{
  "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": "2026-09-27",
  "retrieved_at": "2026-09-27T01:45:11.774Z",
  "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": 2121,
  "rows": [
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "Google",
      "accuracy": 0.952862,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.2003321212121212,
      "avg_duration_per_task_ms": 124514.04545454546,
      "avg_output_tokens_per_task": 16571.29797979798,
      "last_run_timestamp": "2026-09-01T17:50:10.294Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sakana/fugu-ultra",
      "model_slug": "sakana/fugu-ultra",
      "model_name": "Fugu Ultra",
      "provider_name": "auto-routing",
      "accuracy": 0.9461279999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 1.148739303030303,
      "avg_duration_per_task_ms": 328779.10101010103,
      "avg_output_tokens_per_task": 33271.606060606064,
      "last_run_timestamp": "2026-09-15T01:07:47.987Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "OpenAI",
      "accuracy": 0.9461279999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.116425398989899,
      "avg_duration_per_task_ms": 44572.09595959596,
      "avg_output_tokens_per_task": 2176.570707070707,
      "last_run_timestamp": "2026-09-06T03:17:58.152Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "Google AI Studio",
      "accuracy": 0.945286,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.2012271818181818,
      "avg_duration_per_task_ms": 134249.20707070708,
      "avg_output_tokens_per_task": 16644.886363636364,
      "last_run_timestamp": "2026-09-01T17:53:46.786Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.945286,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02981064646464646,
      "avg_duration_per_task_ms": 28082.027777777777,
      "avg_output_tokens_per_task": 7803.272727272727,
      "last_run_timestamp": "2026-09-01T13:28:27.192Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "auto-routing",
      "accuracy": 0.944444,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.20185933333333334,
      "avg_duration_per_task_ms": 126427.48484848485,
      "avg_output_tokens_per_task": 16698.565656565657,
      "last_run_timestamp": "2026-09-01T17:51:04.719Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.944444,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.11967219191919193,
      "avg_duration_per_task_ms": 38197.17171717172,
      "avg_output_tokens_per_task": 2017.3383838383838,
      "last_run_timestamp": "2026-09-26T10:54:50.352Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "Azure",
      "accuracy": 0.9436415317919074,
      "accuracy_std_dev": 0.0020135000000000014,
      "run_count": 2,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.09933485549132949,
      "avg_duration_per_task_ms": 198116.95664739885,
      "avg_output_tokens_per_task": 1767.9046242774566,
      "last_run_timestamp": "2026-09-06T03:05:46.155Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.942761,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.030509156565656566,
      "avg_duration_per_task_ms": 68136.79292929293,
      "avg_output_tokens_per_task": 7993.409090909091,
      "last_run_timestamp": "2026-09-01T13:25:46.082Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.9377104999999999,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.30815713636363634,
      "avg_duration_per_task_ms": 160641.4873737374,
      "avg_output_tokens_per_task": 10142.272727272728,
      "last_run_timestamp": "2026-09-01T13:20:32.033Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.93771,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.4446258484848485,
      "avg_duration_per_task_ms": 116119.66666666667,
      "avg_output_tokens_per_task": 11825.727272727272,
      "last_run_timestamp": "2026-08-01T08:45:01.642Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Parasail",
      "accuracy": 0.9351849999999999,
      "accuracy_std_dev": 0.006162785766193721,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1365803898989899,
      "avg_duration_per_task_ms": 176670.50606060607,
      "avg_output_tokens_per_task": 8896.30606060606,
      "last_run_timestamp": "2026-09-26T09:17:14.173Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google",
      "accuracy": 0.9335082598425197,
      "accuracy_std_dev": 0.011397999999999964,
      "run_count": 2,
      "total_questions": 381,
      "avg_cost_per_task_usd": 0.0213603280839895,
      "avg_duration_per_task_ms": 53235.17585301837,
      "avg_output_tokens_per_task": 7688.425196850394,
      "last_run_timestamp": "2026-09-01T13:29:18.783Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sakana/fugu-ultra-v2",
      "model_slug": "sakana/fugu-ultra-v2",
      "model_name": "Fugu Ultra V2",
      "provider_name": "auto-routing",
      "accuracy": 0.93266,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.5025571414141414,
      "avg_duration_per_task_ms": 194915.41919191918,
      "avg_output_tokens_per_task": 13806.227272727272,
      "last_run_timestamp": "2026-09-15T01:07:48.052Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.93266,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.12205839393939394,
      "avg_duration_per_task_ms": 394338.13636363635,
      "avg_output_tokens_per_task": 20077.80808080808,
      "last_run_timestamp": "2026-09-01T13:22:57.685Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.9324968387096775,
      "accuracy_std_dev": 0.005080239036370865,
      "run_count": 3,
      "total_questions": 558,
      "avg_cost_per_task_usd": 0.0715788476702509,
      "avg_duration_per_task_ms": 64320.11827956989,
      "avg_output_tokens_per_task": 11534.012544802867,
      "last_run_timestamp": "2026-09-01T17:51:06.380Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "OpenAI",
      "accuracy": 0.930976,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.32996825757575754,
      "avg_duration_per_task_ms": 166479.3383838384,
      "avg_output_tokens_per_task": 10868.631313131313,
      "last_run_timestamp": "2026-09-01T13:22:14.451Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Azure",
      "accuracy": 0.9299660000000001,
      "accuracy_std_dev": 0.009009433123121587,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.32134650707070705,
      "avg_duration_per_task_ms": 152235.298989899,
      "avg_output_tokens_per_task": 9994.927272727273,
      "last_run_timestamp": "2026-09-01T13:24:49.600Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.9294545733558178,
      "accuracy_std_dev": 0.004334827511626685,
      "run_count": 3,
      "total_questions": 593,
      "avg_cost_per_task_usd": 0.02681522428330523,
      "avg_duration_per_task_ms": 26517.369308600337,
      "avg_output_tokens_per_task": 2356.256323777403,
      "last_run_timestamp": "2026-09-23T15:41:43.462Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.9292930000000001,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14407252272727272,
      "avg_duration_per_task_ms": 58692.59848484849,
      "avg_output_tokens_per_task": 15884.012626262625,
      "last_run_timestamp": "2026-09-01T17:55:40.117Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "Google",
      "accuracy": 0.92887175,
      "accuracy_std_dev": 0.005239941859171739,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.06947331565656567,
      "avg_duration_per_task_ms": 67244.06060606061,
      "avg_output_tokens_per_task": 11635.200757575758,
      "last_run_timestamp": "2026-09-01T17:56:37.544Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.928451,
      "accuracy_std_dev": 0.004209000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.06683859343434344,
      "avg_duration_per_task_ms": 59600.07575757576,
      "avg_output_tokens_per_task": 11674.472222222223,
      "last_run_timestamp": "2026-09-01T17:57:17.668Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Makora",
      "accuracy": 0.9283648630849219,
      "accuracy_std_dev": 0.017830944261311557,
      "run_count": 4,
      "total_questions": 577,
      "avg_cost_per_task_usd": 0.09151112998266897,
      "avg_duration_per_task_ms": 252153.69497400348,
      "avg_output_tokens_per_task": 7044.975736568457,
      "last_run_timestamp": "2026-09-26T11:09:52.628Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Fireworks",
      "accuracy": 0.9279571938392187,
      "accuracy_std_dev": 0.00835201542708194,
      "run_count": 22,
      "total_questions": 3993,
      "avg_cost_per_task_usd": 0.15530522213874282,
      "avg_duration_per_task_ms": 142386.56699223642,
      "avg_output_tokens_per_task": 8728.15927873779,
      "last_run_timestamp": "2026-09-26T11:50:40.297Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.927609,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14183102777777779,
      "avg_duration_per_task_ms": 79740.97727272728,
      "avg_output_tokens_per_task": 15646.176767676769,
      "last_run_timestamp": "2026-09-01T18:06:42.346Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Relace",
      "accuracy": 0.92613625,
      "accuracy_std_dev": 0.007949293313716653,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.09945460101010102,
      "avg_duration_per_task_ms": 230595.33712121213,
      "avg_output_tokens_per_task": 9430.997474747475,
      "last_run_timestamp": "2026-09-26T08:53:13.109Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "Google",
      "accuracy": 0.9259257500000001,
      "accuracy_std_dev": 0.004610545920766882,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.1489979318181818,
      "avg_duration_per_task_ms": 84816.23358585859,
      "avg_output_tokens_per_task": 15665.375,
      "last_run_timestamp": "2026-09-01T17:51:30.647Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.925347078125,
      "accuracy_std_dev": 0.005810499999999996,
      "run_count": 2,
      "total_questions": 384,
      "avg_cost_per_task_usd": 0.3411057265625,
      "avg_duration_per_task_ms": 143482.140625,
      "avg_output_tokens_per_task": 10207.984375,
      "last_run_timestamp": "2026-09-01T13:21:58.114Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9242420000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.026419161616161615,
      "avg_duration_per_task_ms": 21367.934343434343,
      "avg_output_tokens_per_task": 2243.222222222222,
      "last_run_timestamp": "2026-09-23T15:27:41.867Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "Azure",
      "accuracy": 0.922559,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.4976751313131313,
      "avg_duration_per_task_ms": 91843.67676767676,
      "avg_output_tokens_per_task": 12034.212121212122,
      "last_run_timestamp": "2026-08-01T08:41:32.647Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Modal",
      "accuracy": 0.9224121127739175,
      "accuracy_std_dev": 0.007223723392406446,
      "run_count": 10,
      "total_questions": 1871,
      "avg_cost_per_task_usd": 0.15054908498129344,
      "avg_duration_per_task_ms": 122412.57936932122,
      "avg_output_tokens_per_task": 9889.19080705505,
      "last_run_timestamp": "2026-09-26T08:55:05.759Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9215687647058822,
      "accuracy_std_dev": 0.0038094199325123296,
      "run_count": 3,
      "total_questions": 561,
      "avg_cost_per_task_usd": 0.10569379500891266,
      "avg_duration_per_task_ms": 54609.55971479501,
      "avg_output_tokens_per_task": 3484.6114081996434,
      "last_run_timestamp": "2026-09-01T13:22:25.850Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.9212118,
      "accuracy_std_dev": 0.008675973568424486,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1201016404040404,
      "avg_duration_per_task_ms": 61346.69292929293,
      "avg_output_tokens_per_task": 3579.7959595959596,
      "last_run_timestamp": "2026-09-01T13:28:53.908Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.920875,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.4329253232323232,
      "avg_duration_per_task_ms": 113725.84848484848,
      "avg_output_tokens_per_task": 11493.873737373737,
      "last_run_timestamp": "2026-08-01T08:53:39.678Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Sail Research",
      "accuracy": 0.9205012222222222,
      "accuracy_std_dev": 0.00850227984821038,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.1576766610549944,
      "avg_duration_per_task_ms": 236410.12177328844,
      "avg_output_tokens_per_task": 12595.254208754208,
      "last_run_timestamp": "2026-09-26T08:55:21.775Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Wafer",
      "accuracy": 0.9200335,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07599968055555556,
      "avg_duration_per_task_ms": 289463.85101010103,
      "avg_output_tokens_per_task": 24687.122474747473,
      "last_run_timestamp": "2026-09-26T09:34:25.072Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Alibaba",
      "accuracy": 0.919865,
      "accuracy_std_dev": 0.007939302840930074,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.16831854444444444,
      "avg_duration_per_task_ms": 345204.0070707071,
      "avg_output_tokens_per_task": 9582.316161616161,
      "last_run_timestamp": "2026-09-26T09:45:25.093Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Wafer",
      "accuracy": 0.9192883359550563,
      "accuracy_std_dev": 0.004632004101898027,
      "run_count": 9,
      "total_questions": 1780,
      "avg_cost_per_task_usd": 0.19511707022471914,
      "avg_duration_per_task_ms": 338842.3713483146,
      "avg_output_tokens_per_task": 44188.06966292135,
      "last_run_timestamp": "2026-09-26T08:56:52.320Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "auto-routing",
      "accuracy": 0.9191915,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07126612626262625,
      "avg_duration_per_task_ms": 64525.257575757576,
      "avg_output_tokens_per_task": 3520.8156565656564,
      "last_run_timestamp": "2026-09-01T13:21:08.334Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "auto-routing",
      "accuracy": 0.9188843954372623,
      "accuracy_std_dev": 0.012617923846160294,
      "run_count": 8,
      "total_questions": 1578,
      "avg_cost_per_task_usd": 0.13335884410646387,
      "avg_duration_per_task_ms": 335309.8307984791,
      "avg_output_tokens_per_task": 10503.390367553866,
      "last_run_timestamp": "2026-09-26T08:52:39.134Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "NextBit",
      "accuracy": 0.9186867999999999,
      "accuracy_std_dev": 0.0068919500984844775,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.13311589696969697,
      "avg_duration_per_task_ms": 546514.5777777778,
      "avg_output_tokens_per_task": 36924.677777777775,
      "last_run_timestamp": "2026-09-26T09:34:49.008Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "BaseTen",
      "accuracy": 0.9179851756329115,
      "accuracy_std_dev": 0.013506707135345767,
      "run_count": 10,
      "total_questions": 1896,
      "avg_cost_per_task_usd": 0.12114161814345992,
      "avg_duration_per_task_ms": 120275.13765822785,
      "avg_output_tokens_per_task": 8135.254219409283,
      "last_run_timestamp": "2026-09-26T09:59:38.434Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.9175826670976084,
      "accuracy_std_dev": 0.014195212289272039,
      "run_count": 8,
      "total_questions": 1547,
      "avg_cost_per_task_usd": 0.008065983839689723,
      "avg_duration_per_task_ms": 489574.17323852616,
      "avg_output_tokens_per_task": 20043.881060116353,
      "last_run_timestamp": "2026-09-26T09:23:12.184Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.917508,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.025625949494949497,
      "avg_duration_per_task_ms": 37548.47474747475,
      "avg_output_tokens_per_task": 2404.0757575757575,
      "last_run_timestamp": "2026-09-23T15:32:55.695Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Baidu",
      "accuracy": 0.9164982,
      "accuracy_std_dev": 0.005280767383629021,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.09286130606060607,
      "avg_duration_per_task_ms": 636505.3202020202,
      "avg_output_tokens_per_task": 40298.016161616164,
      "last_run_timestamp": "2026-09-26T09:33:50.566Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Minimax",
      "accuracy": 0.9157045,
      "accuracy_std_dev": 0.012329586256237483,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.029113009740259738,
      "avg_duration_per_task_ms": 215302.51695526697,
      "avg_output_tokens_per_task": 24148.466810966813,
      "last_run_timestamp": "2026-09-26T16:05:15.213Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.914983,
      "accuracy_std_dev": 0.0008420000000000094,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.07349445707070708,
      "avg_duration_per_task_ms": 64386.71464646464,
      "avg_output_tokens_per_task": 3569.0353535353534,
      "last_run_timestamp": "2026-09-01T13:22:37.815Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "AtlasCloud",
      "accuracy": 0.9147023333333334,
      "accuracy_std_dev": 0.011850896290529652,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.1591313367003367,
      "avg_duration_per_task_ms": 538081.3703703703,
      "avg_output_tokens_per_task": 39867.37205387205,
      "last_run_timestamp": "2026-09-26T09:14:59.717Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Moonshot AI",
      "accuracy": 0.9139752900807381,
      "accuracy_std_dev": 0.006586842457168369,
      "run_count": 9,
      "total_questions": 1734,
      "avg_cost_per_task_usd": 0.14062165570934257,
      "avg_duration_per_task_ms": 294457.8246828143,
      "avg_output_tokens_per_task": 9226.012687427912,
      "last_run_timestamp": "2026-09-26T09:18:47.498Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Venice",
      "accuracy": 0.9127381666666666,
      "accuracy_std_dev": 0.0041711655212369175,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.27201272474747473,
      "avg_duration_per_task_ms": 357401.7491582492,
      "avg_output_tokens_per_task": 38560.36363636364,
      "last_run_timestamp": "2026-09-26T09:19:10.819Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "GMICloud",
      "accuracy": 0.9124577500000001,
      "accuracy_std_dev": 0.006628136630117118,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.0378646856060606,
      "avg_duration_per_task_ms": 130990.08964646465,
      "avg_output_tokens_per_task": 28457.30808080808,
      "last_run_timestamp": "2026-09-26T08:54:48.831Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.9107743333333334,
      "accuracy_std_dev": 0.004956316261992276,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03197664478114478,
      "avg_duration_per_task_ms": 194795.97138047137,
      "avg_output_tokens_per_task": 26470.969696969696,
      "last_run_timestamp": "2026-09-26T08:52:40.621Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "InferenceNet",
      "accuracy": 0.9104936666666668,
      "accuracy_std_dev": 0.05077924822427192,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.10505623232323233,
