ms": 474565.40476190473,
      "avg_output_tokens_per_task": 22901.343537414967,
      "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": "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": 0.008193151999999999,
      "avg_duration_per_task_ms": 308148.616,
      "avg_output_tokens_per_task": 14154.342,
      "last_run_timestamp": "2026-09-19T08:01:55.621Z"
    },
    {
      "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": 0.7033335,
      "accuracy_std_dev": 0.016666499999999973,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00633007,
      "avg_duration_per_task_ms": 346835.76,
      "avg_output_tokens_per_task": 4243.93,
      "last_run_timestamp": "2026-09-19T08:15:35.180Z"
    },
    {
      "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": "Inceptron",
      "accuracy": 0.7033333333333334,
      "accuracy_std_dev": 0.04955356249106169,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.02144292,
      "avg_duration_per_task_ms": 65470.65,
      "avg_output_tokens_per_task": 3098.5466666666666,
      "last_run_timestamp": "2026-05-02T08:19:29.051Z"
    },
    {
      "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": 0.7033330000000001,
      "accuracy_std_dev": 0.030000000000000027,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03169822,
      "avg_duration_per_task_ms": 206845.29,
      "avg_output_tokens_per_task": 16212.42,
      "last_run_timestamp": "2026-09-19T08:09:55.413Z"
    },
    {
      "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-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "GMICloud",
      "accuracy": 0.70166675,
      "accuracy_std_dev": 0.03507946138679295,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.155233565,
      "avg_duration_per_task_ms": 432057.795,
      "avg_output_tokens_per_task": 9028.84,
      "last_run_timestamp": "2026-09-19T08:03:46.173Z"
    },
    {
      "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": "Parasail",
      "accuracy": 0.7016666666666667,
      "accuracy_std_dev": 0.045061686115319244,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.022215448333333332,
      "avg_duration_per_task_ms": 72920.13666666667,
      "avg_output_tokens_per_task": 3012.59,
      "last_run_timestamp": "2026-06-06T08:18:22.655Z"
    },
    {
      "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": 0.7015590066815145,
      "accuracy_std_dev": 0.12728152373817306,
      "run_count": 9,
      "total_questions": 449,
      "avg_cost_per_task_usd": 0.049956171492204896,
      "avg_duration_per_task_ms": 149113.59020044544,
      "avg_output_tokens_per_task": 5756.837416481069,
      "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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Novita",
      "accuracy": 0.7008889333333334,
      "accuracy_std_dev": 0.03191351306362591,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.03097094533333333,
      "avg_duration_per_task_ms": 191571.11733333333,
      "avg_output_tokens_per_task": 4884.424,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Together",
      "accuracy": 0.7,
      "accuracy_std_dev": 0.006667000000000034,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.030492239999999997,
      "avg_duration_per_task_ms": 126354.09,
      "avg_output_tokens_per_task": 20020.43,
      "last_run_timestamp": "2026-09-19T08:14:06.252Z"
    },
    {
      "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": "Nebius",
      "accuracy": 0.7,
      "accuracy_std_dev": 0.03999999999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06787243,
      "avg_duration_per_task_ms": 42898.22,
      "avg_output_tokens_per_task": 4562.86,
      "last_run_timestamp": "2026-05-23T10:53:37.096Z"
    },
    {
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Phala",
      "accuracy": 0.7,
      "accuracy_std_dev": 0.020547857418718854,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.11523170499999999,
      "avg_duration_per_task_ms": 699435.85,
      "avg_output_tokens_per_task": 20321.33,
      "last_run_timestamp": "2026-09-19T08:14:58.361Z"
    },
    {
      "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-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-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "StreamLake",
      "accuracy": 0.6995833125,
      "accuracy_std_dev": 0.02938897155975083,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.033650589999999994,
      "avg_duration_per_task_ms": 154427.47,
      "avg_output_tokens_per_task": 6103.86125,
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6993176578947369,
      "accuracy_std_dev": 0.09243577928000629,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.02560508187134503,
      "avg_duration_per_task_ms": 1176961.7573099416,
      "avg_output_tokens_per_task": 38128.90935672515,
      "last_run_timestamp": "2026-09-12T16:00:20.981Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Google",
      "accuracy": 0.6975208464025869,
      "accuracy_std_dev": 0.06687846100100092,
      "run_count": 25,
      "total_questions": 1237,
      "avg_cost_per_task_usd": 0.09316261762328211,
      "avg_duration_per_task_ms": 251175.84963621665,
      "avg_output_tokens_per_task": 6799.790622473727,
      "last_run_timestamp": "2026-08-29T16:44:17.688Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Minimax",
      "accuracy": 0.696923076923077,
      "accuracy_std_dev": 0.07467836376443007,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.023498743076923078,
      "avg_duration_per_task_ms": 88729.55230769231,
      "avg_output_tokens_per_task": 6683.635384615384,
      "last_run_timestamp": "2026-09-19T16:00:28.638Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Parasail",
      "accuracy": 0.6958332500000001,
      "accuracy_std_dev": 0.010374963189693743,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.042198395,
      "avg_duration_per_task_ms": 172150.8225,
      "avg_output_tokens_per_task": 5272.2875,
      "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": "Morph",
      "accuracy": 0.6947618571428572,
      "accuracy_std_dev": 0.029726664088604592,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.03172098571428572,
      "avg_duration_per_task_ms": 126873.46571428572,
      "avg_output_tokens_per_task": 4470.107142857143,
      "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "auto-routing",
      "accuracy": 0.6941175882352941,
      "accuracy_std_dev": 0.04089728697036432,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.04082929647058823,
      "avg_duration_per_task_ms": 422759.3047058823,
      "avg_output_tokens_per_task": 12444.558823529413,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Chutes",
      "accuracy": 0.693333375,
      "accuracy_std_dev": 0.1814600292688017,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.042266376249999994,
      "avg_duration_per_task_ms": 100861.29625,
      "avg_output_tokens_per_task": 3413.32875,
      "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": "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "SiliconFlow",
      "accuracy": 0.692,
      "accuracy_std_dev": 0.04489988864128728,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.046083804,
      "avg_duration_per_task_ms": 295008.8,
      "avg_output_tokens_per_task": 3260.444,
      "last_run_timestamp": "2026-05-02T08:48:58.890Z"
    },
    {
      "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": 0.6915385,
      "accuracy_std_dev": 0.08956798464005299,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.044391347692307695,
      "avg_duration_per_task_ms": 193285.48538461537,
      "avg_output_tokens_per_task": 6191.080769230769,
      "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-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6907936666666666,
      "accuracy_std_dev": 0.03223436396490896,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.023469611428571423,
      "avg_duration_per_task_ms": 207909.9838095238,
      "avg_output_tokens_per_task": 4966.9914285714285,
      "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-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Z.AI",
      "accuracy": 0.6904283835341366,
      "accuracy_std_dev": 0.027558639980730167,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.012428607429718873,
      "avg_duration_per_task_ms": 150833.32128514056,
      "avg_output_tokens_per_task": 5172.266064257028,
      "last_run_timestamp": "2026-08-01T17:38:51.399Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "NextBit",
      "accuracy": 0.6901754210526316,
      "accuracy_std_dev": 0.039015354444811766,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.014771705263157893,
      "avg_duration_per_task_ms": 199050.82631578948,
      "avg_output_tokens_per_task": 12000.06,
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "auto-routing",
      "accuracy": 0.6901235111111111,
      "accuracy_std_dev": 0.05518567527960597,
      "run_count": 11,
      "total_questions": 540,
      "avg_cost_per_task_usd": 0.026533475925925927,
      "avg_duration_per_task_ms": 754731.6203703703,
      "avg_output_tokens_per_task": 34989.23888888889,
      "last_run_timestamp": "2026-09-19T16:00:28.638Z"
    },
    {
      "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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "AtlasCloud",
      "accuracy": 0.69,
      "accuracy_std_dev": 0.04013864859597431,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.04868145222222221,
      "avg_duration_per_task_ms": 105751.87222222223,
      "avg_output_tokens_per_task": 3340.511111111111,
      "last_run_timestamp": "2026-06-20T12:59:17.219Z"
    },
    {
      "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": "Io Net",
      "accuracy": 0.69,
      "accuracy_std_dev": 0.009999999999999953,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01987944,
      "avg_duration_per_task_ms": 47600.21,
      "avg_output_tokens_per_task": 2953.26,
      "last_run_timestamp": "2026-04-18T11:46:57.679Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "Xiaomi",
      "accuracy": 0.69,
      "accuracy_std_dev": 0.040532174168888865,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.005806647142857143,
      "avg_duration_per_task_ms": 77101.00428571428,
      "avg_output_tokens_per_task": 3994.6614285714286,
      "last_run_timestamp": "2026-06-13T08:42:56.762Z"
    },
    {
      "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": 0.6899998333333334,
      "accuracy_std_dev": 0.03415651881860261,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09539628666666666,
      "avg_duration_per_task_ms": 502649.19,
      "avg_output_tokens_per_task": 22095.93666666667,
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Inceptron",
      "accuracy": 0.6896024357798165,
      "accuracy_std_dev": 0.12893777217195382,
      "run_count": 9,
      "total_questions": 436,
      "avg_cost_per_task_usd": 0.05226660550458716,
      "avg_duration_per_task_ms": 99364.6376146789,
      "avg_output_tokens_per_task": 6918.672018348624,
      "last_run_timestamp": "2026-09-19T08:17:29.815Z"
    },
    {
      "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": 0.6895207458193979,
      "accuracy_std_dev": 0.06906980435656501,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.10185452675585285,
      "avg_duration_per_task_ms": 74719.04347826086,
      "avg_output_tokens_per_task": 7900.401337792642,
      "last_run_timestamp": "2026-09-19T08:15:19.093Z"
    },
    {
      "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": 0.689041575052854,
      "accuracy_std_dev": 0.03094305825454303,
      "run_count": 19,
      "total_questions": 946,
      "avg_cost_per_task_usd": 0.021175065539112053,
      "avg_duration_per_task_ms": 438783.39852008456,
      "avg_output_tokens_per_task": 11967.172304439746,
      "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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "auto-routing",
      "accuracy": 0.6882603594890511,
      "accuracy_std_dev": 0.11042330833399057,
      "run_count": 22,
      "total_questions": 1096,
      "avg_cost_per_task_usd": 0.04567830748175184,
      "avg_duration_per_task_ms": 105376.38777372263,
      "avg_output_tokens_per_task": 3958.9105839416056,
      "last_run_timestamp": "2026-08-01T17:38:27.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Mistral",
      "accuracy": 0.6878786363636364,
      "accuracy_std_dev": 0.02442357319280149,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.11701100909090908,
      "avg_duration_per_task_ms": 78384.59636363636,
      "avg_output_tokens_per_task": 9374.68909090909,
      "last_run_timestamp": "2026-09-17T02:10:06.326Z"
    },
    {
      "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": 0.6878611676300578,
      "accuracy_std_dev": 0.042049584814859244,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.013951661849710981,
      "avg_duration_per_task_ms": 210919.60693641618,
      "avg_output_tokens_per_task": 8494.329479768787,
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.6875626118355065,
      "accuracy_std_dev": 0.0373234995892668,
      "run_count": 20,
      "total_questions": 997,
      "avg_cost_per_task_usd": 0.03210182146439318,
      "avg_duration_per_task_ms": 179885.19558676027,
      "avg_output_tokens_per_task": 7259.881644934804,
      "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-5.6-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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Phala",
      "accuracy": 0.6863945836734694,
      "accuracy_std_dev": 0.10682812745363461,
      "run_count": 10,
      "total_questions": 490,
      "avg_cost_per_task_usd": 0.13472426326530615,
      "avg_duration_per_task_ms": 872956.7653061225,
      "avg_output_tokens_per_task": 33505.848979591836,