      "avg_duration_per_task_ms": 246993.5,
      "avg_output_tokens_per_task": 9288.036195286195,
      "last_run_timestamp": "2026-09-26T09:35:57.204Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.9102129999999999,
      "accuracy_std_dev": 0.002861334420627323,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0912948888888889,
      "avg_duration_per_task_ms": 155548.53703703705,
      "avg_output_tokens_per_task": 27869.387205387204,
      "last_run_timestamp": "2026-09-01T13:31:50.869Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Morph",
      "accuracy": 0.9096159477085782,
      "accuracy_std_dev": 0.03256740651027582,
      "run_count": 18,
      "total_questions": 3404,
      "avg_cost_per_task_usd": 0.1496695376028202,
      "avg_duration_per_task_ms": 473661.7364864865,
      "avg_output_tokens_per_task": 8839.687720329024,
      "last_run_timestamp": "2026-09-26T11:50:25.798Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "CoreWeave",
      "accuracy": 0.9091866447534767,
      "accuracy_std_dev": 0.0045401682512876225,
      "run_count": 8,
      "total_questions": 1582,
      "avg_cost_per_task_usd": 0.029416860935524654,
      "avg_duration_per_task_ms": 453238.0587863464,
      "avg_output_tokens_per_task": 30418.960809102402,
      "last_run_timestamp": "2026-09-26T16:21:44.172Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.7-max-20260520",
      "model_slug": "qwen/qwen3.7-max",
      "model_name": "Qwen: Qwen3.7 Max",
      "provider_name": "AtlasCloud",
      "accuracy": 0.9090909999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.2221867676767677,
      "avg_duration_per_task_ms": 298883.5202020202,
      "avg_output_tokens_per_task": 29352.99494949495,
      "last_run_timestamp": "2026-09-27T00:06:19.757Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.9090566691616767,
      "accuracy_std_dev": 0.011036068624917007,
      "run_count": 7,
      "total_questions": 1336,
      "avg_cost_per_task_usd": 0.012388138473053893,
      "avg_duration_per_task_ms": 304706.8982035928,
      "avg_output_tokens_per_task": 16929.82260479042,
      "last_run_timestamp": "2026-09-26T08:52:35.552Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Chutes",
      "accuracy": 0.9087729427112877,
      "accuracy_std_dev": 0.01605553275354013,
      "run_count": 9,
      "total_questions": 1763,
      "avg_cost_per_task_usd": 0.188728044242768,
      "avg_duration_per_task_ms": 517068.7197958026,
      "avg_output_tokens_per_task": 12546.453204764606,
      "last_run_timestamp": "2026-09-26T08:17:01.669Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.9086699999999999,
      "accuracy_std_dev": 0.0024899934738870384,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.035269609848484845,
      "avg_duration_per_task_ms": 184713.31186868687,
      "avg_output_tokens_per_task": 29215.651515151516,
      "last_run_timestamp": "2026-09-26T09:27:25.556Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.9074072500000001,
      "accuracy_std_dev": 0.004610545920766882,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.030798316919191916,
      "avg_duration_per_task_ms": 142600.18308080808,
      "avg_output_tokens_per_task": 28269.325757575756,
      "last_run_timestamp": "2026-09-26T09:35:41.692Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Novita",
      "accuracy": 0.9073833393548387,
      "accuracy_std_dev": 0.021106101149904508,
      "run_count": 12,
      "total_questions": 2325,
      "avg_cost_per_task_usd": 0.03617074838709677,
      "avg_duration_per_task_ms": 259960.5711827957,
      "avg_output_tokens_per_task": 30308.94064516129,
      "last_run_timestamp": "2026-09-26T16:06:12.156Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Alibaba",
      "accuracy": 0.9069262857142857,
      "accuracy_std_dev": 0.01966827796016641,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.13709101803751803,
      "avg_duration_per_task_ms": 571292.8715728716,
      "avg_output_tokens_per_task": 38969.107503607505,
      "last_run_timestamp": "2026-09-26T08:52:35.549Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "auto-routing",
      "accuracy": 0.9066039219088938,
      "accuracy_std_dev": 0.019658930346988745,
      "run_count": 14,
      "total_questions": 2766,
      "avg_cost_per_task_usd": 0.030758929501084598,
      "avg_duration_per_task_ms": 288550.2180043384,
      "avg_output_tokens_per_task": 25664.549168474332,
      "last_run_timestamp": "2026-09-26T16:08:07.059Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.905724,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.054357919191919186,
      "avg_duration_per_task_ms": 31472.4797979798,
      "avg_output_tokens_per_task": 2509.6414141414143,
      "last_run_timestamp": "2026-09-26T08:54:32.056Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "NextBit",
      "accuracy": 0.9057234999999999,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03042029292929293,
      "avg_duration_per_task_ms": 203300.4116161616,
      "avg_output_tokens_per_task": 29610.87373737374,
      "last_run_timestamp": "2026-09-26T10:10:49.311Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Reka",
      "accuracy": 0.9054903294117648,
      "accuracy_std_dev": 0.002950830881105991,
      "run_count": 3,
      "total_questions": 425,
      "avg_cost_per_task_usd": 0.0986494188235294,
      "avg_duration_per_task_ms": 590399.5082352941,
      "avg_output_tokens_per_task": 26203.308235294116,
      "last_run_timestamp": "2026-09-05T08:31:24.281Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Together",
      "accuracy": 0.9052352246520876,
      "accuracy_std_dev": 0.017017287952386474,
      "run_count": 13,
      "total_questions": 2515,
      "avg_cost_per_task_usd": 0.03245243419483101,
      "avg_duration_per_task_ms": 296394.9419483101,
      "avg_output_tokens_per_task": 27033.30656063618,
      "last_run_timestamp": "2026-09-26T16:32:47.455Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.9051626666666666,
      "accuracy_std_dev": 0.00649574344389378,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.028482462962962966,
      "avg_duration_per_task_ms": 244979.07912457912,
      "avg_output_tokens_per_task": 29233.41919191919,
      "last_run_timestamp": "2026-09-26T08:52:19.672Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.9041170761421321,
      "accuracy_std_dev": 0.007233580779177685,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.013811796954314722,
      "avg_duration_per_task_ms": 592278.3874788494,
      "avg_output_tokens_per_task": 28886.412859560067,
      "last_run_timestamp": "2026-09-26T09:15:49.652Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.9040400000000001,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02929719696969697,
      "avg_duration_per_task_ms": 146549.648989899,
      "avg_output_tokens_per_task": 27342.593434343435,
      "last_run_timestamp": "2026-09-01T13:36:23.129Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openrouter/auto",
      "model_slug": "openrouter/auto",
      "model_name": "Auto Router",
      "provider_name": "auto-routing",
      "accuracy": 0.9040399999999998,
      "accuracy_std_dev": 0.003367000000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02179906313131313,
      "avg_duration_per_task_ms": 200555.77272727274,
      "avg_output_tokens_per_task": 25393.28787878788,
      "last_run_timestamp": "2026-09-15T14:00:47.531Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Novita",
      "accuracy": 0.9037085815423515,
      "accuracy_std_dev": 0.00877916433950291,
      "run_count": 4,
      "total_questions": 791,
      "avg_cost_per_task_usd": 0.033949044247787614,
      "avg_duration_per_task_ms": 139050.61061946902,
      "avg_output_tokens_per_task": 28840.025284450065,
      "last_run_timestamp": "2026-09-26T08:51:27.499Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/switchyard",
      "model_slug": "nvidia/switchyard",
      "model_name": "NVIDIA: Switchyard",
      "provider_name": "auto-routing",
      "accuracy": 0.9032489624548736,
      "accuracy_std_dev": 0.03432136607100001,
      "run_count": 7,
      "total_questions": 1385,
      "avg_cost_per_task_usd": 0.017134858483754508,
      "avg_duration_per_task_ms": 71272.78700361011,
      "avg_output_tokens_per_task": 4031.8837545126353,
      "last_run_timestamp": "2026-09-23T21:40:42.470Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "auto-routing",
      "accuracy": 0.9031985,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14676221212121213,
      "avg_duration_per_task_ms": 106397.32070707071,
      "avg_output_tokens_per_task": 9655.194444444445,
      "last_run_timestamp": "2026-09-01T17:54:59.196Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.9031985,
      "accuracy_std_dev": 0.004208500000000004,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.043649184343434344,
      "avg_duration_per_task_ms": 141329.08080808082,
      "avg_output_tokens_per_task": 28078.323232323233,
      "last_run_timestamp": "2026-09-01T13:21:17.284Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Venice",
      "accuracy": 0.902848790174002,
      "accuracy_std_dev": 0.012183292931305548,
      "run_count": 10,
      "total_questions": 1954,
      "avg_cost_per_task_usd": 0.030339729273285567,
      "avg_duration_per_task_ms": 232833.84339815762,
      "avg_output_tokens_per_task": 25157.37871033777,
      "last_run_timestamp": "2026-09-26T16:18:02.009Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "StreamLake",
      "accuracy": 0.90256725,
      "accuracy_std_dev": 0.031891435832171314,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.042016968434343424,
      "avg_duration_per_task_ms": 260202.41792929292,
      "avg_output_tokens_per_task": 31470.176136363636,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "GMICloud",
      "accuracy": 0.9024125477178423,
      "accuracy_std_dev": 0.026053497227534046,
      "run_count": 11,
      "total_questions": 2169,
      "avg_cost_per_task_usd": 0.04220562056247118,
      "avg_duration_per_task_ms": 254266.19502074688,
      "avg_output_tokens_per_task": 31038.156754264637,
      "last_run_timestamp": "2026-09-26T16:06:21.027Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.9023565,
      "accuracy_std_dev": 0.0016835000000000044,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.053309800505050506,
      "avg_duration_per_task_ms": 26495.010101010103,
      "avg_output_tokens_per_task": 2451.373737373737,
      "last_run_timestamp": "2026-09-26T08:59:34.279Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.9013467199999999,
      "accuracy_std_dev": 0.005345045073860445,
      "run_count": 25,
      "total_questions": 4950,
      "avg_cost_per_task_usd": 0.21570517737373737,
      "avg_duration_per_task_ms": 94985.5094949495,
      "avg_output_tokens_per_task": 7862.312121212121,
      "last_run_timestamp": "2026-09-26T08:55:06.809Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Azure",
      "accuracy": 0.901199375,
      "accuracy_std_dev": 0.005700202593713214,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.2040564772727273,
      "avg_duration_per_task_ms": 98886.75126262626,
      "avg_output_tokens_per_task": 7886.399621212121,
      "last_run_timestamp": "2026-09-26T09:36:04.311Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Azure",
      "accuracy": 0.9006730000000001,
      "accuracy_std_dev": 0.007008969420202443,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1542171708754209,
      "avg_duration_per_task_ms": 127676.21296296296,
      "avg_output_tokens_per_task": 9507.647306397306,
      "last_run_timestamp": "2026-09-01T17:57:22.186Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Phala",
      "accuracy": 0.9006104712571067,
      "accuracy_std_dev": 0.058973902356121696,
      "run_count": 8,
      "total_questions": 1583,
      "avg_cost_per_task_usd": 0.14550911939355657,
      "avg_duration_per_task_ms": 301854.2046746684,
      "avg_output_tokens_per_task": 10187.655716993051,
      "last_run_timestamp": "2026-09-26T10:02:09.445Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Together",
      "accuracy": 0.9004862222222222,
      "accuracy_std_dev": 0.044346110676956096,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.14017589562289562,
      "avg_duration_per_task_ms": 217013.79966329967,
      "avg_output_tokens_per_task": 9267.047699214367,
      "last_run_timestamp": "2026-09-26T09:14:25.270Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Alibaba",
      "accuracy": 0.9003924999999999,
      "accuracy_std_dev": 0.01706955011855908,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.02751066750841751,
      "avg_duration_per_task_ms": 211685.4351851852,
      "avg_output_tokens_per_task": 25746.074915824916,
      "last_run_timestamp": "2026-09-26T08:50:52.330Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "SiliconFlow",
      "accuracy": 0.900252375,
      "accuracy_std_dev": 0.024011889023656058,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.15187610795454545,
      "avg_duration_per_task_ms": 607070.2234848485,
      "avg_output_tokens_per_task": 38443.12626262626,
      "last_run_timestamp": "2026-09-26T09:23:06.229Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8999515714285714,
      "accuracy_std_dev": 0.004917020144563225,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.23012572438672438,
      "avg_duration_per_task_ms": 111331.89033189033,
      "avg_output_tokens_per_task": 9001.132756132756,
      "last_run_timestamp": "2026-09-26T08:54:49.538Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "OpenAI",
      "accuracy": 0.8998315,
      "accuracy_std_dev": 0.007575499999999957,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13947029797979799,
      "avg_duration_per_task_ms": 103780.42929292929,
      "avg_output_tokens_per_task": 9170.424242424242,
      "last_run_timestamp": "2026-09-01T17:56:45.430Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepSeek",
      "accuracy": 0.8996633999999999,
      "accuracy_std_dev": 0.005792805834826492,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.007343192929292929,
      "avg_duration_per_task_ms": 219337.0404040404,
      "avg_output_tokens_per_task": 25809.762626262625,
      "last_run_timestamp": "2026-08-08T09:30:52.456Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "NextBit",
      "accuracy": 0.8995508333333332,
      "accuracy_std_dev": 0.004524373673657892,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.031289763468013466,
      "avg_duration_per_task_ms": 215945.1616161616,
      "avg_output_tokens_per_task": 26542.147306397306,
      "last_run_timestamp": "2026-09-26T09:19:06.104Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Baidu",
      "accuracy": 0.8993638888888887,
      "accuracy_std_dev": 0.005599097039194869,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.009368053310886646,
      "avg_duration_per_task_ms": 186510.51515151514,
      "avg_output_tokens_per_task": 26068.07014590348,
      "last_run_timestamp": "2026-09-26T08:51:16.648Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "DeepInfra",
      "accuracy": 0.8993387243502051,
      "accuracy_std_dev": 0.026386926357751014,
      "run_count": 8,
      "total_questions": 1462,
      "avg_cost_per_task_usd": 0.027840691518467853,
      "avg_duration_per_task_ms": 740625.9062927497,
      "avg_output_tokens_per_task": 24883.40560875513,
      "last_run_timestamp": "2026-09-12T16:17:57.063Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.8989897499999999,
      "accuracy_std_dev": 0.007241239789393777,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.12451013636363638,
      "avg_duration_per_task_ms": 29637.002525252527,
      "avg_output_tokens_per_task": 2284.373737373737,
      "last_run_timestamp": "2026-09-26T08:53:11.291Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8987091666666667,
      "accuracy_std_dev": 0.006255284576970791,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.05856475252525253,
      "avg_duration_per_task_ms": 31561.38973063973,
      "avg_output_tokens_per_task": 2534.496632996633,
      "last_run_timestamp": "2026-09-26T09:18:55.607Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Google",
      "accuracy": 0.8986660000000002,
      "accuracy_std_dev": 0.006701794386580333,
      "run_count": 26,
      "total_questions": 5148,
      "avg_cost_per_task_usd": 0.22345247746697744,
      "avg_duration_per_task_ms": 95310.8921911422,
      "avg_output_tokens_per_task": 8135.330031080031,
      "last_run_timestamp": "2026-09-26T09:16:27.499Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Phala",
      "accuracy": 0.8978795217391303,
      "accuracy_std_dev": 0.02199333682676437,
      "run_count": 7,
      "total_questions": 1242,
      "avg_cost_per_task_usd": 0.15500312399355878,
      "avg_duration_per_task_ms": 574496.192431562,
      "avg_output_tokens_per_task": 37507.40821256039,
      "last_run_timestamp": "2026-09-26T09:25:58.145Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Google",
      "accuracy": 0.8978675,
      "accuracy_std_dev": 0.004085377899697062,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.057460788720538725,
      "avg_duration_per_task_ms": 25029.281986531987,
      "avg_output_tokens_per_task": 2489.7180134680134,
      "last_run_timestamp": "2026-09-26T09:20:45.216Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.897348959390863,
      "accuracy_std_dev": 0.009933136206097672,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.021320504230118442,
      "avg_duration_per_task_ms": 601175.5499153977,
      "avg_output_tokens_per_task": 17639.832487309646,
      "last_run_timestamp": "2026-09-26T09:15:29.044Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8973064285714285,
      "accuracy_std_dev": 0.009132616193107635,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.19710293722943723,
      "avg_duration_per_task_ms": 95577.20057720058,
      "avg_output_tokens_per_task": 7683.466089466089,
      "last_run_timestamp": "2026-09-26T09:05:17.040Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8973061999999999,
      "accuracy_std_dev": 0.012997001706547546,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.03622809090909091,
      "avg_duration_per_task_ms": 162801.12121212122,
      "avg_output_tokens_per_task": 27180.768686868687,
      "last_run_timestamp": "2026-09-26T08:53:11.955Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "auto-routing",
      "accuracy": 0.8970957500000001,
      "accuracy_std_dev": 0.013079356862533413,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.20200678535353536,
      "avg_duration_per_task_ms": 97703.5946969697,
      "avg_output_tokens_per_task": 7866.200757575758,
      "last_run_timestamp": "2026-09-26T08:52:41.027Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Azure",
      "accuracy": 0.8967507774936062,
      "accuracy_std_dev": 0.0073637026710600875,
      "run_count": 18,
      "total_questions": 3519,