      "last_run_timestamp": "2026-09-19T16:00:28.641Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Venice",
      "accuracy": 0.685699512345679,
      "accuracy_std_dev": 0.13961279039785202,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.05491914969135803,
      "avg_duration_per_task_ms": 390942.3487654321,
      "avg_output_tokens_per_task": 6689.006172839506,
      "last_run_timestamp": "2026-09-19T08:21:42.831Z"
    },
    {
      "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": 0.6842424545454545,
      "accuracy_std_dev": 0.03293989573412772,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.07802852727272729,
      "avg_duration_per_task_ms": 647259.4745454545,
      "avg_output_tokens_per_task": 16952.354545454546,
      "last_run_timestamp": "2026-09-19T08:02:29.964Z"
    },
    {
      "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": "Io Net",
      "accuracy": 0.684,
      "accuracy_std_dev": 0.02939387691339811,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.011617532,
      "avg_duration_per_task_ms": 184984.468,
      "avg_output_tokens_per_task": 3721.176,
      "last_run_timestamp": "2026-06-13T09:50:28.856Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6836363636363636,
      "accuracy_std_dev": 0.09197555610128431,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.032429353636363636,
      "avg_duration_per_task_ms": 279572.37363636366,
      "avg_output_tokens_per_task": 9297.255454545455,
      "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": "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": 0.6826308620689655,
      "accuracy_std_dev": 0.039302124463935585,
      "run_count": 21,
      "total_questions": 1044,
      "avg_cost_per_task_usd": 0.01650072318007663,
      "avg_duration_per_task_ms": 286180.8256704981,
      "avg_output_tokens_per_task": 13045.386973180077,
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "AkashML",
      "accuracy": 0.6816424057971014,
      "accuracy_std_dev": 0.18410546186501534,
      "run_count": 7,
      "total_questions": 345,
      "avg_cost_per_task_usd": 0.10899141739130434,
      "avg_duration_per_task_ms": 177787.73623188405,
      "avg_output_tokens_per_task": 6552.942028985507,
      "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.6814285714285714,
      "accuracy_std_dev": 0.06300307734628557,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.005582074285714287,
      "avg_duration_per_task_ms": 43375.77857142857,
      "avg_output_tokens_per_task": 3572.1442857142856,
      "last_run_timestamp": "2026-06-13T09:16:25.647Z"
    },
    {
      "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": 0.6813061126126125,
      "accuracy_std_dev": 0.17362449557921197,
      "run_count": 9,
      "total_questions": 444,
      "avg_cost_per_task_usd": 0.04568808558558559,
      "avg_duration_per_task_ms": 182111.46846846846,
      "avg_output_tokens_per_task": 5794.326576576576,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Cloudflare",
      "accuracy": 0.6811110666666667,
      "accuracy_std_dev": 0.12911485116049026,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.1191761946666667,
      "avg_duration_per_task_ms": 130000.048,
      "avg_output_tokens_per_task": 7795.806666666666,
      "last_run_timestamp": "2026-09-19T08:15:16.660Z"
    },
    {
      "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": 0.6808227067307693,
      "accuracy_std_dev": 0.12201399623797919,
      "run_count": 25,
      "total_questions": 1248,
      "avg_cost_per_task_usd": 0.040376496794871784,
      "avg_duration_per_task_ms": 109025.02884615384,
      "avg_output_tokens_per_task": 5601.289262820513,
      "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6805555,
      "accuracy_std_dev": 0.16692092143412302,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.052043624999999996,
      "avg_duration_per_task_ms": 145763.23666666666,
      "avg_output_tokens_per_task": 17883.305,
      "last_run_timestamp": "2026-09-19T16:00:30.862Z"
    },
    {
      "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": 0.6803882349397591,
      "accuracy_std_dev": 0.08651381511485608,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.04140781124497992,
      "avg_duration_per_task_ms": 212525.218875502,
      "avg_output_tokens_per_task": 6209.691767068273,
      "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": "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Krea",
      "accuracy": 0.68,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.036475640000000004,
      "avg_duration_per_task_ms": 622815.32,
      "avg_output_tokens_per_task": 34740.02,
      "last_run_timestamp": "2026-09-16T09:00:02.758Z"
    },
    {
      "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": "Decart",
      "accuracy": 0.68,
      "accuracy_std_dev": 0.03999999999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.028420829999999998,
      "avg_duration_per_task_ms": 35643.83,
      "avg_output_tokens_per_task": 3017.33,
      "last_run_timestamp": "2026-06-20T09:16:42.673Z"
    },
    {
      "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": "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": 0.68,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.15835572,
      "avg_duration_per_task_ms": 627337.14,
      "avg_output_tokens_per_task": 4929.46,
      "last_run_timestamp": "2026-09-14T09:00:02.347Z"
    },
    {
      "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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "SambaNova",
      "accuracy": 0.6783625714285715,
      "accuracy_std_dev": 0.18503792379476175,
      "run_count": 24,
      "total_questions": 1197,
      "avg_cost_per_task_usd": 0.06683292147034252,
      "avg_duration_per_task_ms": 46030.891395154555,
      "avg_output_tokens_per_task": 5800.272347535505,
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.6783333,
      "accuracy_std_dev": 0.029354164951672524,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.049139350000000005,
      "avg_duration_per_task_ms": 440231.73,
      "avg_output_tokens_per_task": 6660.06,
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.6779184535864979,
      "accuracy_std_dev": 0.04493753781451566,
      "run_count": 19,
      "total_questions": 948,
      "avg_cost_per_task_usd": 0.05880673312236288,
      "avg_duration_per_task_ms": 487546.12974683545,
      "avg_output_tokens_per_task": 9480.43670886076,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.6778149866220735,
      "accuracy_std_dev": 0.13741656972108465,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.04846960535117057,
      "avg_duration_per_task_ms": 155359.52173913043,
      "avg_output_tokens_per_task": 7194.195652173913,
      "last_run_timestamp": "2026-09-19T08:15:15.826Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Cloudflare",
      "accuracy": 0.6766999163879599,
      "accuracy_std_dev": 0.167303531565863,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.06077755685618729,
      "avg_duration_per_task_ms": 163271.508361204,
      "avg_output_tokens_per_task": 6916.148829431439,
      "last_run_timestamp": "2026-09-19T08:15:01.043Z"
    },
    {
      "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": "Together",
      "accuracy": 0.6751515454545455,
      "accuracy_std_dev": 0.21232880724838227,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.044593925454545456,
      "avg_duration_per_task_ms": 85905.94363636363,
      "avg_output_tokens_per_task": 3735.2381818181816,
      "last_run_timestamp": "2026-06-30T19:07:40.534Z"
    },
    {
      "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": 0.6742424671717172,
      "accuracy_std_dev": 0.0476536543875113,
      "run_count": 16,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.017849546717171715,
      "avg_duration_per_task_ms": 164206.20707070708,
      "avg_output_tokens_per_task": 12386.181818181818,
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.6734920476190476,
      "accuracy_std_dev": 0.06836312161606557,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.024787600000000003,
      "avg_duration_per_task_ms": 231263.06666666668,
      "avg_output_tokens_per_task": 6580.758095238095,
      "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": "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": 0.6733836574307305,
      "accuracy_std_dev": 0.036202263205970635,
      "run_count": 16,
      "total_questions": 794,
      "avg_cost_per_task_usd": 0.011785144836272042,
      "avg_duration_per_task_ms": 140094.4420654912,
      "avg_output_tokens_per_task": 7374.278337531486,
      "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": "Baidu",
      "accuracy": 0.6733333333333333,
      "accuracy_std_dev": 0.0524933858267454,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.009767394999999998,
      "avg_duration_per_task_ms": 49065.985,
      "avg_output_tokens_per_task": 3511.735,
      "last_run_timestamp": "2026-06-06T17:11:08.802Z"
    },
    {
      "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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Google",
      "accuracy": 0.6731745714285714,
      "accuracy_std_dev": 0.04441834647612931,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.07005776761904764,
      "avg_duration_per_task_ms": 72619.98952380952,
      "avg_output_tokens_per_task": 4610.935238095238,
      "last_run_timestamp": "2026-08-01T17:38:46.618Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Minimax",
      "accuracy": 0.670833375,
      "accuracy_std_dev": 0.061131717422581666,
      "run_count": 40,
      "total_questions": 2000,
      "avg_cost_per_task_usd": 0.023635378499999995,
      "avg_duration_per_task_ms": 110968.634,
      "avg_output_tokens_per_task": 7070.978,
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Alibaba",
      "accuracy": 0.6700102376543211,
      "accuracy_std_dev": 0.10496553242878909,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.027610222222222223,
      "avg_duration_per_task_ms": 162024.1913580247,
      "avg_output_tokens_per_task": 7429.558641975309,
      "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": "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": "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": 0.6695237142857143,
      "accuracy_std_dev": 0.036663425955241954,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11802848000000002,
      "avg_duration_per_task_ms": 120456.28,
      "avg_output_tokens_per_task": 12635.617142857143,
      "last_run_timestamp": "2026-09-19T08:12:38.103Z"
    },
    {
      "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": 0.6692929393939393,
      "accuracy_std_dev": 0.05867076717192114,
      "run_count": 33,
      "total_questions": 1650,
      "avg_cost_per_task_usd": 0.01908582484848485,
      "avg_duration_per_task_ms": 102657.98787878788,
      "avg_output_tokens_per_task": 6209.364848484848,
      "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": "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": 0.6675946726057906,
      "accuracy_std_dev": 0.062223605171164056,
      "run_count": 18,
      "total_questions": 898,
      "avg_cost_per_task_usd": 0.02003099443207127,
      "avg_duration_per_task_ms": 278955.9331848552,
      "avg_output_tokens_per_task": 10682.610244988864,
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Together",
      "accuracy": 0.6672268756302522,
      "accuracy_std_dev": 0.19692376950802565,
      "run_count": 12,
      "total_questions": 595,
      "avg_cost_per_task_usd": 0.04049237815126051,
      "avg_duration_per_task_ms": 57680.32773109244,
      "avg_output_tokens_per_task": 4304.621848739495,
      "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": "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": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "BaseTen",
      "accuracy": 0.6655251643835617,
      "accuracy_std_dev": 0.050919510401766024,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.0655772808219178,
      "avg_duration_per_task_ms": 50910.90410958904,
      "avg_output_tokens_per_task": 3110.958904109589,
      "last_run_timestamp": "2026-09-12T08:11:26.919Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Phala",
      "accuracy": 0.6652707604690117,
      "accuracy_std_dev": 0.19745415809175107,
      "run_count": 12,
      "total_questions": 597,
      "avg_cost_per_task_usd": 0.09212694807370185,
      "avg_duration_per_task_ms": 267596.5443886097,
      "avg_output_tokens_per_task": 7550.497487437186,
      "last_run_timestamp": "2026-09-19T08:12:06.923Z"
    },
    {
      "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": 0.6646666,
      "accuracy_std_dev": 0.03190959382129456,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.12408255800000001,
      "avg_duration_per_task_ms": 116498.477,
      "avg_output_tokens_per_task": 12443.676,
      "last_run_timestamp": "2026-09-19T08:14:21.132Z"
    },
    {
      "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": 0.6644444166666666,
      "accuracy_std_dev": 0.036497949376091285,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.014842355833333333,
      "avg_duration_per_task_ms": 83587.3425,
      "avg_output_tokens_per_task": 5542.576666666667,
      "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": "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": "Z.AI",
      "accuracy": 0.6633333333333333,
      "accuracy_std_dev": 0.046785562825394,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.006290366666666667,
      "avg_duration_per_task_ms": 170099.27333333335,
      "avg_output_tokens_per_task": 5352.633333333333,
      "last_run_timestamp": "2026-05-09T21:59:21.056Z"
    },
    {
      "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": 0.6631578421052632,