      "avg_cost_per_task_usd": 0.12523853253765274,
      "avg_duration_per_task_ms": 58154.57402671214,
      "avg_output_tokens_per_task": 4750.331628303496,
      "last_run_timestamp": "2026-09-26T08:53:17.186Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "OpenAI",
      "accuracy": 0.8964645000000001,
      "accuracy_std_dev": 0.007575499999999957,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03375818181818182,
      "avg_duration_per_task_ms": 56405.70202020202,
      "avg_output_tokens_per_task": 3599.8333333333335,
      "last_run_timestamp": "2026-09-01T13:22:46.235Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "auto-routing",
      "accuracy": 0.8961193926247291,
      "accuracy_std_dev": 0.016375418175062377,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.11571196312364425,
      "avg_duration_per_task_ms": 621953.1062906724,
      "avg_output_tokens_per_task": 36359.131597975414,
      "last_run_timestamp": "2026-09-26T09:20:04.760Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8960958972972972,
      "accuracy_std_dev": 0.005877557334660528,
      "run_count": 3,
      "total_questions": 555,
      "avg_cost_per_task_usd": 0.1555473855855856,
      "avg_duration_per_task_ms": 83534.15135135135,
      "avg_output_tokens_per_task": 9299.446846846848,
      "last_run_timestamp": "2026-09-01T17:56:10.791Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "StreamLake",
      "accuracy": 0.8959085241877256,
      "accuracy_std_dev": 0.035618009581258374,
      "run_count": 7,
      "total_questions": 1385,
      "avg_cost_per_task_usd": 0.11392027942238267,
      "avg_duration_per_task_ms": 563860.525631769,
      "avg_output_tokens_per_task": 38531.1083032491,
      "last_run_timestamp": "2026-09-26T09:10:17.436Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DekaLLM",
      "accuracy": 0.895623,
      "accuracy_std_dev": 0.006733999999999962,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.016944118686868685,
      "avg_duration_per_task_ms": 304656.29545454547,
      "avg_output_tokens_per_task": 21473.98232323232,
      "last_run_timestamp": "2026-09-26T08:26:07.956Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Azure",
      "accuracy": 0.89562275,
      "accuracy_std_dev": 0.0051888143816771525,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.057555659090909095,
      "avg_duration_per_task_ms": 24846.121212121212,
      "avg_output_tokens_per_task": 2521.655303030303,
      "last_run_timestamp": "2026-09-26T09:36:25.776Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Anthropic",
      "accuracy": 0.8954697272727271,
      "accuracy_std_dev": 0.006025463303060488,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.20493744031221303,
      "avg_duration_per_task_ms": 100289.02800734619,
      "avg_output_tokens_per_task": 7987.1299357208445,
      "last_run_timestamp": "2026-09-26T08:51:26.325Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8953789065420559,
      "accuracy_std_dev": 0.010355017014034617,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.012003740654205609,
      "avg_duration_per_task_ms": 284653.5062305296,
      "avg_output_tokens_per_task": 27098.440809968848,
      "last_run_timestamp": "2026-09-26T08:56:10.606Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "Novita",
      "accuracy": 0.8953486936299293,
      "accuracy_std_dev": 0.019178869929169434,
      "run_count": 5,
      "total_questions": 989,
      "avg_cost_per_task_usd": 0.03527696764408493,
      "avg_duration_per_task_ms": 158313.13549039434,
      "avg_output_tokens_per_task": 26466.745197168857,
      "last_run_timestamp": "2026-09-26T09:36:01.171Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Anthropic",
      "accuracy": 0.8950614444444446,
      "accuracy_std_dev": 0.007975755054623836,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.12023811111111112,
      "avg_duration_per_task_ms": 57861.70202020202,
      "avg_output_tokens_per_task": 4595.748597081931,
      "last_run_timestamp": "2026-09-26T08:56:21.122Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Novita",
      "accuracy": 0.8949365645569619,
      "accuracy_std_dev": 0.017816727566741866,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.021203940084388185,
      "avg_duration_per_task_ms": 226768.21476793248,
      "avg_output_tokens_per_task": 26685.461603375526,
      "last_run_timestamp": "2026-09-26T08:50:36.105Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Venice",
      "accuracy": 0.8946420862998922,
      "accuracy_std_dev": 0.007735019467331668,
      "run_count": 5,
      "total_questions": 927,
      "avg_cost_per_task_usd": 0.006440504854368933,
      "avg_duration_per_task_ms": 428399.7303128371,
      "avg_output_tokens_per_task": 13363.50377562028,
      "last_run_timestamp": "2026-09-26T08:57:06.874Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "auto-routing",
      "accuracy": 0.8939478279932548,
      "accuracy_std_dev": 0.009294310315828362,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.22570986565486229,
      "avg_duration_per_task_ms": 108477.89263631254,
      "avg_output_tokens_per_task": 8804.630691399663,
      "last_run_timestamp": "2026-09-26T08:50:43.725Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DeepInfra",
      "accuracy": 0.8934859242520689,
      "accuracy_std_dev": 0.016678015536459194,
      "run_count": 8,
      "total_questions": 1571,
      "avg_cost_per_task_usd": 0.11211757733927435,
      "avg_duration_per_task_ms": 393101.5945257798,
      "avg_output_tokens_per_task": 7662.472310630172,
      "last_run_timestamp": "2026-09-26T09:51:40.881Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Google",
      "accuracy": 0.8931287407407407,
      "accuracy_std_dev": 0.007059243819500409,
      "run_count": 27,
      "total_questions": 5346,
      "avg_cost_per_task_usd": 0.2046095759446315,
      "avg_duration_per_task_ms": 71087.34867190423,
      "avg_output_tokens_per_task": 7454.418069584736,
      "last_run_timestamp": "2026-09-26T09:18:54.068Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Azure",
      "accuracy": 0.8930976666666668,
      "accuracy_std_dev": 0.010182556173290759,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.05509197895622895,
      "avg_duration_per_task_ms": 49604.97474747475,
      "avg_output_tokens_per_task": 3808.635521885522,
      "last_run_timestamp": "2026-09-01T13:23:50.897Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8930975,
      "accuracy_std_dev": 0.0025254999999999583,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.054297787878787876,
      "avg_duration_per_task_ms": 31463.27525252525,
      "avg_output_tokens_per_task": 2503.689393939394,
      "last_run_timestamp": "2026-09-26T08:53:18.330Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "BaseTen",
      "accuracy": 0.892567903763697,
      "accuracy_std_dev": 0.02022831814050264,
      "run_count": 11,
      "total_questions": 2099,
      "avg_cost_per_task_usd": 0.15797696188661267,
      "avg_duration_per_task_ms": 338559.5883754169,
      "avg_output_tokens_per_task": 39638.616484040016,
      "last_run_timestamp": "2026-09-19T08:51:08.108Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Novita",
      "accuracy": 0.8920512903225806,
      "accuracy_std_dev": 0.019911726970242915,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.13115581593927894,
      "avg_duration_per_task_ms": 517762.9652118912,
      "avg_output_tokens_per_task": 37601.20493358634,
      "last_run_timestamp": "2026-09-26T09:00:46.358Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8918440425531915,
      "accuracy_std_dev": 0.0082731599766957,
      "run_count": 3,
      "total_questions": 564,
      "avg_cost_per_task_usd": 0.050205347517730495,
      "avg_duration_per_task_ms": 44870.67021276596,
      "avg_output_tokens_per_task": 3654.1524822695037,
      "last_run_timestamp": "2026-09-01T13:21:31.062Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "ModelRun",
      "accuracy": 0.8914139999999999,
      "accuracy_std_dev": 0.023312104420236266,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.086096711489899,
      "avg_duration_per_task_ms": 172501.05744949495,
      "avg_output_tokens_per_task": 29063.86047979798,
      "last_run_timestamp": "2026-09-26T16:08:44.458Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Parasail",
      "accuracy": 0.8914139999999999,
      "accuracy_std_dev": 0.017677,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04263949494949495,
      "avg_duration_per_task_ms": 361913.06313131313,
      "avg_output_tokens_per_task": 15178.89898989899,
      "last_run_timestamp": "2026-05-23T08:52:13.739Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Inceptron",
      "accuracy": 0.8913139844097995,
      "accuracy_std_dev": 0.016321403571860956,
      "run_count": 12,
      "total_questions": 2245,
      "avg_cost_per_task_usd": 0.09373735100222716,
      "avg_duration_per_task_ms": 534503.8146993319,
      "avg_output_tokens_per_task": 28512.390200445436,
      "last_run_timestamp": "2026-09-26T08:56:57.645Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.8905724999999999,
      "accuracy_std_dev": 0.011784499999999976,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.027024376262626264,
      "avg_duration_per_task_ms": 278293.2196969697,
      "avg_output_tokens_per_task": 26273.641414141413,
      "last_run_timestamp": "2026-09-19T09:00:31.402Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "auto-routing",
      "accuracy": 0.890572,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.21863568181818183,
      "avg_duration_per_task_ms": 106364.56818181818,
      "avg_output_tokens_per_task": 8520.704545454546,
      "last_run_timestamp": "2026-09-01T13:30:25.730Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "SiliconFlow",
      "accuracy": 0.890572,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.1261595202020202,
      "avg_duration_per_task_ms": 738082.0303030303,
      "avg_output_tokens_per_task": 50200.232323232325,
      "last_run_timestamp": "2026-09-23T06:12:17.757Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.8905389900990098,
      "accuracy_std_dev": 0.007335499999999995,
      "run_count": 2,
      "total_questions": 303,
      "avg_cost_per_task_usd": 0.01631713201320132,
      "avg_duration_per_task_ms": 272632.9339933993,
      "avg_output_tokens_per_task": 22974.214521452144,
      "last_run_timestamp": "2026-09-26T09:25:09.058Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "auto-routing",
      "accuracy": 0.8897304999999999,
      "accuracy_std_dev": 0.005892499999999967,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0318945303030303,
      "avg_duration_per_task_ms": 52439.469696969696,
      "avg_output_tokens_per_task": 3386.712121212121,
      "last_run_timestamp": "2026-09-01T13:38:25.421Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepInfra",
      "accuracy": 0.8897192682926828,
      "accuracy_std_dev": 0.012292154908071387,
      "run_count": 13,
      "total_questions": 2542,
      "avg_cost_per_task_usd": 0.0036178985051140834,
      "avg_duration_per_task_ms": 277074.0971675846,
      "avg_output_tokens_per_task": 19808.066876475215,
      "last_run_timestamp": "2026-09-26T09:01:36.155Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.888889,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00832349494949495,
      "avg_duration_per_task_ms": 86259.9494949495,
      "avg_output_tokens_per_task": 13087.217171717171,
      "last_run_timestamp": "2026-09-24T09:11:45.985Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Anthropic",
      "accuracy": 0.8888888430379747,
      "accuracy_std_dev": 0.009622693519820156,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.21254848649789027,
      "avg_duration_per_task_ms": 102777.66371308017,
      "avg_output_tokens_per_task": 8273.845147679325,
      "last_run_timestamp": "2026-09-26T09:15:06.377Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "GMICloud",
      "accuracy": 0.8887592307692308,
      "accuracy_std_dev": 0.013361790161926698,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.012804753690753689,
      "avg_duration_per_task_ms": 213849.3986013986,
      "avg_output_tokens_per_task": 24454.132867132867,
      "last_run_timestamp": "2026-09-26T08:57:03.139Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Anthropic",
      "accuracy": 0.888552,
      "accuracy_std_dev": 0.03117712682079604,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1316316707070707,
      "avg_duration_per_task_ms": 39126.25959595959,
      "avg_output_tokens_per_task": 2423.617171717172,
      "last_run_timestamp": "2026-09-26T08:53:20.595Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Google",
      "accuracy": 0.8883868300395255,
      "accuracy_std_dev": 0.008563760488459709,
      "run_count": 27,
      "total_questions": 5313,
      "avg_cost_per_task_usd": 0.1336053077357425,
      "avg_duration_per_task_ms": 59533.9862601167,
      "avg_output_tokens_per_task": 4780.512516469038,
      "last_run_timestamp": "2026-09-26T09:19:13.119Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Parasail",
      "accuracy": 0.8876862857142855,
      "accuracy_std_dev": 0.018697250017157118,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.13258516161616163,
      "avg_duration_per_task_ms": 538569.6168831169,
      "avg_output_tokens_per_task": 33498.70923520924,
      "last_run_timestamp": "2026-09-26T09:29:09.038Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.8875662777777777,
      "accuracy_std_dev": 0.03225232181080149,
      "run_count": 4,
      "total_questions": 756,
      "avg_cost_per_task_usd": 0.0045816917989418,
      "avg_duration_per_task_ms": 604811.5727513228,
      "avg_output_tokens_per_task": 27586.727513227514,
      "last_run_timestamp": "2026-09-26T08:58:10.498Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8872987222222223,
      "accuracy_std_dev": 0.006285923036662124,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.24028202581369246,
      "avg_duration_per_task_ms": 84935.12850729517,
      "avg_output_tokens_per_task": 8938.721380471381,
      "last_run_timestamp": "2026-09-26T08:53:09.256Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "Anthropic",
      "accuracy": 0.8872055000000001,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20363688131313135,
      "avg_duration_per_task_ms": 99426.9191919192,
      "avg_output_tokens_per_task": 7920.752525252526,
      "last_run_timestamp": "2026-09-01T13:21:53.444Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Azure",
      "accuracy": 0.8868242601351352,
      "accuracy_std_dev": 0.009599924258960477,
      "run_count": 9,
      "total_questions": 1776,
      "avg_cost_per_task_usd": 0.20454106644144146,
      "avg_duration_per_task_ms": 100059.57263513513,
      "avg_output_tokens_per_task": 7955.086148648648,
      "last_run_timestamp": "2026-09-26T08:50:54.262Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Fireworks",
      "accuracy": 0.8867051248554914,
      "accuracy_std_dev": 0.021208571037200977,
      "run_count": 5,
      "total_questions": 865,
      "avg_cost_per_task_usd": 0.16137210289017342,
      "avg_duration_per_task_ms": 430828.78150289017,
      "avg_output_tokens_per_task": 40500.787283237,
      "last_run_timestamp": "2026-09-19T08:53:57.162Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.8866439999999999,
      "accuracy_std_dev": 0.012472582491208459,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.008770978114478115,
      "avg_duration_per_task_ms": 61645.40740740741,
      "avg_output_tokens_per_task": 12897.501683501683,
      "last_run_timestamp": "2026-09-24T09:11:57.282Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8864601803278688,
      "accuracy_std_dev": 0.006211138077858382,
      "run_count": 3,
      "total_questions": 549,
      "avg_cost_per_task_usd": 0.009764420765027322,
      "avg_duration_per_task_ms": 50630.12750455373,
      "avg_output_tokens_per_task": 7246.790528233151,
      "last_run_timestamp": "2026-09-01T17:51:11.087Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.8864530691399664,
      "accuracy_std_dev": 0.01029973730112676,
      "run_count": 6,
      "total_questions": 1186,
      "avg_cost_per_task_usd": 0.023206354974704888,
      "avg_duration_per_task_ms": 81414.45531197301,
      "avg_output_tokens_per_task": 7880.962900505902,
      "last_run_timestamp": "2026-09-01T17:55:01.420Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Morph",
      "accuracy": 0.8863635000000001,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0069983762626262625,
      "avg_duration_per_task_ms": 635651.6565656565,
      "avg_output_tokens_per_task": 21974.18181818182,
      "last_run_timestamp": "2026-09-26T10:42:27.487Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Reka",
      "accuracy": 0.8862431428571429,
      "accuracy_std_dev": 0.008624548475610979,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.016605583694083693,
      "avg_duration_per_task_ms": 131101.75468975469,
      "avg_output_tokens_per_task": 25668.465367965367,
      "last_run_timestamp": "2026-09-26T08:53:15.047Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.8860826666666668,
      "accuracy_std_dev": 0.004199386807089298,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0278726632996633,
      "avg_duration_per_task_ms": 157091.45622895623,
      "avg_output_tokens_per_task": 23053.37037037037,
      "last_run_timestamp": "2026-09-26T09:47:15.375Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sakana/fugu-max",
      "model_slug": "sakana/fugu-max",
      "model_name": "Fugu Max",
      "provider_name": "auto-routing",
      "accuracy": 0.885522,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.07555469191919191,
      "avg_duration_per_task_ms": 225091.94949494948,
      "avg_output_tokens_per_task": 12279.550505050505,
      "last_run_timestamp": "2026-09-15T01:07:47.957Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Cloudflare",
      "accuracy": 0.8849736038095237,
      "accuracy_std_dev": 0.03742100947662959,
      "run_count": 8,
      "total_questions": 1575,
      "avg_cost_per_task_usd": 0.1364461853968254,
      "avg_duration_per_task_ms": 794770.7498412698,
      "avg_output_tokens_per_task": 34155.479365079365,
      "last_run_timestamp": "2026-09-26T10:09:19.111Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Phala",
      "accuracy": 0.8849202417582418,
      "accuracy_std_dev": 0.01938538920596302,
      "run_count": 3,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.04324962454212454,
      "avg_duration_per_task_ms": 205056.15750915752,
      "avg_output_tokens_per_task": 33761.00366300366,
      "last_run_timestamp": "2026-09-26T10:02:16.696Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cohere",
      "accuracy": 0.88468,
      "accuracy_std_dev": 0.005892000000000008,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.009158123737373737,
      "avg_duration_per_task_ms": 142387.94696969696,
      "avg_output_tokens_per_task": 32237.108585858587,
      "last_run_timestamp": "2026-09-26T09:33:13.989Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Sail Research",
      "accuracy": 0.8845707161572053,
      "accuracy_std_dev": 0.02143221713073307,
      "run_count": 6,
      "total_questions": 1145,
      "avg_cost_per_task_usd": 0.16085208820960697,
      "avg_duration_per_task_ms": 756655.7283842795,
      "avg_output_tokens_per_task": 40813.194759825325,
      "last_run_timestamp": "2026-09-16T00:37:42.883Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Relace",
      "accuracy": 0.8841129772439951,
      "accuracy_std_dev": 0.0033388888402431233,
      "run_count": 4,
      "total_questions": 791,
      "avg_cost_per_task_usd": 0.011443802781289506,
      "avg_duration_per_task_ms": 364670.14664981037,
      "avg_output_tokens_per_task": 22004.87104930468,
      "last_run_timestamp": "2026-09-26T12:21:31.561Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Parasail",
      "accuracy": 0.8839537178510998,
      "accuracy_std_dev": 0.013099282603342133,
      "run_count": 12,
      "total_questions": 2364,
      "avg_cost_per_task_usd": 0.00725531429780034,
      "avg_duration_per_task_ms": 489249.17766497465,
      "avg_output_tokens_per_task": 25997.4179357022,
      "last_run_timestamp": "2026-09-26T08:56:41.992Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "Novita",
      "accuracy": 0.883838,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.11161540404040404,
      "avg_duration_per_task_ms": 930549.2727272727,
      "avg_output_tokens_per_task": 44407.63636363636,
      "last_run_timestamp": "2026-09-23T15:38:01.445Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Azure",
      "accuracy": 0.8835014000000001,
      "accuracy_std_dev": 0.025562678362018323,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.14118799797979797,
      "avg_duration_per_task_ms": 41635.47171717172,
      "avg_output_tokens_per_task": 2614.5828282828284,
      "last_run_timestamp": "2026-09-26T08:54:00.492Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Novita",
      "accuracy": 0.8834305684210525,
      "accuracy_std_dev": 0.01461251838899233,
      "run_count": 13,
      "total_questions": 2565,
      "avg_cost_per_task_usd": 0.10276526354775828,
      "avg_duration_per_task_ms": 556976.750877193,
      "avg_output_tokens_per_task": 26780.168031189085,
      "last_run_timestamp": "2026-09-26T09:11:00.979Z"
    },
    {
      "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": "Phala",
      "accuracy": 0.8832771111111111,
      "accuracy_std_dev": 0.009589207202132645,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.02110603142536476,
      "avg_duration_per_task_ms": 350549.90852974186,
      "avg_output_tokens_per_task": 24400.115039281707,
      "last_run_timestamp": "2026-09-26T08:58:18.787Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "auto-routing",
      "accuracy": 0.8829965,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.13198183333333333,
      "avg_duration_per_task_ms": 254155.68181818182,
      "avg_output_tokens_per_task": 43876.38636363636,
      "last_run_timestamp": "2026-09-01T17:57:36.744Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Modal",
      "accuracy": 0.882716,
      "accuracy_std_dev": 0.006496204173720757,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.027370449494949496,
      "avg_duration_per_task_ms": 130101.09932659933,
      "avg_output_tokens_per_task": 22617.942760942762,
      "last_run_timestamp": "2026-09-26T10:08:37.923Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "DeepInfra",
      "accuracy": 0.8824913999999999,
      "accuracy_std_dev": 0.008205959897537906,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.015855806060606062,
      "avg_duration_per_task_ms": 286632.0282828283,
      "avg_output_tokens_per_task": 24219.753535353535,
      "last_run_timestamp": "2026-09-26T09:30:12.511Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8822544579439251,
      "accuracy_std_dev": 0.049358067293064335,
      "run_count": 15,
      "total_questions": 2889,
      "avg_cost_per_task_usd": 0.07833941329179646,
      "avg_duration_per_task_ms": 744926.2416060921,
      "avg_output_tokens_per_task": 34819.21737625476,
      "last_run_timestamp": "2026-08-29T20:15:38.359Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Moonshot AI",
      "accuracy": 0.8822322634862387,
      "accuracy_std_dev": 0.012422583804613795,
      "run_count": 28,
      "total_questions": 5450,
      "avg_cost_per_task_usd": 0.1723975299082569,
      "avg_duration_per_task_ms": 339570.9011009174,
      "avg_output_tokens_per_task": 27806.31211009174,
      "last_run_timestamp": "2026-09-26T08:58:33.306Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Together",
      "accuracy": 0.8821546666666666,
      "accuracy_std_dev": 0.001374572014200141,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.02880540235690236,
      "avg_duration_per_task_ms": 99924.20707070707,
      "avg_output_tokens_per_task": 23871.781144781144,
      "last_run_timestamp": "2026-09-26T09:49:42.288Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "CoreWeave",
      "accuracy": 0.8821545714285715,
      "accuracy_std_dev": 0.014256633562332045,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.18132904473304476,
      "avg_duration_per_task_ms": 297976.3398268398,
      "avg_output_tokens_per_task": 45509.72438672439,
      "last_run_timestamp": "2026-09-26T09:19:29.410Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8818613645976111,
      "accuracy_std_dev": 0.010002367123265743,
      "run_count": 27,
      "total_questions": 5107,
      "avg_cost_per_task_usd": 0.1356296686900333,
      "avg_duration_per_task_ms": 62377.717642451535,
      "avg_output_tokens_per_task": 4931.54278441355,
      "last_run_timestamp": "2026-09-26T08:58:07.474Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "StreamLake",
      "accuracy": 0.8806582222222221,
      "accuracy_std_dev": 0.005109133929167741,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.005640715488215489,
      "avg_duration_per_task_ms": 362315.8209876543,
      "avg_output_tokens_per_task": 28732.147025813694,
      "last_run_timestamp": "2026-09-26T09:34:56.325Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.880471,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0032475000000000004,
      "avg_duration_per_task_ms": 58949.292929292926,
      "avg_output_tokens_per_task": 5746.025252525253,
      "last_run_timestamp": "2026-09-23T06:09:34.299Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "Wafer",
      "accuracy": 0.8799936573033706,
      "accuracy_std_dev": 0.07692276680482653,
      "run_count": 12,
      "total_questions": 2136,
      "avg_cost_per_task_usd": 0.008731384363295881,
      "avg_duration_per_task_ms": 144796.46582397004,
      "avg_output_tokens_per_task": 23515.395599250936,
      "last_run_timestamp": "2026-09-26T08:54:04.117Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Decart",
      "accuracy": 0.8798644568527919,
      "accuracy_std_dev": 0.004527248931623837,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.0029938477157360403,
      "avg_duration_per_task_ms": 316818.90016920475,
      "avg_output_tokens_per_task": 18321.88663282572,
      "last_run_timestamp": "2026-08-22T10:20:26.941Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Inceptron",
      "accuracy": 0.8792515918367347,
      "accuracy_std_dev": 0.00992611788026112,
      "run_count": 9,
      "total_questions": 1764,
      "avg_cost_per_task_usd": 0.004467375283446712,
      "avg_duration_per_task_ms": 383208.2947845805,
      "avg_output_tokens_per_task": 18476.153628117914,
      "last_run_timestamp": "2026-09-26T10:33:25.209Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Reka",
      "accuracy": 0.879189,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 189,
      "avg_cost_per_task_usd": 0.020723571428571428,
      "avg_duration_per_task_ms": 141095.54497354498,
      "avg_output_tokens_per_task": 17734.21164021164,
      "last_run_timestamp": "2026-09-15T00:33:26.568Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.8784510000000001,
      "accuracy_std_dev": 0.011536704780828876,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.022268736363636365,
      "avg_duration_per_task_ms": 229673.4696969697,
      "avg_output_tokens_per_task": 24972.11919191919,
      "last_run_timestamp": "2026-09-26T09:19:28.485Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8784135555555556,
      "accuracy_std_dev": 0.017145903852336865,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.023216065656565654,
      "avg_duration_per_task_ms": 214871.89786756455,
      "avg_output_tokens_per_task": 24869.92312008979,
      "last_run_timestamp": "2026-09-26T09:39:38.744Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8783643322981367,
      "accuracy_std_dev": 0.020055520558938385,
      "run_count": 10,
      "total_questions": 1932,
      "avg_cost_per_task_usd": 0.1293156180124224,
      "avg_duration_per_task_ms": 63254.61180124224,
      "avg_output_tokens_per_task": 4948.346273291925,
      "last_run_timestamp": "2026-09-26T08:52:39.749Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.8779459999999999,
      "accuracy_std_dev": 0.010801196631237997,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.01636711111111111,
      "avg_duration_per_task_ms": 223245.67845117845,
      "avg_output_tokens_per_task": 24540.49494949495,
      "last_run_timestamp": "2026-09-26T10:45:50.086Z"
    },
    {
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8779238157894738,
      "accuracy_std_dev": 0.08964719889246571,
      "run_count": 14,
      "total_questions": 2736,
      "avg_cost_per_task_usd": 0.02495466630116959,
      "avg_duration_per_task_ms": 199904.41374269006,
      "avg_output_tokens_per_task": 21043.42909356725,
      "last_run_timestamp": "2026-09-26T16:07:03.311Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Together",
      "accuracy": 0.8776173653429603,
      "accuracy_std_dev": 0.01555889568113131,
      "run_count": 7,
      "total_questions": 1385,
      "avg_cost_per_task_usd": 0.08483577111913358,
      "avg_duration_per_task_ms": 220187.85559566788,
      "avg_output_tokens_per_task": 21593.077978339352,
      "last_run_timestamp": "2026-09-26T10:06:10.499Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "Google AI Studio",
      "accuracy": 0.8771045000000001,
      "accuracy_std_dev": 0.005050500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.14050905303030303,
      "avg_duration_per_task_ms": 185398.66919191918,
      "avg_output_tokens_per_task": 46713.32070707071,
      "last_run_timestamp": "2026-09-01T17:54:30.227Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Relace",
      "accuracy": 0.8770211787234045,
      "accuracy_std_dev": 0.03195742141593335,
      "run_count": 6,
      "total_questions": 1175,
      "avg_cost_per_task_usd": 0.0032351353191489365,
      "avg_duration_per_task_ms": 354647.9063829787,
      "avg_output_tokens_per_task": 26251.109787234043,
      "last_run_timestamp": "2026-09-19T08:54:45.345Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.8768792503888027,
      "accuracy_std_dev": 0.0068076105343578604,
      "run_count": 4,
      "total_questions": 643,
      "avg_cost_per_task_usd": 0.007664079315707621,
      "avg_duration_per_task_ms": 987610.4805598756,
      "avg_output_tokens_per_task": 30419.066874027994,
      "last_run_timestamp": "2026-09-26T09:19:44.251Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AkashML",
      "accuracy": 0.8765966933797908,
      "accuracy_std_dev": 0.020189124617118887,
      "run_count": 6,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.004125730836236933,
      "avg_duration_per_task_ms": 453615.28397212544,
      "avg_output_tokens_per_task": 16054.564459930314,
      "last_run_timestamp": "2026-09-05T08:24:32.148Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Novita",
      "accuracy": 0.8759223861480078,
      "accuracy_std_dev": 0.007557144477403089,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.005635244149272613,
      "avg_duration_per_task_ms": 187978.53478810878,
      "avg_output_tokens_per_task": 19921.54459203036,
      "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-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Ionstream",
      "accuracy": 0.8754206666666666,
      "accuracy_std_dev": 0.00901385812081723,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.061844909090909096,
      "avg_duration_per_task_ms": 262173.4595959596,
      "avg_output_tokens_per_task": 22965.14478114478,
      "last_run_timestamp": "2026-09-26T09:19:06.449Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.8748596666666667,
      "accuracy_std_dev": 0.006198499675100621,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.003131114478114478,
      "avg_duration_per_task_ms": 32869.14646464647,
      "avg_output_tokens_per_task": 5711.222222222223,
      "last_run_timestamp": "2026-09-23T06:08:44.627Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Novita",
      "accuracy": 0.8746655940919037,
      "accuracy_std_dev": 0.006656857328717377,
      "run_count": 14,
      "total_questions": 2742,
      "avg_cost_per_task_usd": 0.092771409190372,
      "avg_duration_per_task_ms": 766984.4496717724,
      "avg_output_tokens_per_task": 32707.43873085339,
      "last_run_timestamp": "2026-08-29T19:07:02.104Z"
    },
    {
      "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-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Makora",
      "accuracy": 0.8745317312013828,
      "accuracy_std_dev": 0.011186512454241005,
      "run_count": 6,
      "total_questions": 1157,
      "avg_cost_per_task_usd": 0.0043274416594641305,
      "avg_duration_per_task_ms": 185028.52031114954,
      "avg_output_tokens_per_task": 21870.51080380294,
      "last_run_timestamp": "2026-09-26T08:53:59.525Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Venice",
      "accuracy": 0.8744798294036062,
      "accuracy_std_dev": 0.034886602236436054,
      "run_count": 11,
      "total_questions": 2163,
      "avg_cost_per_task_usd": 0.13725122098936662,
      "avg_duration_per_task_ms": 368551.85991678224,
      "avg_output_tokens_per_task": 32192.02172907998,
      "last_run_timestamp": "2026-09-26T08:53:37.509Z"
    },
    {
      "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": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Google",
      "accuracy": 0.8740740000000001,
      "accuracy_std_dev": 0.03472083287595504,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1182012676767677,
      "avg_duration_per_task_ms": 28405.98686868687,
      "avg_output_tokens_per_task": 2155.238383838384,
      "last_run_timestamp": "2026-09-26T08:52:17.920Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "DeepInfra",
      "accuracy": 0.874074,
      "accuracy_std_dev": 0.004178433821421624,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07029311111111111,
      "avg_duration_per_task_ms": 171365.84242424244,
      "avg_output_tokens_per_task": 11443.449494949495,
      "last_run_timestamp": "2026-09-26T09:10:25.930Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "CoreWeave",
      "accuracy": 0.8737373,
      "accuracy_std_dev": 0.01570253542616605,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.11062107323232323,
      "avg_duration_per_task_ms": 229265.31616161615,
      "avg_output_tokens_per_task": 31458.501515151514,
      "last_run_timestamp": "2026-09-26T08:52:34.744Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "auto-routing",
      "accuracy": 0.8737372500000001,
      "accuracy_std_dev": 0.019341996831700202,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.12990520643939396,
      "avg_duration_per_task_ms": 63171.816287878784,
      "avg_output_tokens_per_task": 4978.773358585859,
      "last_run_timestamp": "2026-09-26T08:54:39.535Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8724287134608445,
      "accuracy_std_dev": 0.032129376762123,
      "run_count": 14,
      "total_questions": 2771,
      "avg_cost_per_task_usd": 0.1523031241429087,
      "avg_duration_per_task_ms": 425564.390111873,
      "avg_output_tokens_per_task": 11498.582100324793,
      "last_run_timestamp": "2026-09-26T09:12:17.926Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Z.AI",
      "accuracy": 0.872292240506329,
      "accuracy_std_dev": 0.013194491052666223,
      "run_count": 6,
      "total_questions": 1185,
      "avg_cost_per_task_usd": 0.00790528016877637,
      "avg_duration_per_task_ms": 421318.1687763713,
      "avg_output_tokens_per_task": 21316.46075949367,
      "last_run_timestamp": "2026-09-26T09:05:47.438Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8722870891089111,
      "accuracy_std_dev": 0.011050513265882567,
      "run_count": 17,
      "total_questions": 3333,
      "avg_cost_per_task_usd": 0.005860441644164416,
      "avg_duration_per_task_ms": 188504.78847884788,
      "avg_output_tokens_per_task": 20725.64086408641,
      "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-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepSeek",
      "accuracy": 0.8722740210280374,
      "accuracy_std_dev": 0.016757790257300998,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.004785041277258566,
      "avg_duration_per_task_ms": 167818.7609034268,
      "avg_output_tokens_per_task": 16930.278426791276,
      "last_run_timestamp": "2026-08-08T10:06:09.471Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "AkashML",
      "accuracy": 0.8720536666666666,
      "accuracy_std_dev": 0.008584133982075453,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07818634427609428,
      "avg_duration_per_task_ms": 229996.4814814815,
      "avg_output_tokens_per_task": 19586.67255892256,
      "last_run_timestamp": "2026-06-13T09:38:25.896Z"
    },
    {
      "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-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": "AtlasCloud",
      "accuracy": 0.8714926666666667,
      "accuracy_std_dev": 0.0067804647005611455,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03188070202020202,
      "avg_duration_per_task_ms": 240511.38888888888,
      "avg_output_tokens_per_task": 26418.183501683503,
      "last_run_timestamp": "2026-09-26T08:53:40.686Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "Azure",
      "accuracy": 0.871212,
      "accuracy_std_dev": 0.010943000000000036,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.25165717424242423,
      "avg_duration_per_task_ms": 248967.9292929293,
      "avg_output_tokens_per_task": 23863.48232323232,
      "last_run_timestamp": "2026-09-01T13:53:50.899Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Makora",
      "accuracy": 0.8711222842639593,
      "accuracy_std_dev": 0.0010748023074035804,
      "run_count": 3,
      "total_questions": 591,
      "avg_cost_per_task_usd": 0.04708872588832487,