      "accuracy_std_dev": 0.03436091497198395,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.016520328421052635,
      "avg_duration_per_task_ms": 162137.86210526316,
      "avg_output_tokens_per_task": 6763.676842105263,
      "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": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "Google",
      "accuracy": 0.6622221666666667,
      "accuracy_std_dev": 0.0198761101108564,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.31710583,
      "avg_duration_per_task_ms": 192864.10333333333,
      "avg_output_tokens_per_task": 8814.376666666667,
      "last_run_timestamp": "2026-09-01T13:04:20.137Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.6614287142857143,
      "accuracy_std_dev": 0.021518416940673253,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11292796,
      "avg_duration_per_task_ms": 122765.54,
      "avg_output_tokens_per_task": 12047.011428571428,
      "last_run_timestamp": "2026-09-19T08:19:00.786Z"
    },
    {
      "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": "Phala",
      "accuracy": 0.6611111111111111,
      "accuracy_std_dev": 0.02865288212939472,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.00995818888888889,
      "avg_duration_per_task_ms": 85091.12666666666,
      "avg_output_tokens_per_task": 4232.724444444444,
      "last_run_timestamp": "2026-06-20T21:23:09.799Z"
    },
    {
      "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": 0.6611110833333333,
      "accuracy_std_dev": 0.01760328077120064,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.03028098166666666,
      "avg_duration_per_task_ms": 310414.11,
      "avg_output_tokens_per_task": 9537.056666666667,
      "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": "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": 0.66,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0482247,
      "avg_duration_per_task_ms": 1006365.36,
      "avg_output_tokens_per_task": 46674.68,
      "last_run_timestamp": "2026-09-19T08:08:42.462Z"
    },
    {
      "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": 0.6573610416666666,
      "accuracy_std_dev": 0.035732853578650324,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.013916276666666665,
      "avg_duration_per_task_ms": 148914.81333333332,
      "avg_output_tokens_per_task": 6375.918333333333,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Parasail",
      "accuracy": 0.6573334,
      "accuracy_std_dev": 0.2126070154586626,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.04436226000000001,
      "avg_duration_per_task_ms": 85213.446,
      "avg_output_tokens_per_task": 3863.376,
      "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": "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": 0.6571529281183933,
      "accuracy_std_dev": 0.05360818932369472,
      "run_count": 19,
      "total_questions": 946,
      "avg_cost_per_task_usd": 0.02660617758985201,
      "avg_duration_per_task_ms": 237280.09619450316,
      "avg_output_tokens_per_task": 13635.822410147992,
      "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "CoreWeave",
      "accuracy": 0.6566665,
      "accuracy_std_dev": 0.016666499999999973,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.017312400000000002,
      "avg_duration_per_task_ms": 126237.63,
      "avg_output_tokens_per_task": 21774.52,
      "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": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6554808926174497,
      "accuracy_std_dev": 0.13732673855009286,
      "run_count": 6,
      "total_questions": 298,
      "avg_cost_per_task_usd": 0.04849104697986577,
      "avg_duration_per_task_ms": 2325789.067114094,
      "avg_output_tokens_per_task": 67477.9798657718,
      "last_run_timestamp": "2026-09-19T08:07:37.370Z"
    },
    {
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Novita",
      "accuracy": 0.6548717692307693,
      "accuracy_std_dev": 0.04499981444945309,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.01632250307692308,
      "avg_duration_per_task_ms": 92656.01,
      "avg_output_tokens_per_task": 6382.130769230769,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Inceptron",
      "accuracy": 0.6546666,
      "accuracy_std_dev": 0.14424683227800877,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.041911724,
      "avg_duration_per_task_ms": 194571.3,
      "avg_output_tokens_per_task": 4511.004,
      "last_run_timestamp": "2026-06-30T19:22:10.314Z"
    },
    {
      "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": 0.6542856190476191,
      "accuracy_std_dev": 0.030948730921709834,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.12287592761904761,
      "avg_duration_per_task_ms": 119817.67714285714,
      "avg_output_tokens_per_task": 12557.709523809524,
      "last_run_timestamp": "2026-09-19T08:16:41.224Z"
    },
    {
      "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": 0.652778,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.048472999999999995,
      "avg_duration_per_task_ms": 691836.3541666666,
      "avg_output_tokens_per_task": 18257.291666666668,
      "last_run_timestamp": "2026-08-21T21:00:03.652Z"
    },
    {
      "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": 0.6518518333333333,
      "accuracy_std_dev": 0.03388838939199286,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.026386046666666666,
      "avg_duration_per_task_ms": 174668.08444444445,
      "avg_output_tokens_per_task": 7625.071111111111,
      "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": "thinkingmachines/inkling-small-20260730",
      "model_slug": "thinkingmachines/inkling-small",
      "model_name": "Thinking Machines: Inkling Small",
      "provider_name": "Together",
      "accuracy": 0.65166675,
      "accuracy_std_dev": 0.04506177549972369,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.04474220500000001,
      "avg_duration_per_task_ms": 212693.005,
      "avg_output_tokens_per_task": 3236.785,
      "last_run_timestamp": "2026-09-12T08:12:16.067Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Friendli",
      "accuracy": 0.6512820769230769,
      "accuracy_std_dev": 0.0493641089572974,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.01929119692307693,
      "avg_duration_per_task_ms": 131140.2476923077,
      "avg_output_tokens_per_task": 6527.093076923077,
      "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": "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": 0.6506172962962963,
      "accuracy_std_dev": 0.04126635072919341,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.006890773333333334,
      "avg_duration_per_task_ms": 195510.75185185185,
      "avg_output_tokens_per_task": 7239.452592592593,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Together",
      "accuracy": 0.65,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.013433730000000001,
      "avg_duration_per_task_ms": 30166.39,
      "avg_output_tokens_per_task": 3575.99,
      "last_run_timestamp": "2026-04-25T08:11:30.451Z"
    },
    {
      "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": 0.6485714285714286,
      "accuracy_std_dev": 0.061434100526748275,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.01923216428571428,
      "avg_duration_per_task_ms": 162362.36428571428,
      "avg_output_tokens_per_task": 6530.884285714285,
      "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": "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.6466666818181818,
      "accuracy_std_dev": 0.11090082019181345,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.03871445727272728,
      "avg_duration_per_task_ms": 229660.09454545454,
      "avg_output_tokens_per_task": 6490.092727272728,
      "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": "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": 0.6460429376257546,
      "accuracy_std_dev": 0.1396968497913965,
      "run_count": 20,
      "total_questions": 994,
      "avg_cost_per_task_usd": 0.013013100603621729,
      "avg_duration_per_task_ms": 399798.5160965795,
      "avg_output_tokens_per_task": 11031.460764587526,
      "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": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Azure",
      "accuracy": 0.6457142857142857,
      "accuracy_std_dev": 0.02700476652484195,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.11825372571428569,
      "avg_duration_per_task_ms": 150221.76285714286,
      "avg_output_tokens_per_task": 12779.42857142857,
      "last_run_timestamp": "2026-09-19T08:12:02.575Z"
    },
    {
      "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": 0.644781292929293,
      "accuracy_std_dev": 0.015373999999999999,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.018105767676767675,
      "avg_duration_per_task_ms": 185321.31313131313,
      "avg_output_tokens_per_task": 22168.010101010103,
      "last_run_timestamp": "2026-09-19T08:02:40.719Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.6443191729323308,
      "accuracy_std_dev": 0.049327665162253266,
      "run_count": 16,
      "total_questions": 798,
      "avg_cost_per_task_usd": 0.01105938471177945,
      "avg_duration_per_task_ms": 153875.62656641603,
      "avg_output_tokens_per_task": 4867.625313283208,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.6438461923076922,
      "accuracy_std_dev": 0.04265441692704396,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.01644152,
      "avg_duration_per_task_ms": 177000.07846153845,
      "avg_output_tokens_per_task": 6952.727692307692,
      "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-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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "WandB",
      "accuracy": 0.6427566687763713,
      "accuracy_std_dev": 0.029051197700409103,
      "run_count": 19,
      "total_questions": 948,
      "avg_cost_per_task_usd": 0.01150348523206751,
      "avg_duration_per_task_ms": 156638.9229957806,
      "avg_output_tokens_per_task": 9733.639240506329,
      "last_run_timestamp": "2026-07-18T16:00:10.533Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "DeepInfra",
      "accuracy": 0.6422413563218391,
      "accuracy_std_dev": 0.06412982505652585,
      "run_count": 14,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.02831189511494253,
      "avg_duration_per_task_ms": 340448.8103448276,
      "avg_output_tokens_per_task": 10172.28735632184,
      "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-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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Novita",
      "accuracy": 0.641666625,
      "accuracy_std_dev": 0.03575897388047335,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.014602430000000001,
      "avg_duration_per_task_ms": 80982.3175,
      "avg_output_tokens_per_task": 5601.455833333333,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Phala",
      "accuracy": 0.6411111111111111,
      "accuracy_std_dev": 0.0300000329219984,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.015867753333333335,
      "avg_duration_per_task_ms": 224972.06444444443,
      "avg_output_tokens_per_task": 12173.745555555555,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "AkashML",
      "accuracy": 0.640784294117647,
      "accuracy_std_dev": 0.08323160877777759,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.01927779294117647,
      "avg_duration_per_task_ms": 109439.3894117647,
      "avg_output_tokens_per_task": 3724.7023529411763,
      "last_run_timestamp": "2026-07-01T02:04:30.298Z"
    },
    {
      "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": "SambaNova",
      "accuracy": 0.64,
      "accuracy_std_dev": 0.048989794855663564,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.029686422500000004,
      "avg_duration_per_task_ms": 51733.8325,
      "avg_output_tokens_per_task": 3684.0075,
      "last_run_timestamp": "2026-05-16T16:36:37.985Z"
    },
    {
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Venice",
      "accuracy": 0.63984269527897,
      "accuracy_std_dev": 0.15007041362089646,
      "run_count": 28,
      "total_questions": 1398,
      "avg_cost_per_task_usd": 0.045527532904148786,
      "avg_duration_per_task_ms": 180007.91201716737,
      "avg_output_tokens_per_task": 5421.762517882689,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "WandB",
      "accuracy": 0.6396296111111112,
      "accuracy_std_dev": 0.04605268692021117,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.034697361111111116,
      "avg_duration_per_task_ms": 95669.66222222222,
      "avg_output_tokens_per_task": 4599.92,
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Venice",
      "accuracy": 0.6387463811965812,
      "accuracy_std_dev": 0.07838394965714708,
      "run_count": 12,
      "total_questions": 585,
      "avg_cost_per_task_usd": 0.020067148717948718,
      "avg_duration_per_task_ms": 224214.09743589745,
      "avg_output_tokens_per_task": 22503.152136752138,
      "last_run_timestamp": "2026-09-19T16:00:28.608Z"
    },
    {
      "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": 0.6383547403846154,
      "accuracy_std_dev": 0.034423786918327294,
      "run_count": 25,
      "total_questions": 1248,
      "avg_cost_per_task_usd": 0.012904894230769234,
      "avg_duration_per_task_ms": 200720.20913461538,
      "avg_output_tokens_per_task": 13156.37580128205,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Mara",
      "accuracy": 0.637681,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 46,
      "avg_cost_per_task_usd": 0.04568002173913044,
      "avg_duration_per_task_ms": 70889.17391304347,
      "avg_output_tokens_per_task": 8168.934782608696,
      "last_run_timestamp": "2026-08-01T17:41:39.392Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6356409615384615,
      "accuracy_std_dev": 0.04264513255612286,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.01902188461538462,
      "avg_duration_per_task_ms": 148992.3046153846,
      "avg_output_tokens_per_task": 6265.466923076923,
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.6354166666666666,
      "accuracy_std_dev": 0.07353388359834909,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.060891364999999996,
      "avg_duration_per_task_ms": 149518.805,
      "avg_output_tokens_per_task": 6065.214166666667,
      "last_run_timestamp": "2026-08-22T16:12:20.461Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DeepInfra",
      "accuracy": 0.6339682380952381,
      "accuracy_std_dev": 0.05388220020687527,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.02847615904761905,
      "avg_duration_per_task_ms": 134848.23047619048,
      "avg_output_tokens_per_task": 4950.976190476191,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Mara",
      "accuracy": 0.6337365483870968,
      "accuracy_std_dev": 0.02579029640950258,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.016857699596774196,
      "avg_duration_per_task_ms": 45111.594758064515,
      "avg_output_tokens_per_task": 9635.451612903225,
      "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "DeepInfra",
      "accuracy": 0.6333335,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02418533,
      "avg_duration_per_task_ms": 152335.58,
      "avg_output_tokens_per_task": 37821.34,
      "last_run_timestamp": "2026-09-19T08:06:26.380Z"
    },
    {
      "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": "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Groq",
      "accuracy": 0.6278446465863453,
      "accuracy_std_dev": 0.02683498380174657,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.03220683132530121,
      "avg_duration_per_task_ms": 87364.718875502,
      "avg_output_tokens_per_task": 22117.32730923695,
      "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": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6251851666666667,
      "accuracy_std_dev": 0.04896739375935106,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.058607118888888886,
      "avg_duration_per_task_ms": 290047.4222222222,
      "avg_output_tokens_per_task": 13976.531111111111,
      "last_run_timestamp": "2026-07-01T03:04:05.088Z"
    },
    {
      "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": 0.6234782173913044,
      "accuracy_std_dev": 0.04417228647797023,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.008193006086956521,
      "avg_duration_per_task_ms": 187011.97043478262,
      "avg_output_tokens_per_task": 11024.112173913043,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DeepInfra",
      "accuracy": 0.6230798605990784,
      "accuracy_std_dev": 0.04010606346417775,
      "run_count": 35,
      "total_questions": 1736,
      "avg_cost_per_task_usd": 0.014455639976958522,
      "avg_duration_per_task_ms": 224122.75806451612,
      "avg_output_tokens_per_task": 14008.978110599079,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Venice",
      "accuracy": 0.6221364138996139,
      "accuracy_std_dev": 0.12792573087258097,
      "run_count": 26,
      "total_questions": 1295,
      "avg_cost_per_task_usd": 0.013064416216216216,
      "avg_duration_per_task_ms": 169022.18301158302,
      "avg_output_tokens_per_task": 5297.602316602317,
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Venice",
      "accuracy": 0.6214286071428571,
      "accuracy_std_dev": 0.14501615904180656,
      "run_count": 28,
      "total_questions": 1400,
      "avg_cost_per_task_usd": 0.028539994285714288,
      "avg_duration_per_task_ms": 185000.78214285715,
      "avg_output_tokens_per_task": 5080.563571428572,
      "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6203515192697769,
      "accuracy_std_dev": 0.2303888022849418,
      "run_count": 10,
      "total_questions": 493,
      "avg_cost_per_task_usd": 0.020992683569979714,
      "avg_duration_per_task_ms": 345295.43204868154,
      "avg_output_tokens_per_task": 11072.965517241379,
      "last_run_timestamp": "2026-09-19T08:18:35.767Z"
    },
    {
      "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": 0.620325766025641,
      "accuracy_std_dev": 0.05438397244850727,
      "run_count": 25,
      "total_questions": 1248,
      "avg_cost_per_task_usd": 0.03392773717948717,
      "avg_duration_per_task_ms": 185250.95272435897,
      "avg_output_tokens_per_task": 5975.309294871795,
      "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": "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6200679918367347,
      "accuracy_std_dev": 0.08681822122573118,
      "run_count": 10,
      "total_questions": 490,
      "avg_cost_per_task_usd": 0.017946885714285717,
      "avg_duration_per_task_ms": 578490.112244898,
      "avg_output_tokens_per_task": 21616.84081632653,
      "last_run_timestamp": "2026-09-19T16:00:30.834Z"
    },
    {
      "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": "Ambient",
      "accuracy": 0.62,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01353782,
      "avg_duration_per_task_ms": 89487.42,
      "avg_output_tokens_per_task": 4959.24,
      "last_run_timestamp": "2026-06-13T15:23:43.944Z"
    },
    {
      "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": 0.6191536436525612,
      "accuracy_std_dev": 0.050396043955175315,
      "run_count": 27,
      "total_questions": 1347,
      "avg_cost_per_task_usd": 0.007358628062360802,
      "avg_duration_per_task_ms": 148190.81885671863,
      "avg_output_tokens_per_task": 7990.572383073497,
      "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": "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": 0.6176811739130434,
      "accuracy_std_dev": 0.043985459303543334,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.05628466086956523,
      "avg_duration_per_task_ms": 389421.3947826087,
      "avg_output_tokens_per_task": 19986.14956521739,
      "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": "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": "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": 0.6175489270386266,
      "accuracy_std_dev": 0.06455018061760406,
      "run_count": 28,
      "total_questions": 1398,
      "avg_cost_per_task_usd": 0.008319118025751073,
      "avg_duration_per_task_ms": 160872.3583690987,
      "avg_output_tokens_per_task": 8106.090128755365,
      "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": "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Parasail",
      "accuracy": 0.6156657788369876,
      "accuracy_std_dev": 0.06881313004209644,
      "run_count": 21,
      "total_questions": 1049,
      "avg_cost_per_task_usd": 0.018457796949475698,
      "avg_duration_per_task_ms": 105384.89418493804,
      "avg_output_tokens_per_task": 8663.15061963775,
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "auto-routing",
      "accuracy": 0.6156620170015455,
      "accuracy_std_dev": 0.09743333851751483,
      "run_count": 26,
      "total_questions": 1294,
      "avg_cost_per_task_usd": 0.05918442503863987,
      "avg_duration_per_task_ms": 375639.1128284389,
      "avg_output_tokens_per_task": 9550.46367851623,
      "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-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "SambaNova",
      "accuracy": 0.6154540773993807,
      "accuracy_std_dev": 0.05444941500976228,
      "run_count": 26,
      "total_questions": 1292,
      "avg_cost_per_task_usd": 0.025835168730650157,
      "avg_duration_per_task_ms": 55520.85758513932,
      "avg_output_tokens_per_task": 14734.878482972135,
      "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": "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "BaseTen",
      "accuracy": 0.613333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.04676464,
      "avg_duration_per_task_ms": 28214.94,
      "avg_output_tokens_per_task": 3108.98,
      "last_run_timestamp": "2026-09-19T08:06:47.377Z"
    },
    {
      "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": 0.6113782756410256,
      "accuracy_std_dev": 0.036669727145404285,
      "run_count": 25,
      "total_questions": 1248,
      "avg_cost_per_task_usd": 0.06140813701923077,
      "avg_duration_per_task_ms": 418606.3149038461,
      "avg_output_tokens_per_task": 21705.047275641027,
      "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": "Novita",
      "accuracy": 0.6110183873121869,
      "accuracy_std_dev": 0.05300622363142776,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.09442565108514192,
      "avg_duration_per_task_ms": 376469.8146911519,
      "avg_output_tokens_per_task": 18202.715358931553,
      "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": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Fireworks",
      "accuracy": 0.608,
      "accuracy_std_dev": 0.25443270230062803,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.005877256,
      "avg_duration_per_task_ms": 51102.604,
      "avg_output_tokens_per_task": 3478.088,
      "last_run_timestamp": "2026-06-30T19:12:02.101Z"
    },
    {
      "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": 0.6077776666666667,
      "accuracy_std_dev": 0.08590813209793095,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.01715707,
      "avg_duration_per_task_ms": 128260.91,
      "avg_output_tokens_per_task": 13292.39,
      "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": "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Cerebras",
      "accuracy": 0.60599996,
      "accuracy_std_dev": 0.03504916217141859,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.038558521600000006,
      "avg_duration_per_task_ms": 20769.1408,
      "avg_output_tokens_per_task": 12079.4896,
      "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": "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "Novita",
      "accuracy": 0.6025,
      "accuracy_std_dev": 0.03799671038392668,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.005295742500000001,
      "avg_duration_per_task_ms": 41164.3,
      "avg_output_tokens_per_task": 1845.8825,
      "last_run_timestamp": "2026-05-23T08:43:48.681Z"
    },
    {
      "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": 0.6,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.26466022,
      "avg_duration_per_task_ms": 947997.3,
      "avg_output_tokens_per_task": 14069,
      "last_run_timestamp": "2026-07-31T00:04:29.459Z"
    },
    {
      "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": "Io Net",
      "accuracy": 0.6,
      "accuracy_std_dev": 0.03651483716701107,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.014077083333333332,
      "avg_duration_per_task_ms": 306993.37,
      "avg_output_tokens_per_task": 10503.493333333334,
      "last_run_timestamp": "2026-05-09T16:36:33.638Z"
    },
    {
      "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": 0.5985185555555556,
      "accuracy_std_dev": 0.06724027958681894,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.05260444666666667,
      "avg_duration_per_task_ms": 181957.32444444444,
      "avg_output_tokens_per_task": 8015.968888888889,
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "AtlasCloud",
      "accuracy": 0.5980392352941176,
      "accuracy_std_dev": 0.04037501463781108,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.02223544235294118,
      "avg_duration_per_task_ms": 43177.25411764706,
      "avg_output_tokens_per_task": 1978.5894117647058,
      "last_run_timestamp": "2026-07-01T01:59:48.107Z"
    },
    {
      "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": 0.5966666428571429,
      "accuracy_std_dev": 0.06755366708414358,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.04054190857142857,
      "avg_duration_per_task_ms": 231295.97714285715,
      "avg_output_tokens_per_task": 10843.084285714285,
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Venice",
      "accuracy": 0.5942841618589744,
      "accuracy_std_dev": 0.10277414844439432,
      "run_count": 25,
      "total_questions": 1248,
      "avg_cost_per_task_usd": 0.0771629423076923,
      "avg_duration_per_task_ms": 262057.74919871794,
      "avg_output_tokens_per_task": 6180.838141025641,
      "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-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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DigitalOcean",
      "accuracy": 0.5922862479338843,
      "accuracy_std_dev": 0.1002871639271946,
      "run_count": 5,
      "total_questions": 242,
      "avg_cost_per_task_usd": 0.5630070619834712,
      "avg_duration_per_task_ms": 3123571.8512396696,
      "avg_output_tokens_per_task": 96760.30578512397,
      "last_run_timestamp": "2026-09-19T08:12:55.047Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.5908772105263158,
      "accuracy_std_dev": 0.1441605041942301,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.009657277894736842,
      "avg_duration_per_task_ms": 71760.33368421052,
      "avg_output_tokens_per_task": 3688.9136842105263,
      "last_run_timestamp": "2026-07-01T02:00:24.123Z"
    },
    {
      "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": 0.5893471237113401,
      "accuracy_std_dev": 0.30515532350212654,
      "run_count": 4,
      "total_questions": 194,
      "avg_cost_per_task_usd": 0.02748909793814433,
      "avg_duration_per_task_ms": 98623.75773195876,
      "avg_output_tokens_per_task": 4244.2525773195875,
      "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": "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Parasail",
      "accuracy": 0.5861581787365177,
      "accuracy_std_dev": 0.08951304546845104,
      "run_count": 26,