      "avg_duration_per_task_ms": 168365.97292724196,
      "avg_output_tokens_per_task": 39058.98984771573,
      "last_run_timestamp": "2026-09-26T10:21:09.898Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "ModelRun",
      "accuracy": 0.8710750330630069,
      "accuracy_std_dev": 0.028110456226890915,
      "run_count": 9,
      "total_questions": 1603,
      "avg_cost_per_task_usd": 0.09870775982532752,
      "avg_duration_per_task_ms": 221158.11790393014,
      "avg_output_tokens_per_task": 26829.437305053027,
      "last_run_timestamp": "2026-09-26T09:19:13.012Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "CoreWeave",
      "accuracy": 0.8710435000000001,
      "accuracy_std_dev": 0.015469940324707153,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.007902119696969698,
      "avg_duration_per_task_ms": 203714.64696969697,
      "avg_output_tokens_per_task": 28019.556565656567,
      "last_run_timestamp": "2026-09-26T09:36:46.529Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Together",
      "accuracy": 0.8709859808450705,
      "accuracy_std_dev": 0.00995822887470397,
      "run_count": 9,
      "total_questions": 1775,
      "avg_cost_per_task_usd": 0.007974392676056338,
      "avg_duration_per_task_ms": 311914.2202816901,
      "avg_output_tokens_per_task": 28209.84957746479,
      "last_run_timestamp": "2026-09-26T09:58:11.632Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8709169574970483,
      "accuracy_std_dev": 0.025335242123251705,
      "run_count": 13,
      "total_questions": 2541,
      "avg_cost_per_task_usd": 0.09988292601338057,
      "avg_duration_per_task_ms": 696367.0350255804,
      "avg_output_tokens_per_task": 26502.97284533648,
      "last_run_timestamp": "2026-09-26T09:22:13.122Z"
    },
    {
      "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": 0.8705803019038983,
      "accuracy_std_dev": 0.09978272834395396,
      "run_count": 13,
      "total_questions": 2206,
      "avg_cost_per_task_usd": 0.01978426609247507,
      "avg_duration_per_task_ms": 262537.4936536718,
      "avg_output_tokens_per_task": 16415.930190389845,
      "last_run_timestamp": "2026-09-26T17:07:03.598Z"
    },
    {
      "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-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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Alibaba",
      "accuracy": 0.8700495714285714,
      "accuracy_std_dev": 0.014023880162648517,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.0307443746993747,
      "avg_duration_per_task_ms": 207892.97162097163,
      "avg_output_tokens_per_task": 13057.428811928812,
      "last_run_timestamp": "2026-08-29T18:36:08.027Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "BaseTen",
      "accuracy": 0.8699951448047163,
      "accuracy_std_dev": 0.03969212027542979,
      "run_count": 14,
      "total_questions": 2714,
      "avg_cost_per_task_usd": 0.00531859358879882,
      "avg_duration_per_task_ms": 114563.89867354458,
      "avg_output_tokens_per_task": 20095.400515843772,
      "last_run_timestamp": "2026-09-26T08:53:37.618Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Ionstream",
      "accuracy": 0.8699495,
      "accuracy_std_dev": 0.002104500000000009,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.00502325,
      "avg_duration_per_task_ms": 389434.8611111111,
      "avg_output_tokens_per_task": 17944.10101010101,
      "last_run_timestamp": "2026-08-08T10:05:07.443Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.8695795853658537,
      "accuracy_std_dev": 0.014776870412912198,
      "run_count": 5,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.010841884146341461,
      "avg_duration_per_task_ms": 398387.18699186994,
      "avg_output_tokens_per_task": 22967.41768292683,
      "last_run_timestamp": "2026-09-26T09:27:32.975Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.8695285000000001,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.10538721464646464,
      "avg_duration_per_task_ms": 581981.0101010101,
      "avg_output_tokens_per_task": 34695.7398989899,
      "last_run_timestamp": "2026-08-08T09:03:54.762Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "GMICloud",
      "accuracy": 0.869318125,
      "accuracy_std_dev": 0.03295688160474795,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.10640666414141413,
      "avg_duration_per_task_ms": 513815.1142676768,
      "avg_output_tokens_per_task": 34567.331439393936,
      "last_run_timestamp": "2026-09-26T08:56:11.901Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Sail Research",
      "accuracy": 0.8692809647058825,
      "accuracy_std_dev": 0.009104176033557385,
      "run_count": 4,
      "total_questions": 765,
      "avg_cost_per_task_usd": 0.0047775359477124175,
      "avg_duration_per_task_ms": 482494.045751634,
      "avg_output_tokens_per_task": 26353.58431372549,
      "last_run_timestamp": "2026-08-22T16:48:48.690Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Parasail",
      "accuracy": 0.8691380874524715,
      "accuracy_std_dev": 0.016741224006923745,
      "run_count": 16,
      "total_questions": 3156,
      "avg_cost_per_task_usd": 0.005154015209125474,
      "avg_duration_per_task_ms": 318150.1036121673,
      "avg_output_tokens_per_task": 18230.222116603294,
      "last_run_timestamp": "2026-08-29T18:13:33.416Z"
    },
    {
      "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": "Inceptron",
      "accuracy": 0.8689024817073171,
      "accuracy_std_dev": 0.04756926554684933,
      "run_count": 3,
      "total_questions": 328,
      "avg_cost_per_task_usd": 0.014946170731707318,
      "avg_duration_per_task_ms": 479610.80792682926,
      "avg_output_tokens_per_task": 32324.884146341465,
      "last_run_timestamp": "2026-09-26T08:55:13.371Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Ambient",
      "accuracy": 0.868687,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.13888562626262627,
      "avg_duration_per_task_ms": 1028019.2020202021,
      "avg_output_tokens_per_task": 46184.969696969696,
      "last_run_timestamp": "2026-07-18T16:16:27.436Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.868687,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0028392929292929293,
      "avg_duration_per_task_ms": 49008.020202020205,
      "avg_output_tokens_per_task": 5520.065656565656,
      "last_run_timestamp": "2026-09-23T06:08:55.567Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "auto-routing",
      "accuracy": 0.8683500000000001,
      "accuracy_std_dev": 0.014020687172888494,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.08152287070707072,
      "avg_duration_per_task_ms": 190241.53434343435,
      "avg_output_tokens_per_task": 13202.136363636364,
      "last_run_timestamp": "2026-09-26T08:55:09.429Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Together",
      "accuracy": 0.86835,
      "accuracy_std_dev": 0.01103428752570821,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07775004545454545,
      "avg_duration_per_task_ms": 107497.15858585859,
      "avg_output_tokens_per_task": 12580.208080808081,
      "last_run_timestamp": "2026-09-26T08:57:01.878Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Moonshot AI",
      "accuracy": 0.8682967644787642,
      "accuracy_std_dev": 0.012405441826741467,
      "run_count": 12,
      "total_questions": 2331,
      "avg_cost_per_task_usd": 0.15647866452166453,
      "avg_duration_per_task_ms": 840962.8275418276,
      "avg_output_tokens_per_task": 39244.66795366795,
      "last_run_timestamp": "2026-08-29T18:24:25.224Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Venice",
      "accuracy": 0.8681256666666667,
      "accuracy_std_dev": 0.02130941144710999,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.02939427441077441,
      "avg_duration_per_task_ms": 110502.21548821549,
      "avg_output_tokens_per_task": 19394.752525252527,
      "last_run_timestamp": "2026-09-26T09:38:56.771Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8677674782608696,
      "accuracy_std_dev": 0.016613732149306157,
      "run_count": 15,
      "total_questions": 2967,
      "avg_cost_per_task_usd": 0.0769334031007752,
      "avg_duration_per_task_ms": 376119.02392989554,
      "avg_output_tokens_per_task": 22811.041118975398,
      "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-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Venice",
      "accuracy": 0.8676579999999999,
      "accuracy_std_dev": 0.007874153076151538,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.006459232323232324,
      "avg_duration_per_task_ms": 253155.9539842873,
      "avg_output_tokens_per_task": 18088.584736251403,
      "last_run_timestamp": "2026-09-26T08:54:07.885Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Novita",
      "accuracy": 0.8670031666666665,
      "accuracy_std_dev": 0.004661340237051534,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07032303872053872,
      "avg_duration_per_task_ms": 224493.27861952863,
      "avg_output_tokens_per_task": 11452.973905723906,
      "last_run_timestamp": "2026-09-26T09:37:35.352Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "StreamLake",
      "accuracy": 0.8669794312026003,
      "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": "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-k2.7-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": "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/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Modal",
      "accuracy": 0.8659932,
      "accuracy_std_dev": 0.008010237759267841,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07810327272727272,
      "avg_duration_per_task_ms": 87737.49797979798,
      "avg_output_tokens_per_task": 12770.258585858586,
      "last_run_timestamp": "2026-09-26T08:55:49.839Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8658808333333333,
      "accuracy_std_dev": 0.0104532872886422,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.07000543097643097,
      "avg_duration_per_task_ms": 223270.64393939395,
      "avg_output_tokens_per_task": 11398.105218855218,
      "last_run_timestamp": "2026-09-26T08:58:06.175Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Fireworks",
      "accuracy": 0.8655813274418604,
      "accuracy_std_dev": 0.019123465253362573,
      "run_count": 31,
      "total_questions": 5375,
      "avg_cost_per_task_usd": 0.15140991627906977,
      "avg_duration_per_task_ms": 236648.42009302325,
      "avg_output_tokens_per_task": 27703.360372093022,
      "last_run_timestamp": "2026-09-26T08:54:22.466Z"
    },
    {
      "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": 0.8655346706281833,
      "accuracy_std_dev": 0.02323406358048937,
      "run_count": 15,
      "total_questions": 2945,
      "avg_cost_per_task_usd": 0.09417275280135823,
      "avg_duration_per_task_ms": 559483.7314091681,
      "avg_output_tokens_per_task": 31466.001697792868,
      "last_run_timestamp": "2026-08-29T18:25:54.650Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.8655206622851365,
      "accuracy_std_dev": 0.014576764442083846,
      "run_count": 15,
      "total_questions": 2967,
      "avg_cost_per_task_usd": 0.04512668891135827,
      "avg_duration_per_task_ms": 460619.7182339063,
      "avg_output_tokens_per_task": 23222.32389619144,
      "last_run_timestamp": "2026-08-29T19:37:04.387Z"
    },
    {
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.865319761904762,
      "accuracy_std_dev": 0.012651559952394205,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.04589295334295333,
      "avg_duration_per_task_ms": 204017.00841750842,
      "avg_output_tokens_per_task": 13051.606782106783,
      "last_run_timestamp": "2026-08-29T18:39:14.500Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "InferenceNet",
      "accuracy": 0.8651804917197452,
      "accuracy_std_dev": 0.013354329136557154,
      "run_count": 4,
      "total_questions": 785,
      "avg_cost_per_task_usd": 0.060597563057324846,
      "avg_duration_per_task_ms": 205391.9694267516,
      "avg_output_tokens_per_task": 23879.336305732482,
      "last_run_timestamp": "2026-09-26T09:20:06.107Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "GMICloud",
      "accuracy": 0.8649570000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.020998656410256413,
      "avg_duration_per_task_ms": 619238.4615384615,
      "avg_output_tokens_per_task": 23774.34358974359,
      "last_run_timestamp": "2026-09-26T01:32:20.726Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Phala",
      "accuracy": 0.8648248729281768,
      "accuracy_std_dev": 0.02631682070007696,
      "run_count": 5,
      "total_questions": 905,
      "avg_cost_per_task_usd": 0.010246941436464088,
      "avg_duration_per_task_ms": 355027.6861878453,
      "avg_output_tokens_per_task": 21157.914917127073,
      "last_run_timestamp": "2026-09-26T09:56:32.081Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Inceptron",
      "accuracy": 0.8636842273684211,
      "accuracy_std_dev": 0.01763815223202249,
      "run_count": 5,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.08584259578947369,
      "avg_duration_per_task_ms": 395986.4378947368,
      "avg_output_tokens_per_task": 24656.30947368421,
      "last_run_timestamp": "2026-09-26T08:57:34.809Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Mancer 2",
      "accuracy": 0.8636519239465571,
      "accuracy_std_dev": 0.028707041857903168,
      "run_count": 11,
      "total_questions": 1946,
      "avg_cost_per_task_usd": 0.010248814491264132,
      "avg_duration_per_task_ms": 324581.0220966084,
      "avg_output_tokens_per_task": 18963.51336073998,
      "last_run_timestamp": "2026-09-26T08:55:27.375Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Chutes",
      "accuracy": 0.8631489999999998,
      "accuracy_std_dev": 0.03249701661124569,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.03565506804891016,
      "avg_duration_per_task_ms": 175446.75730994152,
      "avg_output_tokens_per_task": 12155.395534290272,
      "last_run_timestamp": "2026-08-15T15:55:02.752Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Inceptron",
      "accuracy": 0.8631252833052278,
      "accuracy_std_dev": 0.03388611321929843,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.09767507644744237,
      "avg_duration_per_task_ms": 708453.107925801,
      "avg_output_tokens_per_task": 37470.564362001125,
      "last_run_timestamp": "2026-09-21T15:33:37.909Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Reka",
      "accuracy": 0.8631102531752752,
      "accuracy_std_dev": 0.01428710704815903,
      "run_count": 6,
      "total_questions": 1181,
      "avg_cost_per_task_usd": 0.006545317527519051,
      "avg_duration_per_task_ms": 246133.70618120237,
      "avg_output_tokens_per_task": 12855.507197290432,
      "last_run_timestamp": "2026-09-26T09:18:13.150Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Novita",
      "accuracy": 0.8630882576271186,
      "accuracy_std_dev": 0.015697174214519886,
      "run_count": 9,
      "total_questions": 1770,
      "avg_cost_per_task_usd": 0.16561930621468926,
      "avg_duration_per_task_ms": 967589.7152542373,
      "avg_output_tokens_per_task": 49108.66779661017,
      "last_run_timestamp": "2026-08-29T20:16:17.828Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DeepInfra",
      "accuracy": 0.8628256666666666,
      "accuracy_std_dev": 0.019902461007624172,
      "run_count": 5,
      "total_questions": 972,
      "avg_cost_per_task_usd": 0.08342864197530865,
      "avg_duration_per_task_ms": 454814.01543209876,
      "avg_output_tokens_per_task": 23652.057613168723,
      "last_run_timestamp": "2026-09-26T08:53:09.429Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Google",
      "accuracy": 0.8627943749999999,
      "accuracy_std_dev": 0.00929733078546604,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.855192389520202,
      "avg_duration_per_task_ms": 483915.25,
      "avg_output_tokens_per_task": 34017.0119949495,
      "last_run_timestamp": "2026-09-26T08:53:16.951Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Baidu",
      "accuracy": 0.8626742857142858,
      "accuracy_std_dev": 0.017858490299448887,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.08792987914862915,
      "avg_duration_per_task_ms": 406224.28066378064,
      "avg_output_tokens_per_task": 39434.920634920636,
      "last_run_timestamp": "2026-09-26T08:56:57.800Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.862584125,
      "accuracy_std_dev": 0.009408632345318578,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.8442858333333334,
      "avg_duration_per_task_ms": 415897.4154040404,
      "avg_output_tokens_per_task": 33580.344696969696,
      "last_run_timestamp": "2026-09-26T08:53:36.382Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "BaseTen",
      "accuracy": 0.8622473714739068,
      "accuracy_std_dev": 0.021847945469408447,
      "run_count": 29,
      "total_questions": 5672,
      "avg_cost_per_task_usd": 0.2541298106488012,
      "avg_duration_per_task_ms": 337436.6962270804,
      "avg_output_tokens_per_task": 46060.357369534555,
      "last_run_timestamp": "2026-09-26T11:00:26.140Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Morph",
      "accuracy": 0.8620522502732241,
      "accuracy_std_dev": 0.014311015984733796,
      "run_count": 14,
      "total_questions": 2745,
      "avg_cost_per_task_usd": 0.031429440072859735,
      "avg_duration_per_task_ms": 137144.43606557377,
      "avg_output_tokens_per_task": 8964.896903460838,
      "last_run_timestamp": "2026-07-04T18:14:56.425Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.8619530000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.008309641414141413,
      "avg_duration_per_task_ms": 730613.3737373737,
      "avg_output_tokens_per_task": 29358.666666666668,
      "last_run_timestamp": "2026-07-31T04:09:58.076Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.8619525,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.223144601010101,
      "avg_duration_per_task_ms": 273332.86363636365,
      "avg_output_tokens_per_task": 22220.545454545456,
      "last_run_timestamp": "2026-09-01T13:23:04.034Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "auto-routing",
      "accuracy": 0.8618233076923076,
      "accuracy_std_dev": 0.012536811386930849,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.0038161161616161623,
      "avg_duration_per_task_ms": 321804.18686868687,
      "avg_output_tokens_per_task": 18513.827505827507,
      "last_run_timestamp": "2026-08-22T17:16:03.168Z"
    },
    {
      "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": "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8608830257913248,
      "accuracy_std_dev": 0.018271073393392782,
      "run_count": 13,
      "total_questions": 2559,
      "avg_cost_per_task_usd": 0.06846565533411489,
      "avg_duration_per_task_ms": 377785.2809691286,
      "avg_output_tokens_per_task": 21384.166471277844,