      "total_questions": 1298,
      "avg_cost_per_task_usd": 0.013956122496147918,
      "avg_duration_per_task_ms": 66319.37981510015,
      "avg_output_tokens_per_task": 2672.7750385208014,
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "auto-routing",
      "accuracy": 0.581538423076923,
      "accuracy_std_dev": 0.06680269140718871,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.025179401538461547,
      "avg_duration_per_task_ms": 63238.23923076923,
      "avg_output_tokens_per_task": 2868.8115384615385,
      "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": "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": "openai/gpt-5.2-chat-20251211",
      "model_slug": "openai/gpt-5.2-chat",
      "model_name": "OpenAI: GPT-5.2 Chat",
      "provider_name": "OpenAI",
      "accuracy": 0.58,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.10846708,
      "avg_duration_per_task_ms": 79571.2,
      "avg_output_tokens_per_task": 4199.82,
      "last_run_timestamp": "2026-08-01T08:05:35.385Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Together",
      "accuracy": 0.58,
      "accuracy_std_dev": 0.03577708763999662,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.04891755600000001,
      "avg_duration_per_task_ms": 28136.316,
      "avg_output_tokens_per_task": 1804.596,
      "last_run_timestamp": "2026-05-09T08:27:19.378Z"
    },
    {
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "StreamLake",
      "accuracy": 0.5747916875,
      "accuracy_std_dev": 0.09687443552720627,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.033850981249999995,
      "avg_duration_per_task_ms": 62474.61875,
      "avg_output_tokens_per_task": 3318.425,
      "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": "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": 0.5716783164335665,
      "accuracy_std_dev": 0.05890284538551935,
      "run_count": 23,
      "total_questions": 1144,
      "avg_cost_per_task_usd": 0.09208451311188812,
      "avg_duration_per_task_ms": 689217.8505244756,
      "avg_output_tokens_per_task": 15762.438811188811,
      "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-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Mancer 2",
      "accuracy": 0.5712366451612902,
      "accuracy_std_dev": 0.0858999966086146,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.038580427419354836,
      "avg_duration_per_task_ms": 2634832.3427419355,
      "avg_output_tokens_per_task": 80935.72983870968,
      "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": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "AtlasCloud",
      "accuracy": 0.5706666666666667,
      "accuracy_std_dev": 0.07187643718369895,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.059964562666666665,
      "avg_duration_per_task_ms": 452906.82933333336,
      "avg_output_tokens_per_task": 11189.822666666667,
      "last_run_timestamp": "2026-06-20T22:23:15.613Z"
    },
    {
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "SiliconFlow",
      "accuracy": 0.5666666666666667,
      "accuracy_std_dev": 0.05120763831912405,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03156545333333333,
      "avg_duration_per_task_ms": 131519.01666666666,
      "avg_output_tokens_per_task": 1421.2433333333333,
      "last_run_timestamp": "2026-05-09T11:38:50.156Z"
    },
    {
      "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": 0.5658332500000001,
      "accuracy_std_dev": 0.1112399227646599,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.03004894333333333,
      "avg_duration_per_task_ms": 247199.82,
      "avg_output_tokens_per_task": 13306.571666666667,
      "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": "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": "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": 0.5615522976098689,
      "accuracy_std_dev": 0.13428503772778175,
      "run_count": 26,
      "total_questions": 1297,
      "avg_cost_per_task_usd": 0.0262945235158057,
      "avg_duration_per_task_ms": 147034.7632999229,
      "avg_output_tokens_per_task": 7801.311488049345,
      "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": "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "Novita",
      "accuracy": 0.560303,
      "accuracy_std_dev": 0.04800994961179683,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.12068492181818179,
      "avg_duration_per_task_ms": 185307.95727272728,
      "avg_output_tokens_per_task": 3636.26,
      "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": "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "auto-routing",
      "accuracy": 0.5531269180327869,
      "accuracy_std_dev": 0.05461221697062189,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.12032896812386154,
      "avg_duration_per_task_ms": 164673.07012750456,
      "avg_output_tokens_per_task": 3443.247723132969,
      "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": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "auto-routing",
      "accuracy": 0.5500352600422833,
      "accuracy_std_dev": 0.058938257867956145,
      "run_count": 19,
      "total_questions": 946,
      "avg_cost_per_task_usd": 0.0899786046511628,
      "avg_duration_per_task_ms": 885342.9714587738,
      "avg_output_tokens_per_task": 18576.52959830867,
      "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-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": "Nebius",
      "accuracy": 0.5488554020100502,
      "accuracy_std_dev": 0.08234461848628022,
      "run_count": 24,
      "total_questions": 1194,
      "avg_cost_per_task_usd": 0.07921593467336686,
      "avg_duration_per_task_ms": 188771.27135678392,
      "avg_output_tokens_per_task": 31712.71691792295,
      "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "AtlasCloud",
      "accuracy": 0.5379486923076923,
      "accuracy_std_dev": 0.2535341663450706,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.02374335076923077,
      "avg_duration_per_task_ms": 109221.52846153846,
      "avg_output_tokens_per_task": 2656.123076923077,
      "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": "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-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Ambient",
      "accuracy": 0.53,
      "accuracy_std_dev": 0.030000000000000027,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00720358,
      "avg_duration_per_task_ms": 62972.84,
      "avg_output_tokens_per_task": 2706.77,
      "last_run_timestamp": "2026-05-23T16:34:17.047Z"
    },
    {
      "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": "Fireworks",
      "accuracy": 0.53,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.004457689999999999,
      "avg_duration_per_task_ms": 29494.83,
      "avg_output_tokens_per_task": 2949.9,
      "last_run_timestamp": "2026-04-25T10:10:06.379Z"
    },
    {
      "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": 0.5213333333333333,
      "accuracy_std_dev": 0.019352465629876622,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.027455891999999996,
      "avg_duration_per_task_ms": 341646.812,
      "avg_output_tokens_per_task": 11097.765333333333,
      "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-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Clarifai",
      "accuracy": 0.52,
      "accuracy_std_dev": 0.1,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02067781,
      "avg_duration_per_task_ms": 133609.62,
      "avg_output_tokens_per_task": 82701.09,
      "last_run_timestamp": "2026-04-25T09:59:33.629Z"
    },
    {
      "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": 0.51916675,
      "accuracy_std_dev": 0.02591380209343084,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.025645805,
      "avg_duration_per_task_ms": 448484.4,
      "avg_output_tokens_per_task": 52443.9575,
      "last_run_timestamp": "2026-09-19T16:00:22.539Z"
    },
    {
      "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": 0.5151851666666667,
      "accuracy_std_dev": 0.028528172728900812,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.03173818,
      "avg_duration_per_task_ms": 738927.32,
      "avg_output_tokens_per_task": 116854.36666666667,
      "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": "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": 0.513371664756447,
      "accuracy_std_dev": 0.024300137045049023,
      "run_count": 14,
      "total_questions": 698,
      "avg_cost_per_task_usd": 0.05487210888252148,
      "avg_duration_per_task_ms": 254698.24212034384,
      "avg_output_tokens_per_task": 3318.022922636103,
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "auto-routing",
      "accuracy": 0.5126262474747475,
      "accuracy_std_dev": 0.02515381026771839,
      "run_count": 16,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.023066161616161616,
      "avg_duration_per_task_ms": 1118619.621212121,
      "avg_output_tokens_per_task": 98415.15909090909,
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Ionstream",
      "accuracy": 0.510644818430657,
      "accuracy_std_dev": 0.19786794978818611,
      "run_count": 22,
      "total_questions": 1096,
      "avg_cost_per_task_usd": 0.010844100364963506,
      "avg_duration_per_task_ms": 37477.74726277372,
      "avg_output_tokens_per_task": 1870.0903284671533,
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "NextBit",
      "accuracy": 0.5083333333333333,
      "accuracy_std_dev": 0.02229848026709944,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.007687528333333332,
      "avg_duration_per_task_ms": 35548.63166666667,
      "avg_output_tokens_per_task": 2391.6783333333333,
      "last_run_timestamp": "2026-06-20T16:41:40.557Z"
    },
    {
      "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": 0.5081871543859648,
      "accuracy_std_dev": 0.018863102843005027,
      "run_count": 23,
      "total_questions": 1140,
      "avg_cost_per_task_usd": 0.037433712280701754,
      "avg_duration_per_task_ms": 648881.6219298246,
      "avg_output_tokens_per_task": 187737.9605263158,
      "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": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Novita",
      "accuracy": 0.5054545454545455,
      "accuracy_std_dev": 0.02709029895591154,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.01571898,
      "avg_duration_per_task_ms": 333669.9781818182,
      "avg_output_tokens_per_task": 17186.79090909091,
      "last_run_timestamp": "2026-05-23T13:37:43.176Z"
    },
    {
      "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": 0.504816971098266,
      "accuracy_std_dev": 0.0288748445833789,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.024907991329479767,
      "avg_duration_per_task_ms": 563438.3439306358,
      "avg_output_tokens_per_task": 53889.92774566474,
      "last_run_timestamp": "2026-09-19T16:00:24.106Z"
    },
    {
      "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": 0.5025926111111111,
      "accuracy_std_dev": 0.05174193415740047,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.10888994777777779,
      "avg_duration_per_task_ms": 112476.33777777778,
      "avg_output_tokens_per_task": 1597.5555555555557,
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DeepInfra",
      "accuracy": 0.5017007653061224,
      "accuracy_std_dev": 0.02191038312649964,
      "run_count": 8,
      "total_questions": 392,
      "avg_cost_per_task_usd": 0.02296841326530612,
      "avg_duration_per_task_ms": 1353579.7653061224,
      "avg_output_tokens_per_task": 120488.71173469388,
      "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": "Ambient",
      "accuracy": 0.5,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00390972,
      "avg_duration_per_task_ms": 84786.05,
      "avg_output_tokens_per_task": 2818.58,
      "last_run_timestamp": "2026-06-13T17:26:55.519Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.49960782352941174,
      "accuracy_std_dev": 0.06175941664254676,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.015916549411764704,
      "avg_duration_per_task_ms": 33738.58705882353,
      "avg_output_tokens_per_task": 1680.0847058823529,
      "last_run_timestamp": "2026-06-30T19:02:38.080Z"
    },
    {
      "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": 0.49888883333333334,
      "accuracy_std_dev": 0.03680103829792791,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.026936693333333334,
      "avg_duration_per_task_ms": 564954.42,
      "avg_output_tokens_per_task": 52653.473333333335,
      "last_run_timestamp": "2026-09-19T16:00:24.106Z"
    },
    {
      "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": 0.49830913043478264,
      "accuracy_std_dev": 0.23149644601826339,
      "run_count": 14,
      "total_questions": 690,
      "avg_cost_per_task_usd": 0.005944484057971015,
      "avg_duration_per_task_ms": 247630.89855072464,
      "avg_output_tokens_per_task": 4340.623188405797,
      "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": "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DekaLLM",
      "accuracy": 0.49479164285714283,
      "accuracy_std_dev": 0.21842675350179366,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.0037404129464285716,
      "avg_duration_per_task_ms": 123800.87723214286,
      "avg_output_tokens_per_task": 3922.089285714286,
      "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": "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "SiliconFlow",
      "accuracy": 0.49308644444444444,
      "accuracy_std_dev": 0.05507886856091561,
      "run_count": 27,
      "total_questions": 1350,
      "avg_cost_per_task_usd": 0.04064708518518519,
      "avg_duration_per_task_ms": 188126.99925925926,
      "avg_output_tokens_per_task": 2360.5,
      "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": "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Parasail",
      "accuracy": 0.4864496585365854,
      "accuracy_std_dev": 0.04830717864914074,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.030671650406504065,
      "avg_duration_per_task_ms": 1720485.825203252,
      "avg_output_tokens_per_task": 131204.84552845528,
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Parasail",
      "accuracy": 0.4855528467336683,
      "accuracy_std_dev": 0.22212037609522364,
      "run_count": 16,
      "total_questions": 796,
      "avg_cost_per_task_usd": 0.011942556532663315,
      "avg_duration_per_task_ms": 259270.25502512563,
      "avg_output_tokens_per_task": 6411.776381909548,
      "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": "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": 0.485269303030303,
      "accuracy_std_dev": 0.023007442468416624,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.028894598484848485,
      "avg_duration_per_task_ms": 389374.452020202,
      "avg_output_tokens_per_task": 68395.02777777778,
      "last_run_timestamp": "2026-09-19T16:00:24.088Z"
    },
    {
      "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-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Ambient",
      "accuracy": 0.48,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00732523,
      "avg_duration_per_task_ms": 33542.46,
      "avg_output_tokens_per_task": 1578.75,
      "last_run_timestamp": "2026-06-13T08:36:56.093Z"
    },
    {
      "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": "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": 0.4775,
      "accuracy_std_dev": 0.03857702972301004,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.0438162675,
      "avg_duration_per_task_ms": 100713.26625,
      "avg_output_tokens_per_task": 2946.27125,
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Google",
      "accuracy": 0.4762204823232323,
      "accuracy_std_dev": 0.04091233889137237,
      "run_count": 16,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.09208659090909091,
      "avg_duration_per_task_ms": 53154.96338383838,
      "avg_output_tokens_per_task": 3752.4343434343436,
      "last_run_timestamp": "2026-08-15T14:15:58.674Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "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": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4756506221662469,
      "accuracy_std_dev": 0.009147616682220762,
      "run_count": 16,
      "total_questions": 794,
      "avg_cost_per_task_usd": 0.11002369899244333,
      "avg_duration_per_task_ms": 1736618.065491184,
      "avg_output_tokens_per_task": 28866.458438287154,
      "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "DeepInfra",
      "accuracy": 0.4749275652173913,
      "accuracy_std_dev": 0.036802997798296266,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.016968564347826086,
      "avg_duration_per_task_ms": 57427.46,
      "avg_output_tokens_per_task": 1869.2747826086957,
      "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": "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": 0.47472220833333334,
      "accuracy_std_dev": 0.052598362074449914,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.021700947499999998,
      "avg_duration_per_task_ms": 28847.52666666667,
      "avg_output_tokens_per_task": 2211.5291666666667,
      "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "Fireworks",
      "accuracy": 0.4744445,
      "accuracy_std_dev": 0.03557299398771489,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.023446433333333332,
      "avg_duration_per_task_ms": 1018917.68,
      "avg_output_tokens_per_task": 62981.19,
      "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-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": 0.47376544907407403,
      "accuracy_std_dev": 0.0497035553280053,
      "run_count": 26,
      "total_questions": 1296,
      "avg_cost_per_task_usd": 0.014925031635802471,
      "avg_duration_per_task_ms": 36653.99614197531,
      "avg_output_tokens_per_task": 2309.6319444444443,
      "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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4736589626436782,
      "accuracy_std_dev": 0.0308370259740376,
      "run_count": 14,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.029854971264367817,
      "avg_duration_per_task_ms": 1270502.7399425288,
      "avg_output_tokens_per_task": 22673.870689655174,
      "last_run_timestamp": "2026-08-15T14:12:30.435Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.473333,
      "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": "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": "Cerebras",
      "accuracy": 0.472,
      "accuracy_std_dev": 0.048332183894378274,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.05911650399999999,
      "avg_duration_per_task_ms": 16473.438,
      "avg_output_tokens_per_task": 1544.504,
      "last_run_timestamp": "2026-05-23T08:11:43.884Z"
    },
    {
      "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": 0.4710981791907515,
      "accuracy_std_dev": 0.2776117089494547,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.01732042774566474,
      "avg_duration_per_task_ms": 101207.72832369943,
      "avg_output_tokens_per_task": 18057.771676300577,
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Phala",
      "accuracy": 0.47,
      "accuracy_std_dev": 0.03,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06848815,
      "avg_duration_per_task_ms": 119560.68,
      "avg_output_tokens_per_task": 4089.64,
      "last_run_timestamp": "2026-04-25T09:53:01.657Z"
    },
    {
      "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": 0.46805773261205563,
      "accuracy_std_dev": 0.038430249999417805,
      "run_count": 26,
      "total_questions": 1294,
      "avg_cost_per_task_usd": 0.01566730989180835,
      "avg_duration_per_task_ms": 431720.4644513138,
      "avg_output_tokens_per_task": 13130.05486862442,
      "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Together",
      "accuracy": 0.468,
      "accuracy_std_dev": 0.020396078054371138,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.17626782199999996,
      "avg_duration_per_task_ms": 36863.33,
      "avg_output_tokens_per_task": 1283.172,
      "last_run_timestamp": "2026-05-23T08:17:34.003Z"
    },
    {
      "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": 0.46799996,
      "accuracy_std_dev": 0.03661208446562965,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.0154081888,
      "avg_duration_per_task_ms": 64765.2344,
      "avg_output_tokens_per_task": 2351.1312,
      "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": "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": 0.4675,
      "accuracy_std_dev": 0.009682458365518523,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0121487525,
      "avg_duration_per_task_ms": 77116.52,
      "avg_output_tokens_per_task": 1985.6575,
      "last_run_timestamp": "2026-05-16T08:26:24.421Z"
    },
    {
      "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": 0.46717946153846157,
      "accuracy_std_dev": 0.15927102080132052,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.02094849153846154,
      "avg_duration_per_task_ms": 36856.35692307692,
      "avg_output_tokens_per_task": 1813.543076923077,
      "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-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Together",
      "accuracy": 0.4660245762439807,
      "accuracy_std_dev": 0.04158627548980073,
      "run_count": 25,
      "total_questions": 1246,
      "avg_cost_per_task_usd": 0.016147356340288924,
      "avg_duration_per_task_ms": 152395.11235955055,
      "avg_output_tokens_per_task": 11304.92455858748,
      "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-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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "auto-routing",
      "accuracy": 0.46319996,
      "accuracy_std_dev": 0.0514715031235576,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.05008044400000001,
      "avg_duration_per_task_ms": 66373.1232,
      "avg_output_tokens_per_task": 1925.0056,
      "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": "DeepInfra",
      "accuracy": 0.46244856790123456,
      "accuracy_std_dev": 0.03790693205554764,
      "run_count": 26,
      "total_questions": 1296,
      "avg_cost_per_task_usd": 0.015229477623456784,
      "avg_duration_per_task_ms": 96153.46373456791,
      "avg_output_tokens_per_task": 2336.3819444444443,
      "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Google",
      "accuracy": 0.46140347368421053,
      "accuracy_std_dev": 0.04392064180142758,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.028559272631578946,
      "avg_duration_per_task_ms": 26772.692631578946,
      "avg_output_tokens_per_task": 1583.818947368421,
      "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": "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": "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": 0.46,
      "accuracy_std_dev": 0.042365957548484615,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.01855554615384615,
      "avg_duration_per_task_ms": 109691.66846153846,
      "avg_output_tokens_per_task": 2329.1007692307694,
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "SambaNova",
      "accuracy": 0.4585964736842105,
      "accuracy_std_dev": 0.11550210320544337,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.08024197263157894,
      "avg_duration_per_task_ms": 119560.23052631579,
      "avg_output_tokens_per_task": 7054.113684210526,
      "last_run_timestamp": "2026-07-04T16:00:25.000Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.45777799999999996,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.00622334,
      "avg_duration_per_task_ms": 39170.17333333333,
      "avg_output_tokens_per_task": 1174.36,
      "last_run_timestamp": "2026-08-15T08:16:13.926Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Google",
      "accuracy": 0.4570513076923077,
      "accuracy_std_dev": 0.04206556644533465,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.03599279999999999,
      "avg_duration_per_task_ms": 44357.65923076923,
      "avg_output_tokens_per_task": 2306.9653846153847,
      "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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Alibaba",
      "accuracy": 0.45699995,
      "accuracy_std_dev": 0.039635830171039686,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.016694145000000007,
      "avg_duration_per_task_ms": 191163.145,
      "avg_output_tokens_per_task": 15722.233,
      "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": "AtlasCloud",
      "accuracy": 0.45686276470588233,
      "accuracy_std_dev": 0.05780818724787877,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.08079224352941176,
      "avg_duration_per_task_ms": 50778.730588235296,
      "avg_output_tokens_per_task": 1461.3635294117646,
      "last_run_timestamp": "2026-06-30T19:13:39.299Z"
    },
    {
      "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": 0.4561111666666667,
      "accuracy_std_dev": 0.041137423636783854,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.01247452916666667,
      "avg_duration_per_task_ms": 49982.72,
      "avg_output_tokens_per_task": 2226.931666666667,
      "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": "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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.45450982352941177,
      "accuracy_std_dev": 0.03875415290723167,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.01727678352941177,
      "avg_duration_per_task_ms": 273100.5176470588,
      "avg_output_tokens_per_task": 7362.865882352941,
      "last_run_timestamp": "2026-07-01T02:31:49.051Z"
    },
    {
      "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": 0.45413336,
      "accuracy_std_dev": 0.03916098911506705,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.17327989679999997,
      "avg_duration_per_task_ms": 59985.74,
      "avg_output_tokens_per_task": 1913.856,
      "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": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4533333333333333,
      "accuracy_std_dev": 0.03771236166328253,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.01162648,
      "avg_duration_per_task_ms": 116752.73333333334,
      "avg_output_tokens_per_task": 1542.3466666666666,
      "last_run_timestamp": "2026-05-09T11:02:11.791Z"
    },
    {
      "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": 0.45200004,
      "accuracy_std_dev": 0.04410092989720739,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.06653196960000002,
      "avg_duration_per_task_ms": 66418.1632,
      "avg_output_tokens_per_task": 1909.792,
      "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": "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": 0.450416625,
      "accuracy_std_dev": 0.05038566298794107,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.026321354999999994,
      "avg_duration_per_task_ms": 84986.70625,
      "avg_output_tokens_per_task": 1810.525,
      "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-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": 0.44898461069836554,
      "accuracy_std_dev": 0.07415219664337017,
      "run_count": 27,
      "total_questions": 1346,
      "avg_cost_per_task_usd": 0.020572284546805354,
      "avg_duration_per_task_ms": 49604.04457652303,
      "avg_output_tokens_per_task": 2003.719910846954,
      "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": "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.448811634369287,
      "accuracy_std_dev": 0.059025095658296695,
      "run_count": 11,