      "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-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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Novita",
      "accuracy": 0.860524491362764,
      "accuracy_std_dev": 0.01647408876083893,
      "run_count": 8,
      "total_questions": 1563,
      "avg_cost_per_task_usd": 0.05773977287268074,
      "avg_duration_per_task_ms": 575351.7728726808,
      "avg_output_tokens_per_task": 23045.906589891234,
      "last_run_timestamp": "2026-08-01T17:46:48.312Z"
    },
    {
      "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": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8603745,
      "accuracy_std_dev": 0.008427361219563313,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.8821099195075758,
      "avg_duration_per_task_ms": 416796.10890151514,
      "avg_output_tokens_per_task": 33415.74179292929,
      "last_run_timestamp": "2026-09-26T09:19:04.990Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "auto-routing",
      "accuracy": 0.860269,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.20759804545454547,
      "avg_duration_per_task_ms": 271144.0252525252,
      "avg_output_tokens_per_task": 20665.12626262626,
      "last_run_timestamp": "2026-09-01T13:31:11.391Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.8601822310030396,
      "accuracy_std_dev": 0.013547575585321513,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.007783161094224923,
      "avg_duration_per_task_ms": 526386.7112462006,
      "avg_output_tokens_per_task": 27793.246200607904,
      "last_run_timestamp": "2026-08-29T18:49:16.416Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.859638,
      "accuracy_std_dev": 0.008071103007024502,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.815575909090909,
      "avg_duration_per_task_ms": 465568.93876262626,
      "avg_output_tokens_per_task": 32451.6672979798,
      "last_run_timestamp": "2026-09-26T09:01:02.717Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.8594275,
      "accuracy_std_dev": 0.013424092492604471,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0069255740740740744,
      "avg_duration_per_task_ms": 366969.99410774413,
      "avg_output_tokens_per_task": 23129.368686868685,
      "last_run_timestamp": "2026-09-26T09:34:13.855Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepSeek",
      "accuracy": 0.85925909122807,
      "accuracy_std_dev": 0.016791103126745952,
      "run_count": 13,
      "total_questions": 2565,
      "avg_cost_per_task_usd": 0.016009586354775826,
      "avg_duration_per_task_ms": 298851.8736842105,
      "avg_output_tokens_per_task": 18611.48382066277,
      "last_run_timestamp": "2026-08-08T09:24:32.269Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "auto-routing",
      "accuracy": 0.8587961250000001,
      "accuracy_std_dev": 0.026570942984948328,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0075656887626262635,
      "avg_duration_per_task_ms": 497586.5126262626,
      "avg_output_tokens_per_task": 24566.76767676768,
      "last_run_timestamp": "2026-09-26T08:53:09.295Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Novita",
      "accuracy": 0.8584595735950045,
      "accuracy_std_dev": 0.02051306669344079,
      "run_count": 17,
      "total_questions": 3363,
      "avg_cost_per_task_usd": 0.0565897927445733,
      "avg_duration_per_task_ms": 340557.95628902764,
      "avg_output_tokens_per_task": 21041.84418673803,
      "last_run_timestamp": "2026-08-29T19:21:48.221Z"
    },
    {
      "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": 0.8580247037037037,
      "accuracy_std_dev": 0.03117987540449769,
      "run_count": 15,
      "total_questions": 2916,
      "avg_cost_per_task_usd": 0.06713021399176955,
      "avg_duration_per_task_ms": 850575.8528806585,
      "avg_output_tokens_per_task": 29951.768861454046,
      "last_run_timestamp": "2026-08-29T18:36:10.818Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Alibaba",
      "accuracy": 0.8580246666666665,
      "accuracy_std_dev": 0.007694244595078101,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.06644984511784513,
      "avg_duration_per_task_ms": 213157.5361952862,
      "avg_output_tokens_per_task": 10806.75925925926,
      "last_run_timestamp": "2026-09-26T09:18:24.826Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.857954375,
      "accuracy_std_dev": 0.013038726633930733,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.7458450347222222,
      "avg_duration_per_task_ms": 493231.4678030303,
      "avg_output_tokens_per_task": 29674.352904040403,
      "last_run_timestamp": "2026-09-26T09:34:52.748Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Inceptron",
      "accuracy": 0.8577439999999998,
      "accuracy_std_dev": 0.02077245929590428,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.12316677878787877,
      "avg_duration_per_task_ms": 891518.1712121213,
      "avg_output_tokens_per_task": 39909.007575757576,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Venice",
      "accuracy": 0.8577028343558282,
      "accuracy_std_dev": 0.011904490003355866,
      "run_count": 5,
      "total_questions": 978,
      "avg_cost_per_task_usd": 0.14577651738241312,
      "avg_duration_per_task_ms": 258434.56646216768,
      "avg_output_tokens_per_task": 28881.36707566462,
      "last_run_timestamp": "2026-09-26T09:21:12.329Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.8575786315028903,
      "accuracy_std_dev": 0.022850648176995846,
      "run_count": 21,
      "total_questions": 4152,
      "avg_cost_per_task_usd": 0.03863133718689789,
      "avg_duration_per_task_ms": 314388.73578998074,
      "avg_output_tokens_per_task": 13377.785886319845,
      "last_run_timestamp": "2026-08-29T19:19:13.521Z"
    },
    {
      "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": "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": "Decart",
      "accuracy": 0.857191527607362,
      "accuracy_std_dev": 0.013500948849617942,
      "run_count": 5,
      "total_questions": 978,
      "avg_cost_per_task_usd": 0.1052768057259714,
      "avg_duration_per_task_ms": 186949.86605316974,
      "avg_output_tokens_per_task": 27345.398773006134,
      "last_run_timestamp": "2026-09-26T08:51:46.229Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.8571427393675027,
      "accuracy_std_dev": 0.024697655769633415,
      "run_count": 14,
      "total_questions": 2751,
      "avg_cost_per_task_usd": 0.08261968629589239,
      "avg_duration_per_task_ms": 564800.7659033079,
      "avg_output_tokens_per_task": 35564.60014540167,
      "last_run_timestamp": "2026-09-26T08:59:56.037Z"
    },
    {
      "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": 0.8569615784966699,
      "accuracy_std_dev": 0.053359451282668766,
      "run_count": 16,
      "total_questions": 3153,
      "avg_cost_per_task_usd": 0.004682514747859181,
      "avg_duration_per_task_ms": 225937.11703139867,
      "avg_output_tokens_per_task": 16787.288931176656,
      "last_run_timestamp": "2026-08-29T18:30:11.054Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.8569022,
      "accuracy_std_dev": 0.010212534228094395,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.8410627575757575,
      "avg_duration_per_task_ms": 450697.24494949495,
      "avg_output_tokens_per_task": 33467.82424242424,
      "last_run_timestamp": "2026-09-26T09:29:09.285Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "auto-routing",
      "accuracy": 0.8565655999999999,
      "accuracy_std_dev": 0.01503514097838793,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.09365999898989899,
      "avg_duration_per_task_ms": 671877.2212121212,
      "avg_output_tokens_per_task": 28686.372727272726,
      "last_run_timestamp": "2026-09-26T08:52:43.449Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Z.AI",
      "accuracy": 0.8564577918781726,
      "accuracy_std_dev": 0.028181512947320644,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.10150419204737732,
      "avg_duration_per_task_ms": 330925.20050761424,
      "avg_output_tokens_per_task": 22836.06937394247,
      "last_run_timestamp": "2026-09-26T09:19:52.896Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "BaseTen",
      "accuracy": 0.8563503216174183,
      "accuracy_std_dev": 0.016757180106276432,
      "run_count": 17,
      "total_questions": 3215,
      "avg_cost_per_task_usd": 0.1523061856920684,
      "avg_duration_per_task_ms": 200594.1632970451,
      "avg_output_tokens_per_task": 27265.382270606533,
      "last_run_timestamp": "2026-09-26T09:17:01.108Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Together",
      "accuracy": 0.8563168223350255,
      "accuracy_std_dev": 0.012593349140849445,
      "run_count": 6,
      "total_questions": 1182,
      "avg_cost_per_task_usd": 0.01789011252115059,
      "avg_duration_per_task_ms": 343652.3730964467,
      "avg_output_tokens_per_task": 35569.74111675127,
      "last_run_timestamp": "2026-09-26T08:50:11.879Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "CoreWeave",
      "accuracy": 0.8562288,
      "accuracy_std_dev": 0.011497073373689477,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.1913848090909091,
      "avg_duration_per_task_ms": 350329.27575757576,
      "avg_output_tokens_per_task": 55974.5,
      "last_run_timestamp": "2026-08-29T18:20:49.176Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Wafer",
      "accuracy": 0.85590693376174,
      "accuracy_std_dev": 0.1796436210829807,
      "run_count": 12,
      "total_questions": 2023,
      "avg_cost_per_task_usd": 0.14563704152249132,
      "avg_duration_per_task_ms": 271299.9115175482,
      "avg_output_tokens_per_task": 11255.300543746911,
      "last_run_timestamp": "2026-09-26T09:35:00.392Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.8555556,
      "accuracy_std_dev": 0.0176375124050984,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.011364495959595959,
      "avg_duration_per_task_ms": 652022.0676767677,
      "avg_output_tokens_per_task": 33080.43131313131,
      "last_run_timestamp": "2026-09-26T08:57:03.952Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "StreamLake",
      "accuracy": 0.8553997803738318,
      "accuracy_std_dev": 0.03094177075749652,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.06754141316199376,
      "avg_duration_per_task_ms": 582176.5778816199,
      "avg_output_tokens_per_task": 42569.91900311527,
      "last_run_timestamp": "2026-09-26T09:04:08.258Z"
    },
    {
      "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": 0.8553822048846674,
      "accuracy_std_dev": 0.018862412436841176,
      "run_count": 8,
      "total_questions": 1474,
      "avg_cost_per_task_usd": 0.05061357462686567,
      "avg_duration_per_task_ms": 845573.0067842605,
      "avg_output_tokens_per_task": 25929.323609226594,
      "last_run_timestamp": "2026-07-18T16:22:39.277Z"
    },
    {
      "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": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8547375714285715,
      "accuracy_std_dev": 0.04301245273458486,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.1766843759018759,
      "avg_duration_per_task_ms": 45692.21067821068,
      "avg_output_tokens_per_task": 3336.1709956709956,
      "last_run_timestamp": "2026-09-26T09:24:09.411Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "Near AI",
      "accuracy": 0.8543179200968524,
      "accuracy_std_dev": 0.02075318154457823,
      "run_count": 3,
      "total_questions": 413,
      "avg_cost_per_task_usd": 0.004252978208232445,
      "avg_duration_per_task_ms": 444698.63196125906,
      "avg_output_tokens_per_task": 8315.878934624698,
      "last_run_timestamp": "2026-09-26T09:19:00.094Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Azure",
      "accuracy": 0.8541664999999999,
      "accuracy_std_dev": 0.020077776601257404,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.8391522569444444,
      "avg_duration_per_task_ms": 453772.9583333333,
      "avg_output_tokens_per_task": 33398.183712121216,
      "last_run_timestamp": "2026-09-26T09:34:32.494Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Parasail",
      "accuracy": 0.8541419724300905,
      "accuracy_std_dev": 0.03420096150574074,
      "run_count": 13,
      "total_questions": 2539,
      "avg_cost_per_task_usd": 0.17093107877116975,
      "avg_duration_per_task_ms": 297710.6187475384,
      "avg_output_tokens_per_task": 38736.540764080346,
      "last_run_timestamp": "2026-09-26T08:54:40.874Z"
    },
    {
      "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": 0.8539230009478672,
      "accuracy_std_dev": 0.01405882861282702,
      "run_count": 16,
      "total_questions": 3165,
      "avg_cost_per_task_usd": 0.01950426066350711,
      "avg_duration_per_task_ms": 171550.25529225907,
      "avg_output_tokens_per_task": 12814.184834123223,
      "last_run_timestamp": "2026-08-01T17:57:47.536Z"
    },
    {
      "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": 0.8537537303044951,
      "accuracy_std_dev": 0.0606172032579932,
      "run_count": 21,
      "total_questions": 4138,
      "avg_cost_per_task_usd": 0.045678762928951185,
      "avg_duration_per_task_ms": 198289.18946350896,
      "avg_output_tokens_per_task": 12610.571532141132,
      "last_run_timestamp": "2026-08-29T18:30:43.161Z"
    },
    {
      "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": 0.8536342352941175,
      "accuracy_std_dev": 0.01673036914795934,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.013844663398692813,
      "avg_duration_per_task_ms": 136198.07783719548,
      "avg_output_tokens_per_task": 11446.42899584076,
      "last_run_timestamp": "2026-08-01T17:57:21.740Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.8535522308340354,
      "accuracy_std_dev": 0.010038567733441294,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.01690193850042123,
      "avg_duration_per_task_ms": 472592.7733782645,
      "avg_output_tokens_per_task": 33567.315080033695,
      "last_run_timestamp": "2026-09-26T08:54:58.400Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Minimax",
      "accuracy": 0.8529878011634057,
      "accuracy_std_dev": 0.011201391026239146,
      "run_count": 29,
      "total_questions": 5673,
      "avg_cost_per_task_usd": 0.04108059227921735,
      "avg_duration_per_task_ms": 443819.24202362064,
      "avg_output_tokens_per_task": 23556.23320994183,
      "last_run_timestamp": "2026-08-29T19:16:17.710Z"
    },
    {
      "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": "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Ionstream",
      "accuracy": 0.851852,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.16585097979797978,
      "avg_duration_per_task_ms": 389107.60101010103,
      "avg_output_tokens_per_task": 37536.631313131315,
      "last_run_timestamp": "2026-08-04T03:12:10.622Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Azure",
      "accuracy": 0.851851625,
      "accuracy_std_dev": 0.018151435989595288,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.7516495170454545,
      "avg_duration_per_task_ms": 500493.72222222225,
      "avg_output_tokens_per_task": 29958.29797979798,
      "last_run_timestamp": "2026-09-26T08:55:10.633Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cloudflare",
      "accuracy": 0.851679139925373,
      "accuracy_std_dev": 0.03687109443675745,
      "run_count": 28,
      "total_questions": 4824,
      "avg_cost_per_task_usd": 0.011843935737976785,
      "avg_duration_per_task_ms": 411338.9653814262,
      "avg_output_tokens_per_task": 16150.16169154229,
      "last_run_timestamp": "2026-09-26T08:52:37.124Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Novita",
      "accuracy": 0.8516331806375443,
      "accuracy_std_dev": 0.05745577028591585,
      "run_count": 13,
      "total_questions": 2541,
      "avg_cost_per_task_usd": 0.06650903935458481,
      "avg_duration_per_task_ms": 606257.2499016136,
      "avg_output_tokens_per_task": 37357.11216056671,
      "last_run_timestamp": "2026-09-26T09:41:13.780Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8516111428571428,
      "accuracy_std_dev": 0.032073947329028685,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.831446111111111,
      "avg_duration_per_task_ms": 427691.9718614719,
      "avg_output_tokens_per_task": 33200.131313131315,
      "last_run_timestamp": "2026-09-26T08:56:21.997Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Anthropic",
      "accuracy": 0.8511034444444444,
      "accuracy_std_dev": 0.03586305000387234,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.17666739506172838,
      "avg_duration_per_task_ms": 44619.677328843994,
      "avg_output_tokens_per_task": 3328.8120089786757,
      "last_run_timestamp": "2026-09-26T08:53:27.502Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Azure",
      "accuracy": 0.8510597058823527,
      "accuracy_std_dev": 0.02313439249450838,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.1586523764111705,
      "avg_duration_per_task_ms": 153997.80986333927,
      "avg_output_tokens_per_task": 15572.684789067142,
      "last_run_timestamp": "2026-09-26T09:21:26.479Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.8510099000000001,
      "accuracy_std_dev": 0.015100002764569275,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.7538109368686869,
      "avg_duration_per_task_ms": 501864.8787878788,
      "avg_output_tokens_per_task": 30068.36717171717,
      "last_run_timestamp": "2026-09-26T09:17:18.349Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8508604502868068,
      "accuracy_std_dev": 0.02478573563844512,
      "run_count": 6,
      "total_questions": 1046,
      "avg_cost_per_task_usd": 0.010149271510516252,
      "avg_duration_per_task_ms": 344062.4971319312,
      "avg_output_tokens_per_task": 20091.341300191205,
      "last_run_timestamp": "2026-09-26T09:16:37.664Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8507282936078008,
      "accuracy_std_dev": 0.03297843398312891,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.05085055074033947,
      "avg_duration_per_task_ms": 213234.5695196822,
      "avg_output_tokens_per_task": 21095.91910436981,
      "last_run_timestamp": "2026-09-26T09:01:34.657Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Google",
      "accuracy": 0.8506943125000002,
      "accuracy_std_dev": 0.0170931360994653,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.8004648680555555,
      "avg_duration_per_task_ms": 479005.81407828286,
      "avg_output_tokens_per_task": 30341.69191919192,
      "last_run_timestamp": "2026-09-26T09:19:41.155Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Novita",
      "accuracy": 0.8505651288135594,
      "accuracy_std_dev": 0.014603362238425309,
      "run_count": 6,
      "total_questions": 1180,
      "avg_cost_per_task_usd": 0.008425746610169493,
      "avg_duration_per_task_ms": 368412.6118644068,
      "avg_output_tokens_per_task": 24250.365254237287,
      "last_run_timestamp": "2026-09-26T09:22:08.527Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8505522105263156,
      "accuracy_std_dev": 0.04204091224118485,
      "run_count": 13,
      "total_questions": 2565,
      "avg_cost_per_task_usd": 0.11401860623781677,
      "avg_duration_per_task_ms": 498314.546588694,
      "avg_output_tokens_per_task": 33037.62729044834,
      "last_run_timestamp": "2026-09-26T08:50:56.054Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.8503937480314961,
      "accuracy_std_dev": 0.00023499999999998522,
      "run_count": 2,
      "total_questions": 381,
      "avg_cost_per_task_usd": 0.023863905511811022,
      "avg_duration_per_task_ms": 137111.656167979,
      "avg_output_tokens_per_task": 36528.412073490814,
      "last_run_timestamp": "2026-09-26T09:52:38.111Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8501715195876289,
      "accuracy_std_dev": 0.027779548602524146,
      "run_count": 5,
      "total_questions": 970,
      "avg_cost_per_task_usd": 0.13975330515463918,
      "avg_duration_per_task_ms": 383469.5556701031,
      "avg_output_tokens_per_task": 31875.5793814433,
      "last_run_timestamp": "2026-09-26T08:53:10.326Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "auto-routing",
      "accuracy": 0.8501683333333333,
      "accuracy_std_dev": 0.041790840030375614,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.17284996857463525,
      "avg_duration_per_task_ms": 43624.98316498317,
      "avg_output_tokens_per_task": 3253.4483726150393,
      "last_run_timestamp": "2026-09-26T08:54:57.321Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.8498074999999999,
      "accuracy_std_dev": 0.008564809236054237,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.06874417352092352,
      "avg_duration_per_task_ms": 217155.5533910534,
      "avg_output_tokens_per_task": 21389.247474747473,
      "last_run_timestamp": "2026-09-26T08:53:37.648Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Phala",
      "accuracy": 0.8498050435706695,
      "accuracy_std_dev": 0.03351477910773096,
      "run_count": 5,
      "total_questions": 941,
      "avg_cost_per_task_usd": 0.11714972582359193,
      "avg_duration_per_task_ms": 505045.9829968119,
      "avg_output_tokens_per_task": 33189.51115834219,
      "last_run_timestamp": "2026-09-26T08:54:17.442Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Chutes",
      "accuracy": 0.8497266914754098,
      "accuracy_std_dev": 0.0316272928565439,
      "run_count": 16,
      "total_questions": 3050,
      "avg_cost_per_task_usd": 0.10351839540983607,
      "avg_duration_per_task_ms": 840685.2724590164,
      "avg_output_tokens_per_task": 32158.24557377049,
      "last_run_timestamp": "2026-08-29T18:13:11.491Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Baidu",
      "accuracy": 0.8495983231939164,
      "accuracy_std_dev": 0.010674015891406563,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.09633560709759188,
      "avg_duration_per_task_ms": 314235.76045627374,
      "avg_output_tokens_per_task": 30867.523447401774,
      "last_run_timestamp": "2026-09-26T09:23:32.894Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "StreamLake",
      "accuracy": 0.8492335655172413,
      "accuracy_std_dev": 0.026170202058877375,
      "run_count": 9,
      "total_questions": 1740,
      "avg_cost_per_task_usd": 0.029375192528735626,
      "avg_duration_per_task_ms": 542113.9201149425,
      "avg_output_tokens_per_task": 28046.42528735632,
      "last_run_timestamp": "2026-09-26T08:57:03.866Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Z.AI",
      "accuracy": 0.8489460772833725,
      "accuracy_std_dev": 0.06927676739771367,
      "run_count": 13,
      "total_questions": 2562,
      "avg_cost_per_task_usd": 0.15593850858704136,
      "avg_duration_per_task_ms": 473082.737704918,
      "avg_output_tokens_per_task": 35500.65378610461,
      "last_run_timestamp": "2026-09-26T08:52:39.171Z"
    },
    {
      "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": 0.848817300658376,
      "accuracy_std_dev": 0.015530404525060451,
      "run_count": 21,
      "total_questions": 4101,
      "avg_cost_per_task_usd": 0.05035830358449157,
      "avg_duration_per_task_ms": 546041.7607900512,
      "avg_output_tokens_per_task": 25470.43672275055,
      "last_run_timestamp": "2026-08-30T00:19:02.095Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.8486918037542662,
      "accuracy_std_dev": 0.014893804491204496,
      "run_count": 6,
      "total_questions": 1172,
      "avg_cost_per_task_usd": 0.017789553754266208,
      "avg_duration_per_task_ms": 299607.09129692835,
      "avg_output_tokens_per_task": 35353.55375426621,
      "last_run_timestamp": "2026-09-26T08:54:42.351Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8486816067172265,
      "accuracy_std_dev": 0.00921244198054446,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.0456744940411701,
      "avg_duration_per_task_ms": 261759.05597688697,
      "avg_output_tokens_per_task": 41683.36944745395,
      "last_run_timestamp": "2026-09-26T16:05:38.016Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Sail Research",
      "accuracy": 0.8482611842239186,
      "accuracy_std_dev": 0.011728653128556577,
      "run_count": 10,
      "total_questions": 1965,
      "avg_cost_per_task_usd": 0.12833305750636134,
      "avg_duration_per_task_ms": 389662.5424936387,
      "avg_output_tokens_per_task": 33100.7262086514,
      "last_run_timestamp": "2026-09-26T08:55:48.508Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Alibaba",
      "accuracy": 0.848238375,
      "accuracy_std_dev": 0.029464977306795932,
      "run_count": 20,
      "total_questions": 3936,
      "avg_cost_per_task_usd": 0.22566900152439032,
      "avg_duration_per_task_ms": 509900.2350101626,
      "avg_output_tokens_per_task": 42987.483485772354,
      "last_run_timestamp": "2026-09-26T08:55:16.971Z"
    },
    {
      "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": 0.8481524347826088,
      "accuracy_std_dev": 0.010044369703747568,
      "run_count": 13,
      "total_questions": 2553,
      "avg_cost_per_task_usd": 0.026485490795142973,
      "avg_duration_per_task_ms": 580834.6169212691,
      "avg_output_tokens_per_task": 27399.61222091657,
      "last_run_timestamp": "2026-09-26T09:15:06.166Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "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": "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Parasail",
      "accuracy": 0.8479360327102803,
      "accuracy_std_dev": 0.008095688383477137,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.03592315147975078,
      "avg_duration_per_task_ms": 355113.16238317755,
      "avg_output_tokens_per_task": 36306.2511682243,
      "last_run_timestamp": "2026-09-26T16:12:42.605Z"
    },
    {
      "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": 0.8476976540731995,
      "accuracy_std_dev": 0.03320482882435078,
      "run_count": 13,
      "total_questions": 2541,
      "avg_cost_per_task_usd": 0.08207621723730815,
      "avg_duration_per_task_ms": 875595.4120425029,
      "avg_output_tokens_per_task": 34778.203856749315,
      "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-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.847643,
      "accuracy_std_dev": 0.01430999999999999,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.02130802777777778,
      "avg_duration_per_task_ms": 487650.29545454547,
      "avg_output_tokens_per_task": 42398.98737373737,
      "last_run_timestamp": "2026-09-26T09:15:51.836Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8475419999999999,
      "accuracy_std_dev": 0.02235490172109912,
      "run_count": 25,
      "total_questions": 4950,
      "avg_cost_per_task_usd": 0.16931346101010097,
      "avg_duration_per_task_ms": 157193.26565656567,
      "avg_output_tokens_per_task": 15870.87696969697,
      "last_run_timestamp": "2026-09-26T09:21:17.407Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Friendli",
      "accuracy": 0.8473202765017668,
      "accuracy_std_dev": 0.011288322955111134,
      "run_count": 6,
      "total_questions": 1132,
      "avg_cost_per_task_usd": 0.016132280918727912,
      "avg_duration_per_task_ms": 230171.71908127208,
      "avg_output_tokens_per_task": 32052.750883392226,
      "last_run_timestamp": "2026-09-26T08:49:42.815Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Together",
      "accuracy": 0.8472220833333332,
      "accuracy_std_dev": 0.01989710415051367,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.02657060563973064,
      "avg_duration_per_task_ms": 47738.242424242424,
      "avg_output_tokens_per_task": 7342.873316498317,
      "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": "Inceptron",
      "accuracy": 0.8470345574324325,
      "accuracy_std_dev": 0.011874871521592171,
      "run_count": 18,
      "total_questions": 3552,
      "avg_cost_per_task_usd": 0.011837505067567568,
      "avg_duration_per_task_ms": 328739.2460585586,
      "avg_output_tokens_per_task": 13305.292511261261,
      "last_run_timestamp": "2026-08-15T14:57:08.773Z"
    },
    {
      "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": 0.8469764115044247,
      "accuracy_std_dev": 0.012587309196211092,
      "run_count": 14,
      "total_questions": 2712,
      "avg_cost_per_task_usd": 0.02411464749262537,
      "avg_duration_per_task_ms": 588302.2522123894,
      "avg_output_tokens_per_task": 27531.24004424779,
      "last_run_timestamp": "2026-09-26T09:22:28.862Z"
    },
    {
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Alibaba",
      "accuracy": 0.8462764299380979,
      "accuracy_std_dev": 0.015770782183096275,
      "run_count": 9,
      "total_questions": 1777,
      "avg_cost_per_task_usd": 0.08913706978052896,
      "avg_duration_per_task_ms": 416673.65053460887,
      "avg_output_tokens_per_task": 22175.449071468767,
      "last_run_timestamp": "2026-09-26T09:15:30.970Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Makora",
      "accuracy": 0.8460472371705964,
      "accuracy_std_dev": 0.020750268207603947,
      "run_count": 4,
      "total_questions": 721,
      "avg_cost_per_task_usd": 0.1417408987517337,
      "avg_duration_per_task_ms": 223845.73370319002,
      "avg_output_tokens_per_task": 32811.28294036061,
      "last_run_timestamp": "2026-09-26T09:01:35.352Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "DekaLLM",
      "accuracy": 0.8459595,
      "accuracy_std_dev": 0.002525500000000014,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0029617095959595958,
      "avg_duration_per_task_ms": 113309.4494949495,
      "avg_output_tokens_per_task": 2921.4873737373737,
      "last_run_timestamp": "2026-05-16T08:55:42.302Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Azure",
      "accuracy": 0.8458658888888889,
      "accuracy_std_dev": 0.03972899765618852,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.17558623344556676,
      "avg_duration_per_task_ms": 45322.60044893378,
      "avg_output_tokens_per_task": 3305.652076318743,
      "last_run_timestamp": "2026-09-26T08:54:44.118Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Anthropic",
      "accuracy": 0.8458657777777778,
      "accuracy_std_dev": 0.025409284277723633,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.15791502805836138,
      "avg_duration_per_task_ms": 153529.50112233445,
      "avg_output_tokens_per_task": 15577.850168350169,
      "last_run_timestamp": "2026-09-26T09:19:06.330Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "GMICloud",
      "accuracy": 0.8451178333333332,
      "accuracy_std_dev": 0.062380510090403164,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.10489853282828282,
      "avg_duration_per_task_ms": 490349.9595959596,
      "avg_output_tokens_per_task": 26821.65404040404,
      "last_run_timestamp": "2026-09-26T08:57:17.745Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "GMICloud",
      "accuracy": 0.8451176666666667,
      "accuracy_std_dev": 0.014740528220597193,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1084811388888889,
      "avg_duration_per_task_ms": 330556.0925925926,
      "avg_output_tokens_per_task": 28808.55387205387,
      "last_run_timestamp": "2026-09-26T09:15:10.787Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "AkashML",
      "accuracy": 0.8448117272727271,
      "accuracy_std_dev": 0.03913291354785384,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.0030450344352617085,
      "avg_duration_per_task_ms": 533396.9063360882,
      "avg_output_tokens_per_task": 12009.300275482094,
      "last_run_timestamp": "2026-07-18T08:42:46.001Z"
    },
    {
      "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": 0.8445602826784282,
      "accuracy_std_dev": 0.023734129927349024,
      "run_count": 13,
      "total_questions": 2494,
      "avg_cost_per_task_usd": 0.14965243223736974,
      "avg_duration_per_task_ms": 310432.4819566961,
      "avg_output_tokens_per_task": 38965.461106655974,
      "last_run_timestamp": "2026-09-26T08:55:26.766Z"
    },
    {
      "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": "Mistral",
      "accuracy": 0.8443695555555555,
      "accuracy_std_dev": 0.00866138720107314,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.14565384680134677,
      "avg_duration_per_task_ms": 171228.07014590348,
      "avg_output_tokens_per_task": 32219.744668911335,
      "last_run_timestamp": "2026-09-26T09:17:52.892Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Relace",
      "accuracy": 0.844276,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.059444419191919194,
      "avg_duration_per_task_ms": 215120.31818181818,
      "avg_output_tokens_per_task": 34746.81060606061,
      "last_run_timestamp": "2026-09-26T10:57:32.238Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8440735980528513,
      "accuracy_std_dev": 0.022142710827570043,
      "run_count": 11,
      "total_questions": 2157,
      "avg_cost_per_task_usd": 0.1289552053778396,
      "avg_duration_per_task_ms": 442026.2132591562,
      "avg_output_tokens_per_task": 33600.40519239685,
      "last_run_timestamp": "2026-09-26T10:36:35.700Z"
    },
    {
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Friendli",
      "accuracy": 0.8431678722157092,
      "accuracy_std_dev": 0.03311619265359203,
      "run_count": 13,
      "total_questions": 2559,
      "avg_cost_per_task_usd": 0.182964666275889,
      "avg_duration_per_task_ms": 327141.2329034779,
      "avg_output_tokens_per_task": 41509.383743649865,
      "last_run_timestamp": "2026-09-26T08:56:02.276Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "WandB",
      "accuracy": 0.8430508404840484,
      "accuracy_std_dev": 0.050661991069463994,
      "run_count": 14,
      "total_questions": 2727,
      "avg_cost_per_task_usd": 0.13254691932526588,
      "avg_duration_per_task_ms": 173628.933626696,
      "avg_output_tokens_per_task": 33041.062706270626,
      "last_run_timestamp": "2026-07-18T08:26:40.483Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Fireworks",
      "accuracy": 0.842761,
      "accuracy_std_dev": 0.013366603308245514,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.11932980505050504,
      "avg_duration_per_task_ms": 317023.23434343433,
      "avg_output_tokens_per_task": 26929.522222222222,
      "last_run_timestamp": "2026-09-26T08:52:53.172Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8427608000000001,
      "accuracy_std_dev": 0.011595230077924271,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.08432281212121212,
      "avg_duration_per_task_ms": 322252.27373737376,
      "avg_output_tokens_per_task": 20908.12626262626,
      "last_run_timestamp": "2026-09-26T08:55:48.606Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "auto-routing",
      "accuracy": 0.8427522467005077,
      "accuracy_std_dev": 0.018612571921389303,
      "run_count": 15,
      "total_questions": 2955,
      "avg_cost_per_task_usd": 0.030284539763113367,
      "avg_duration_per_task_ms": 443994.4812182741,
      "avg_output_tokens_per_task": 24288.26192893401,
      "last_run_timestamp": "2026-08-29T19:06:57.766Z"
    },
    {
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Crusoe",
      "accuracy": 0.8425925,
      "accuracy_std_dev": 0.008788056170166437,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.14231570202020202,
      "avg_duration_per_task_ms": 163901.45454545456,
      "avg_output_tokens_per_task": 32169.752525252527,
      "last_run_timestamp": "2026-09-26T09:14:46.960Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Together",
      "accuracy": 0.8425191583372257,
      "accuracy_std_dev": 0.019868589783045544,
      "run_count": 11,
      "total_questions": 2141,
      "avg_cost_per_task_usd": 0.19488549135917793,
      "avg_duration_per_task_ms": 339625.2237272303,
      "avg_output_tokens_per_task": 44550.5791686128,
      "last_run_timestamp": "2026-09-26T09:23:36.793Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Z.AI",
      "accuracy": 0.842400931034483,
      "accuracy_std_dev": 0.020596427323446166,
      "run_count": 20,
      "total_questions": 3915,
      "avg_cost_per_task_usd": 0.055737818135376774,
      "avg_duration_per_task_ms": 573338.7954022988,
      "avg_output_tokens_per_task": 25592.180587484036,
      "last_run_timestamp": "2026-08-30T00:28:09.702Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Friendli",
      "accuracy": 0.8419116213235294,
      "accuracy_std_dev": 0.013588325247266868,
      "run_count": 14,
      "total_questions": 2720,
      "avg_cost_per_task_usd": 0.1566976856617647,
      "avg_duration_per_task_ms": 305686.83897058823,
      "avg_output_tokens_per_task": 35495.729044117645,
      "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Together",
      "accuracy": 0.8417577430497052,
      "accuracy_std_dev": 0.014641061103015125,
      "run_count": 6,
      "total_questions": 1187,
      "avg_cost_per_task_usd": 0.13158069755686608,
      "avg_duration_per_task_ms": 164554.15669755687,
      "avg_output_tokens_per_task": 29682.097725358046,
      "last_run_timestamp": "2026-09-26T08:53:13.623Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Relace",
      "accuracy": 0.8415826000000001,
      "accuracy_std_dev": 0.01677086684223569,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.013069406060606059,
      "avg_duration_per_task_ms": 474851.301010101,
      "avg_output_tokens_per_task": 38641.41818181818,
      "last_run_timestamp": "2026-09-26T08:53:10.669Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8414969770992364,
      "accuracy_std_dev": 0.03970141746131365,
      "run_count": 16,
      "total_questions": 3144,
      "avg_cost_per_task_usd": 0.0710119901399491,
      "avg_duration_per_task_ms": 630148.5734732824,
      "avg_output_tokens_per_task": 28421.371501272264,
      "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": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8410292857142857,
      "accuracy_std_dev": 0.03771800444544936,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.6859342640692642,