      "total_questions": 547,
      "avg_cost_per_task_usd": 0.030702687385740404,
      "avg_duration_per_task_ms": 605640.2120658135,
      "avg_output_tokens_per_task": 6772.7239488117,
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "WandB",
      "accuracy": 0.44875,
      "accuracy_std_dev": 0.04580870550452174,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.011756012500000001,
      "avg_duration_per_task_ms": 28247.49,
      "avg_output_tokens_per_task": 1531.14625,
      "last_run_timestamp": "2026-06-20T16:39:11.919Z"
    },
    {
      "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": "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": 0.44642858482142855,
      "accuracy_std_dev": 0.022519182833890147,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.04059176116071428,
      "avg_duration_per_task_ms": 108020.11607142857,
      "avg_output_tokens_per_task": 1939.064732142857,
      "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": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.4452632105263158,
      "accuracy_std_dev": 0.057249987520641206,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.01557863684210526,
      "avg_duration_per_task_ms": 192621.90315789473,
      "avg_output_tokens_per_task": 10602.088421052631,
      "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": "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": "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": 0.44375,
      "accuracy_std_dev": 0.04539754949333719,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.021965795,
      "avg_duration_per_task_ms": 36665.47375,
      "avg_output_tokens_per_task": 1608.6575,
      "last_run_timestamp": "2026-06-20T17:18:35.419Z"
    },
    {
      "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": "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": "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": 0.4389744230769231,
      "accuracy_std_dev": 0.05707069678320921,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.03386180076923076,
      "avg_duration_per_task_ms": 88270.66923076923,
      "avg_output_tokens_per_task": 2280.8869230769233,
      "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/qwen3.6-35b-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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Venice",
      "accuracy": 0.43711742237061774,
      "accuracy_std_dev": 0.06180167813928354,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.010960462437395655,
      "avg_duration_per_task_ms": 96791.42904841402,
      "avg_output_tokens_per_task": 1832.7679465776293,
      "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": "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": "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": 0.43621399845679015,
      "accuracy_std_dev": 0.04684420405366707,
      "run_count": 26,
      "total_questions": 1296,
      "avg_cost_per_task_usd": 0.015120972993827161,
      "avg_duration_per_task_ms": 167672.98225308643,
      "avg_output_tokens_per_task": 2265.6689814814813,
      "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": "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": "Cerebras",
      "accuracy": 0.436,
      "accuracy_std_dev": 0.026532998322843202,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.012890084,
      "avg_duration_per_task_ms": 17630.374,
      "avg_output_tokens_per_task": 1208.354,
      "last_run_timestamp": "2026-05-23T08:10:41.406Z"
    },
    {
      "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": 0.43537074448897795,
      "accuracy_std_dev": 0.11336612641816779,
      "run_count": 20,
      "total_questions": 998,
      "avg_cost_per_task_usd": 0.009285377755511023,
      "avg_duration_per_task_ms": 54460.993987975955,
      "avg_output_tokens_per_task": 1751.9448897795592,
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4346106916058394,
      "accuracy_std_dev": 0.024005000661578763,
      "run_count": 22,
      "total_questions": 1096,
      "avg_cost_per_task_usd": 0.07036007481751827,
      "avg_duration_per_task_ms": 656963.5930656934,
      "avg_output_tokens_per_task": 12332.072080291971,
      "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": "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": 0.43401008121827417,
      "accuracy_std_dev": 0.029416728833002063,
      "run_count": 8,
      "total_questions": 394,
      "avg_cost_per_task_usd": 0.035802939086294415,
      "avg_duration_per_task_ms": 50594.96446700508,
      "avg_output_tokens_per_task": 1953.4467005076142,
      "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": "Alibaba",
      "accuracy": 0.43333334615384617,
      "accuracy_std_dev": 0.04355954957291871,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.025601543076923073,
      "avg_duration_per_task_ms": 51646.70153846154,
      "avg_output_tokens_per_task": 2288.9223076923076,
      "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": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "DeepInfra",
      "accuracy": 0.43215811057692305,
      "accuracy_std_dev": 0.0625616725008659,
      "run_count": 25,
      "total_questions": 1248,
      "avg_cost_per_task_usd": 0.01311161698717949,
      "avg_duration_per_task_ms": 240895.54807692306,
      "avg_output_tokens_per_task": 12017.117788461539,
      "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": "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "Novita",
      "accuracy": 0.43,
      "accuracy_std_dev": 0.035355339059327376,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.02250267,
      "avg_duration_per_task_ms": 103645.2525,
      "avg_output_tokens_per_task": 2220.14375,
      "last_run_timestamp": "2026-06-20T17:24:59.951Z"
    },
    {
      "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": 0.42910253846153845,
      "accuracy_std_dev": 0.049782059536644926,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.016837386923076923,
      "avg_duration_per_task_ms": 118019.43153846153,
      "avg_output_tokens_per_task": 2356.7115384615386,
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "auto-routing",
      "accuracy": 0.42883213321167885,
      "accuracy_std_dev": 0.038713233772523226,
      "run_count": 22,
      "total_questions": 1096,
      "avg_cost_per_task_usd": 0.04187416332116788,
      "avg_duration_per_task_ms": 349556.7518248175,
      "avg_output_tokens_per_task": 8660.958941605839,
      "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": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.4281170687022901,
      "accuracy_std_dev": 0.06363488770219329,
      "run_count": 21,
      "total_questions": 1048,
      "avg_cost_per_task_usd": 0.019827515267175574,
      "avg_duration_per_task_ms": 147943.55916030533,
      "avg_output_tokens_per_task": 10525.584923664122,
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "DeepInfra",
      "accuracy": 0.4234779606458123,
      "accuracy_std_dev": 0.039658085032279355,
      "run_count": 20,
      "total_questions": 991,
      "avg_cost_per_task_usd": 0.019681630676084763,
      "avg_duration_per_task_ms": 234062.66094853682,
      "avg_output_tokens_per_task": 2454.673057517659,
      "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/qwen3-14b-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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "auto-routing",
      "accuracy": 0.42064714021571653,
      "accuracy_std_dev": 0.10279858751976688,
      "run_count": 26,
      "total_questions": 1298,
      "avg_cost_per_task_usd": 0.015630807395993835,
      "avg_duration_per_task_ms": 485399.4352850539,
      "avg_output_tokens_per_task": 13574.13405238829,
      "last_run_timestamp": "2026-08-29T16:16:50.456Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.41933335,
      "accuracy_std_dev": 0.045161409534330295,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.016045132,
      "avg_duration_per_task_ms": 211276.496,
      "avg_output_tokens_per_task": 10822.63,
      "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": "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-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.41333329166666666,
      "accuracy_std_dev": 0.05206835251257009,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.025643208333333337,
      "avg_duration_per_task_ms": 118146.86333333333,
      "avg_output_tokens_per_task": 2518.2691666666665,
      "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/qwen3-14b-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": "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "WandB",
      "accuracy": 0.4103852261306533,
      "accuracy_std_dev": 0.0554530731629442,
      "run_count": 16,
      "total_questions": 796,
      "avg_cost_per_task_usd": 0.20432325502512563,
      "avg_duration_per_task_ms": 305226.5891959799,
      "avg_output_tokens_per_task": 15810.201005025127,
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "AtlasCloud",
      "accuracy": 0.4086021532258065,
      "accuracy_std_dev": 0.04283406661478688,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.771351502016129,
      "avg_duration_per_task_ms": 371618.85282258067,
      "avg_output_tokens_per_task": 14652.147177419354,
      "last_run_timestamp": "2026-07-01T04:28:59.565Z"
    },
    {
      "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": 0.4074358846153846,
      "accuracy_std_dev": 0.045038470881379175,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.03544446692307693,
      "avg_duration_per_task_ms": 143373.31461538462,
      "avg_output_tokens_per_task": 3607.816923076923,
      "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": "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": 0.4072934727592267,
      "accuracy_std_dev": 0.021976352713623176,
      "run_count": 23,
      "total_questions": 1138,
      "avg_cost_per_task_usd": 0.006092479789103691,
      "avg_duration_per_task_ms": 18758.165202108965,
      "avg_output_tokens_per_task": 321.83391915641477,
      "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": "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": "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": "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": 0.40160004,
      "accuracy_std_dev": 0.03598909780639687,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.017235716,
      "avg_duration_per_task_ms": 165323.3096,
      "avg_output_tokens_per_task": 2670.9176,
      "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": "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": 0.4015928839590444,
      "accuracy_std_dev": 0.04374660891326003,
      "run_count": 12,
      "total_questions": 586,
      "avg_cost_per_task_usd": 0.05105657679180887,
      "avg_duration_per_task_ms": 176971.67918088738,
      "avg_output_tokens_per_task": 3873.042662116041,
      "last_run_timestamp": "2026-09-19T08:13:57.688Z"
    },
    {
      "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": "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": 0.3997436153846154,
      "accuracy_std_dev": 0.01995555515307593,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.009033361538461534,
      "avg_duration_per_task_ms": 40294.35923076923,
      "avg_output_tokens_per_task": 386.8792307692308,
      "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Venice",
      "accuracy": 0.39916650000000004,
      "accuracy_std_dev": 0.32402718893134574,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0238333725,
      "avg_duration_per_task_ms": 73030.4175,
      "avg_output_tokens_per_task": 12951.32,
      "last_run_timestamp": "2026-09-19T16:00:30.841Z"
    },
    {
      "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": "Inceptron",
      "accuracy": 0.3975,
      "accuracy_std_dev": 0.029047375096555618,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.005109264999999999,
      "avg_duration_per_task_ms": 12326.42,
      "avg_output_tokens_per_task": 249.89,
      "last_run_timestamp": "2026-05-16T08:28:13.794Z"
    },
    {
      "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": 0.3953846153846154,
      "accuracy_std_dev": 0.02553822881196092,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.008089532307692306,
      "avg_duration_per_task_ms": 21213.126153846155,
      "avg_output_tokens_per_task": 369.43384615384616,
      "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": "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-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.3894871153846154,
      "accuracy_std_dev": 0.03518801188159107,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.047560097692307686,
      "avg_duration_per_task_ms": 76824.58384615385,
      "avg_output_tokens_per_task": 1990.8353846153846,
      "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": "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": 0.389445249614792,
      "accuracy_std_dev": 0.0657706643988149,
      "run_count": 26,
      "total_questions": 1298,
      "avg_cost_per_task_usd": 0.035619542372881355,
      "avg_duration_per_task_ms": 164533.79506933744,
      "avg_output_tokens_per_task": 2309.8736517719567,
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "AtlasCloud",
      "accuracy": 0.3875,
      "accuracy_std_dev": 0.035266839949164706,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.01950373,
      "avg_duration_per_task_ms": 100582.8275,
      "avg_output_tokens_per_task": 2448.57875,
      "last_run_timestamp": "2026-06-20T16:46:21.602Z"
    },
    {
      "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": 0.3852564230769231,
      "accuracy_std_dev": 0.0754143744899182,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.038005519230769245,
      "avg_duration_per_task_ms": 57650.9,
      "avg_output_tokens_per_task": 1802.6207692307692,
      "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": "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": 0.3848532040185471,
      "accuracy_std_dev": 0.034396038458699964,
      "run_count": 26,
      "total_questions": 1294,
      "avg_cost_per_task_usd": 0.012941758887171563,
      "avg_duration_per_task_ms": 45655.90262751159,
      "avg_output_tokens_per_task": 569.8686244204018,
      "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/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "AtlasCloud",
      "accuracy": 0.3839215882352941,
      "accuracy_std_dev": 0.05204618331834641,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.01229372588235294,
      "avg_duration_per_task_ms": 64996.49411764706,
      "avg_output_tokens_per_task": 1870.1082352941175,
      "last_run_timestamp": "2026-06-30T19:11:05.329Z"
    },
    {
      "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": 0.38385963157894737,
      "accuracy_std_dev": 0.04338409272291609,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.01576738736842105,
      "avg_duration_per_task_ms": 37820.52210526316,
      "avg_output_tokens_per_task": 2258.8694736842103,
      "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": "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": "StreamLake",
      "accuracy": 0.3822342530282638,
      "accuracy_std_dev": 0.06305786163883313,
      "run_count": 15,
      "total_questions": 743,
      "avg_cost_per_task_usd": 0.009679043068640645,
      "avg_duration_per_task_ms": 96270.86137281293,
      "avg_output_tokens_per_task": 2609.127860026918,
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Venice",
      "accuracy": 0.3822221666666667,
      "accuracy_std_dev": 0.15283338826307627,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.030608421666666673,
      "avg_duration_per_task_ms": 214180.72166666668,
      "avg_output_tokens_per_task": 2714.745,
      "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": "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": 0.38166665,
      "accuracy_std_dev": 0.02484398541151358,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.038916812999999995,
      "avg_duration_per_task_ms": 13463.68,
      "avg_output_tokens_per_task": 499.736,
      "last_run_timestamp": "2026-07-18T16:00:09.280Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.38106668,
      "accuracy_std_dev": 0.032258962827369383,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.0164048448,
      "avg_duration_per_task_ms": 118964.3272,
      "avg_output_tokens_per_task": 2708.0272,
      "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": "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": 0.3793939090909091,
      "accuracy_std_dev": 0.05088740122422613,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.039116065454545455,
      "avg_duration_per_task_ms": 58462.97090909091,
      "avg_output_tokens_per_task": 3651.5763636363636,
      "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": "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-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Venice",
      "accuracy": 0.3725,
      "accuracy_std_dev": 0.09312108008170508,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.040488017499999994,
      "avg_duration_per_task_ms": 139617.32,
      "avg_output_tokens_per_task": 4754.3475,
      "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/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Venice",
      "accuracy": 0.368,
      "accuracy_std_dev": 0.04489988864128728,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.04196784600000001,
      "avg_duration_per_task_ms": 61894.268,
      "avg_output_tokens_per_task": 2505.452,
      "last_run_timestamp": "2026-05-23T09:30:59.805Z"
    },
    {
      "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": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.36075754545454547,
      "accuracy_std_dev": 0.053009762429135535,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.019241303636363637,
      "avg_duration_per_task_ms": 42582.33090909091,
      "avg_output_tokens_per_task": 2226.5054545454545,
      "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": "DeepInfra",
      "accuracy": 0.35989578125,
      "accuracy_std_dev": 0.027985955459943063,
      "run_count": 13,
      "total_questions": 640,
      "avg_cost_per_task_usd": 0.06471081249999999,
      "avg_duration_per_task_ms": 330465.9046875,
      "avg_output_tokens_per_task": 6852.8765625,
      "last_run_timestamp": "2026-09-19T08:08:41.469Z"
    },
    {
      "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": 0.34972216666666667,
      "accuracy_std_dev": 0.043118179904833125,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.023779785,
      "avg_duration_per_task_ms": 75866.685,
      "avg_output_tokens_per_task": 2981.369166666667,
      "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": "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": 0.34503904682274245,
      "accuracy_std_dev": 0.03117646397380117,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.07262195986622073,
      "avg_duration_per_task_ms": 212425.83779264215,
      "avg_output_tokens_per_task": 6685.140468227425,
      "last_run_timestamp": "2026-09-19T08:12:53.333Z"
    },
    {
      "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": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.33351874638487206,
      "accuracy_std_dev": 0.08113788331325603,
      "run_count": 18,
      "total_questions": 899,
      "avg_cost_per_task_usd": 0.02762981312569522,
      "avg_duration_per_task_ms": 64964.9043381535,
      "avg_output_tokens_per_task": 3953.5528364849833,
      "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": 0.3266667692307692,
      "accuracy_std_dev": 0.04280749195598978,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.07343648153846152,
      "avg_duration_per_task_ms": 129206.12923076923,
      "avg_output_tokens_per_task": 6940.74,
      "last_run_timestamp": "2026-09-19T08:13:24.163Z"
    },
    {
      "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": 0.3263781514683153,
      "accuracy_std_dev": 0.11733471265135623,
      "run_count": 26,
      "total_questions": 1294,
      "avg_cost_per_task_usd": 0.09428212982998455,
      "avg_duration_per_task_ms": 283568.8972179289,
      "avg_output_tokens_per_task": 5857.006182380216,
      "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": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.3262745294117647,
      "accuracy_std_dev": 0.06326493483766375,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.010698870588235293,
      "avg_duration_per_task_ms": 44460.25529411765,
      "avg_output_tokens_per_task": 1664.0011764705882,
      "last_run_timestamp": "2026-07-01T01:59:56.255Z"
    },
    {
      "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": 0.3262120909090909,
      "accuracy_std_dev": 0.04420222727719117,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.019649248181818187,
      "avg_duration_per_task_ms": 67216.11454545455,
      "avg_output_tokens_per_task": 2692.0563636363636,
      "last_run_timestamp": "2026-08-01T17:38:36.457Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "SiliconFlow",
      "accuracy": 0.31547627678571427,
      "accuracy_std_dev": 0.18940222003098767,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.11378589732142856,
      "avg_duration_per_task_ms": 450016.14285714284,
      "avg_output_tokens_per_task": 4166.011160714285,
      "last_run_timestamp": "2026-09-19T08:15:10.642Z"
    },
    {
      "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": 0.3133333,
      "accuracy_std_dev": 0.067462010709065,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.006098155,
      "avg_duration_per_task_ms": 65881.432,
      "avg_output_tokens_per_task": 1526.893,
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Fireworks",
      "accuracy": 0.29714285714285715,
      "accuracy_std_dev": 0.1097678440702774,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.16768704,
      "avg_duration_per_task_ms": 485101.77714285714,
      "avg_output_tokens_per_task": 17757.942857142858,
      "last_run_timestamp": "2026-05-09T17:11:48.691Z"
    },
    {
      "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": 0.29666649999999994,
      "accuracy_std_dev": 0.023333500000000007,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.05697704,
      "avg_duration_per_task_ms": 17166.81,
      "avg_output_tokens_per_task": 1588.56,
      "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": "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": 0.29243700336134454,
      "accuracy_std_dev": 0.037781726827337526,
      "run_count": 24,
      "total_questions": 1190,
      "avg_cost_per_task_usd": 0.012292368067226894,
      "avg_duration_per_task_ms": 153931.20336134452,
      "avg_output_tokens_per_task": 2026.116806722689,
      "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": 0.29205126923076924,
      "accuracy_std_dev": 0.03702711278089413,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.09791408615384616,
      "avg_duration_per_task_ms": 63451.20538461539,
      "avg_output_tokens_per_task": 1227.14,
      "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": "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": 0.2862745294117647,
      "accuracy_std_dev": 0.022107336631874436,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.006061691764705881,
      "avg_duration_per_task_ms": 76495.49647058823,
      "avg_output_tokens_per_task": 1836.8976470588236,
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Chutes",
      "accuracy": 0.2733333333333333,
      "accuracy_std_dev": 0.009428090415820642,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09871155999999999,
      "avg_duration_per_task_ms": 1053542.8,
      "avg_output_tokens_per_task": 16485.386666666665,
      "last_run_timestamp": "2026-05-02T21:11:19.074Z"
    },
    {
      "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": "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/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": "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": "Friendli",
      "accuracy": 0.2375,
      "accuracy_std_dev": 0.030720514318611274,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0706259825,
      "avg_duration_per_task_ms": 47810.7625,
      "avg_output_tokens_per_task": 9708.8825,
      "last_run_timestamp": "2026-05-16T08:51:28.206Z"
    },
    {
      "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": "Nebius",
      "accuracy": 0.22,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.5653728,
      "avg_duration_per_task_ms": 371898.36,
      "avg_output_tokens_per_task": 8626.18,
      "last_run_timestamp": "2026-04-18T22:28:10.840Z"
    },
    {
      "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": "WandB",
      "accuracy": 0.215,
      "accuracy_std_dev": 0.05361902647381805,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.08529526375,
      "avg_duration_per_task_ms": 35242.32625,
      "avg_output_tokens_per_task": 1680.2625,
      "last_run_timestamp": "2026-06-20T08:23:22.767Z"
    },
    {
      "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": "mistralai/mistral-nemo",
      "model_slug": "mistralai/mistral-nemo",
      "model_name": "Mistral: Mistral Nemo",
      "provider_name": "auto-routing",
      "accuracy": 0.18712804910714284,
      "accuracy_std_dev": 0.063489276040705,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.019668446428571427,
      "avg_duration_per_task_ms": 108649.33705357143,
      "avg_output_tokens_per_task": 2703.4866071428573,
      "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": 0.17888888888888888,
      "accuracy_std_dev": 0.06174408353742521,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021073126666666667,
      "avg_duration_per_task_ms": 119786.34888888888,
      "avg_output_tokens_per_task": 2837.891111111111,
      "last_run_timestamp": "2026-07-18T08:04:25.389Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen-2.5-7b-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": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Chutes",
      "accuracy": 0.155,
      "accuracy_std_dev": 0.043301270189221926,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.015117625,
      "avg_duration_per_task_ms": 773376.1,
      "avg_output_tokens_per_task": 13813.88,
      "last_run_timestamp": "2026-05-02T10:35:28.048Z"
    },
    {
      "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "SambaNova",
      "accuracy": 0.033333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0007856200000000001,
      "avg_duration_per_task_ms": 1400.1,
      "avg_output_tokens_per_task": 126.5,
      "last_run_timestamp": "2026-06-30T19:06:38.991Z"
    },
    {
      "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": "SiliconFlow",
      "accuracy": 0.02,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.19539478,
      "avg_duration_per_task_ms": 618160.14,
      "avg_output_tokens_per_task": 1815.22,
      "last_run_timestamp": "2026-04-18T10:55:30.513Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.013333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00098206,
      "avg_duration_per_task_ms": 2461.54,
      "avg_output_tokens_per_task": 61.9,
      "last_run_timestamp": "2026-07-01T01:58:01.413Z"
    },
    {
      "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": "AtlasCloud",
      "accuracy": 0.006667000000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00013588,
      "avg_duration_per_task_ms": 446,
      "avg_output_tokens_per_task": 9.72,
      "last_run_timestamp": "2026-06-30T19:02:54.924Z"
    }
  ]
}