      "avg_duration_per_task_ms": 470200.2972582973,
      "avg_output_tokens_per_task": 27535.334054834057,
      "last_run_timestamp": "2026-09-26T08:58:47.240Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "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": "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Reka",
      "accuracy": 0.840404,
      "accuracy_std_dev": 0.008869764145680555,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.07076540101010101,
      "avg_duration_per_task_ms": 297317.25757575757,
      "avg_output_tokens_per_task": 20594.954545454544,
      "last_run_timestamp": "2026-09-26T08:53:05.194Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Together",
      "accuracy": 0.8402299275862066,
      "accuracy_std_dev": 0.03863386507191665,
      "run_count": 9,
      "total_questions": 1740,
      "avg_cost_per_task_usd": 0.16238871264367818,
      "avg_duration_per_task_ms": 375858.65287356323,
      "avg_output_tokens_per_task": 36786.99310344827,
      "last_run_timestamp": "2026-08-15T08:57:45.830Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.8401273660017347,
      "accuracy_std_dev": 0.00647490120044743,
      "run_count": 6,
      "total_questions": 1153,
      "avg_cost_per_task_usd": 0.022382765828274063,
      "avg_duration_per_task_ms": 71469.82740676496,
      "avg_output_tokens_per_task": 18471.962705984388,
      "last_run_timestamp": "2026-09-26T10:30:17.147Z"
    },
    {
      "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": 0.8398268571428571,
      "accuracy_std_dev": 0.01764351869205047,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.03009014502164502,
      "avg_duration_per_task_ms": 211303.82539682538,
      "avg_output_tokens_per_task": 13627.257575757576,
      "last_run_timestamp": "2026-06-06T17:25:31.571Z"
    },
    {
      "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": 0.8395753932584269,
      "accuracy_std_dev": 0.02635514356497756,
      "run_count": 11,
      "total_questions": 2136,
      "avg_cost_per_task_usd": 0.054479627808988774,
      "avg_duration_per_task_ms": 414791.2270599251,
      "avg_output_tokens_per_task": 21872.586142322096,
      "last_run_timestamp": "2026-08-01T17:53:10.275Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.8392968267419962,
      "accuracy_std_dev": 0.0037396706405778625,
      "run_count": 4,
      "total_questions": 531,
      "avg_cost_per_task_usd": 0.12392348210922788,
      "avg_duration_per_task_ms": 100276.54613935969,
      "avg_output_tokens_per_task": 27948.913370998118,
      "last_run_timestamp": "2026-09-26T09:05:37.095Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "ModelRun",
      "accuracy": 0.8390480151770657,
      "accuracy_std_dev": 0.04955882783964433,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.04416806408094435,
      "avg_duration_per_task_ms": 290822.7987633502,
      "avg_output_tokens_per_task": 23210.459808881395,
      "last_run_timestamp": "2026-07-11T08:27:26.027Z"
    },
    {
      "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": "NextBit",
      "accuracy": 0.8387204,
      "accuracy_std_dev": 0.02117180060930105,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.020740932323232323,
      "avg_duration_per_task_ms": 437294.63434343436,
      "avg_output_tokens_per_task": 37324.31616161616,
      "last_run_timestamp": "2026-09-26T09:10:17.414Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.8386324061538462,
      "accuracy_std_dev": 0.032593817304240734,
      "run_count": 15,
      "total_questions": 2925,
      "avg_cost_per_task_usd": 0.05106006564102564,
      "avg_duration_per_task_ms": 250208.00615384616,
      "avg_output_tokens_per_task": 22368.053333333333,
      "last_run_timestamp": "2026-09-26T10:00:05.608Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Morph",
      "accuracy": 0.838579522693997,
      "accuracy_std_dev": 0.017426082982771814,
      "run_count": 7,
      "total_questions": 1366,
      "avg_cost_per_task_usd": 0.0794436691068814,
      "avg_duration_per_task_ms": 602046.7005856516,
      "avg_output_tokens_per_task": 29605.27232796486,
      "last_run_timestamp": "2026-09-26T09:19:10.825Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Venice",
      "accuracy": 0.8383037142857144,
      "accuracy_std_dev": 0.03628775495967758,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.05614424146224147,
      "avg_duration_per_task_ms": 269081.58946608944,
      "avg_output_tokens_per_task": 12677.165464165464,
      "last_run_timestamp": "2026-08-29T18:50:46.982Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Google",
      "accuracy": 0.838284705882353,
      "accuracy_std_dev": 0.05170697825901454,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.18276167468805704,
      "avg_duration_per_task_ms": 40754.59982174688,
      "avg_output_tokens_per_task": 3290.764111705288,
      "last_run_timestamp": "2026-09-26T08:53:17.875Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.838277288056206,
      "accuracy_std_dev": 0.03168625568690031,
      "run_count": 13,
      "total_questions": 2562,
      "avg_cost_per_task_usd": 0.10126575644028102,
      "avg_duration_per_task_ms": 880551.862607338,
      "avg_output_tokens_per_task": 31631.52537080406,
      "last_run_timestamp": "2026-09-26T10:24:21.190Z"
    },
    {
      "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": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.8381625397574123,
      "accuracy_std_dev": 0.007901272276048144,
      "run_count": 9,
      "total_questions": 1484,
      "avg_cost_per_task_usd": 0.017624041105121297,
      "avg_duration_per_task_ms": 189126.76886792452,
      "avg_output_tokens_per_task": 35057.77156334232,
      "last_run_timestamp": "2026-09-26T09:29:08.510Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "StreamLake",
      "accuracy": 0.8380671171597633,
      "accuracy_std_dev": 0.02773812208207325,
      "run_count": 5,
      "total_questions": 845,
      "avg_cost_per_task_usd": 0.07336154792899409,
      "avg_duration_per_task_ms": 417593.0852071006,
      "avg_output_tokens_per_task": 24105.62721893491,
      "last_run_timestamp": "2026-09-26T08:54:14.754Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8377145373291274,
      "accuracy_std_dev": 0.08364258942907019,
      "run_count": 5,
      "total_questions": 951,
      "avg_cost_per_task_usd": 0.09799871188222924,
      "avg_duration_per_task_ms": 610509.8033648791,
      "avg_output_tokens_per_task": 25161.19242902208,
      "last_run_timestamp": "2026-09-26T08:53:09.416Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.8376065934065935,
      "accuracy_std_dev": 0.00571412028808169,
      "run_count": 3,
      "total_questions": 546,
      "avg_cost_per_task_usd": 0.036677783882783886,
      "avg_duration_per_task_ms": 47951.65201465201,
      "avg_output_tokens_per_task": 14583.362637362638,
      "last_run_timestamp": "2026-09-01T17:50:52.743Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8375589327830189,
      "accuracy_std_dev": 0.020056898198209345,
      "run_count": 9,
      "total_questions": 1696,
      "avg_cost_per_task_usd": 0.054286748820754724,
      "avg_duration_per_task_ms": 282159.9534198113,
      "avg_output_tokens_per_task": 22660.685141509435,
      "last_run_timestamp": "2026-09-26T09:50:39.842Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8373233845303869,
      "accuracy_std_dev": 0.00973152140401156,
      "run_count": 14,
      "total_questions": 2715,
      "avg_cost_per_task_usd": 0.06885601362799264,
      "avg_duration_per_task_ms": 716196.5233885819,
      "avg_output_tokens_per_task": 41564.71123388582,
      "last_run_timestamp": "2026-09-26T16:00:20.168Z"
    },
    {
      "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": "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "auto-routing",
      "accuracy": 0.8368515680218921,
      "accuracy_std_dev": 0.03421836730617917,
      "run_count": 13,
      "total_questions": 2558,
      "avg_cost_per_task_usd": 0.12014493940578577,
      "avg_duration_per_task_ms": 750575.0445660673,
      "avg_output_tokens_per_task": 36877.55043002345,
      "last_run_timestamp": "2026-08-29T18:43:32.988Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.8367000000000001,
      "accuracy_std_dev": 0.006734000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.06203523232323233,
      "avg_duration_per_task_ms": 92697.9696969697,
      "avg_output_tokens_per_task": 13658.787878787878,
      "last_run_timestamp": "2026-09-01T17:49:43.168Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Reka",
      "accuracy": 0.8366999999999999,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04958680555555556,
      "avg_duration_per_task_ms": 536787.6388888889,
      "avg_output_tokens_per_task": 49485.26515151515,
      "last_run_timestamp": "2026-09-26T16:26:15.411Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Crusoe",
      "accuracy": 0.8366909151079136,
      "accuracy_std_dev": 0.01869795941406443,
      "run_count": 4,
      "total_questions": 695,
      "avg_cost_per_task_usd": 0.020156113669064748,
      "avg_duration_per_task_ms": 242800.49064748202,
      "avg_output_tokens_per_task": 40096.50647482014,
      "last_run_timestamp": "2026-09-26T09:48:09.064Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DekaLLM",
      "accuracy": 0.8366301736263737,
      "accuracy_std_dev": 0.012674357340709624,
      "run_count": 3,
      "total_questions": 455,
      "avg_cost_per_task_usd": 0.04995332967032967,
      "avg_duration_per_task_ms": 107459.2945054945,
      "avg_output_tokens_per_task": 15090.074725274726,
      "last_run_timestamp": "2026-09-26T08:53:14.372Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Morph",
      "accuracy": 0.8365793063670413,
      "accuracy_std_dev": 0.023337994672784947,
      "run_count": 7,
      "total_questions": 1335,
      "avg_cost_per_task_usd": 0.01897062172284644,
      "avg_duration_per_task_ms": 793025.5857677902,
      "avg_output_tokens_per_task": 41115.47715355805,
      "last_run_timestamp": "2026-09-26T12:41:27.111Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8362919349112425,
      "accuracy_std_dev": 0.01693241709290274,
      "run_count": 18,
      "total_questions": 3549,
      "avg_cost_per_task_usd": 0.04999668385460693,
      "avg_duration_per_task_ms": 467378.05945336714,
      "avg_output_tokens_per_task": 26951.275852352774,
      "last_run_timestamp": "2026-08-22T17:01:35.600Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Azure",
      "accuracy": 0.83627925,
      "accuracy_std_dev": 0.037120416716242575,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.015161977272727275,
      "avg_duration_per_task_ms": 425327.9482323232,
      "avg_output_tokens_per_task": 27500.13005050505,
      "last_run_timestamp": "2026-08-29T19:58:27.585Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Google",
      "accuracy": 0.836234166760247,
      "accuracy_std_dev": 0.02982784551020275,
      "run_count": 27,
      "total_questions": 5343,
      "avg_cost_per_task_usd": 0.1657105358412877,
      "avg_duration_per_task_ms": 150009.68556990454,
      "avg_output_tokens_per_task": 15316.003930376193,
      "last_run_timestamp": "2026-09-26T08:59:40.295Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Nebius",
      "accuracy": 0.836139,
      "accuracy_std_dev": 0.05678065343641852,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.005011018518518518,
      "avg_duration_per_task_ms": 214818.47643097644,
      "avg_output_tokens_per_task": 17529.722222222223,
      "last_run_timestamp": "2026-09-26T11:06:41.988Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Novita",
      "accuracy": 0.8358585000000001,
      "accuracy_std_dev": 0.02753853053904659,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1297035244107744,
      "avg_duration_per_task_ms": 386511.07744107745,
      "avg_output_tokens_per_task": 35313.07323232323,
      "last_run_timestamp": "2026-09-26T08:54:10.966Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "CoreWeave",
      "accuracy": 0.8355473623188407,
      "accuracy_std_dev": 0.012606082718439233,
      "run_count": 11,
      "total_questions": 1656,
      "avg_cost_per_task_usd": 0.03472509842995169,
      "avg_duration_per_task_ms": 236602.51086956522,
      "avg_output_tokens_per_task": 27679.197463768116,
      "last_run_timestamp": "2026-09-26T16:00:25.681Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8354201847715735,
      "accuracy_std_dev": 0.021899629119538376,
      "run_count": 15,
      "total_questions": 2955,
      "avg_cost_per_task_usd": 0.1346993539763113,
      "avg_duration_per_task_ms": 635579.6727580372,
      "avg_output_tokens_per_task": 34556.54077834179,
      "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": "Parasail",
      "accuracy": 0.8353531999999999,
      "accuracy_std_dev": 0.00989691835674118,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.03142896161616161,
      "avg_duration_per_task_ms": 541746.1080808081,
      "avg_output_tokens_per_task": 15052.834343434344,
      "last_run_timestamp": "2026-06-06T16:02:18.201Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openrouter/auto-beta",
      "model_slug": "openrouter/auto-beta",
      "model_name": "Auto Router (Beta)",
      "provider_name": "auto-routing",
      "accuracy": 0.835017,
      "accuracy_std_dev": 0.062145249638783075,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.025753547138047138,
      "avg_duration_per_task_ms": 248297.88383838383,
      "avg_output_tokens_per_task": 30752.395622895623,
      "last_run_timestamp": "2026-09-15T14:04:40.150Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.835017,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.036737896464646466,
      "avg_duration_per_task_ms": 49536.31060606061,
      "avg_output_tokens_per_task": 14606.348484848484,
      "last_run_timestamp": "2026-09-01T17:49:40.120Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.8348945269320842,
      "accuracy_std_dev": 0.07887946023533471,
      "run_count": 13,
      "total_questions": 2562,
      "avg_cost_per_task_usd": 0.07108125370804061,
      "avg_duration_per_task_ms": 478944.9949258392,
      "avg_output_tokens_per_task": 31685.776346604216,
      "last_run_timestamp": "2026-09-26T09:14:27.364Z"
    },
    {
      "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": 0.8348081017699116,
      "accuracy_std_dev": 0.015122908520600433,
      "run_count": 7,
      "total_questions": 1356,
      "avg_cost_per_task_usd": 0.034566790560471974,
      "avg_duration_per_task_ms": 347006.30088495574,
      "avg_output_tokens_per_task": 13778.697640117995,
      "last_run_timestamp": "2026-06-30T22:55:56.214Z"
    },
    {
      "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": "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/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.834175,
      "accuracy_std_dev": 0.009259000000000017,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.028012333333333334,
      "avg_duration_per_task_ms": 352518.654040404,
      "avg_output_tokens_per_task": 14875.09595959596,
      "last_run_timestamp": "2026-09-26T09:19:39.685Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Venice",
      "accuracy": 0.8339535412719892,
      "accuracy_std_dev": 0.08422964195615594,
      "run_count": 15,
      "total_questions": 2956,
      "avg_cost_per_task_usd": 0.09844996684709066,
      "avg_duration_per_task_ms": 493001.6951962111,
      "avg_output_tokens_per_task": 30313.425236806495,
      "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": "Cloudflare",
      "accuracy": 0.8339505777777776,
      "accuracy_std_dev": 0.01071049161258874,
      "run_count": 6,
      "total_questions": 1080,
      "avg_cost_per_task_usd": 0.025328656481481478,
      "avg_duration_per_task_ms": 650579.8324074075,
      "avg_output_tokens_per_task": 40296.40185185185,
      "last_run_timestamp": "2026-09-26T09:00:45.770Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8338943333333333,
      "accuracy_std_dev": 0.011690806112307073,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.03773407239057239,
      "avg_duration_per_task_ms": 276396.1228956229,
      "avg_output_tokens_per_task": 28290.345117845118,
      "last_run_timestamp": "2026-09-12T09:56:44.384Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Venice",
      "accuracy": 0.8338056683673469,
      "accuracy_std_dev": 0.013414099005395167,
      "run_count": 18,
      "total_questions": 3528,
      "avg_cost_per_task_usd": 0.004950865079365079,
      "avg_duration_per_task_ms": 179413.4075963719,
      "avg_output_tokens_per_task": 16477.43820861678,
      "last_run_timestamp": "2026-08-29T18:49:32.713Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8337061610738257,
      "accuracy_std_dev": 0.0693376467642218,
      "run_count": 5,
      "total_questions": 894,
      "avg_cost_per_task_usd": 0.009961087248322148,
      "avg_duration_per_task_ms": 513095.80536912754,
      "avg_output_tokens_per_task": 19701.401565995526,
      "last_run_timestamp": "2026-09-26T08:54:24.468Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-flash-20260826",
      "model_slug": "qwen/qwen3.8-flash",
      "model_name": "Qwen: Qwen3.8 Flash",
      "provider_name": "Makora",
      "accuracy": 0.8333329999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 156,
      "avg_cost_per_task_usd": 0.003982064102564103,
      "avg_duration_per_task_ms": 34178.44871794872,
      "avg_output_tokens_per_task": 8341.692307692309,
      "last_run_timestamp": "2026-09-09T15:03:02.824Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Novita",
      "accuracy": 0.8332027746478873,
      "accuracy_std_dev": 0.026066320644252704,
      "run_count": 13,
      "total_questions": 2556,
      "avg_cost_per_task_usd": 0.06387245735524257,
      "avg_duration_per_task_ms": 894067.6103286385,
      "avg_output_tokens_per_task": 29485.598591549297,
      "last_run_timestamp": "2026-08-29T19:12:56.521Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Parasail",
      "accuracy": 0.833159912591051,
      "accuracy_std_dev": 0.01602299589215448,
      "run_count": 5,
      "total_questions": 961,
      "avg_cost_per_task_usd": 0.14537621123829345,
      "avg_duration_per_task_ms": 277832.8116545265,
      "avg_output_tokens_per_task": 32850.81373569199,
      "last_run_timestamp": "2026-09-26T09:19:06.323Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Mistral",
      "accuracy": 0.8331351764705882,
      "accuracy_std_dev": 0.011719105633588751,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.25533701039809864,
      "avg_duration_per_task_ms": 277560.3027332145,
      "avg_output_tokens_per_task": 55800.19964349376,
      "last_run_timestamp": "2026-09-26T09:16:57.240Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.8326118571428571,
      "accuracy_std_dev": 0.010792989132217443,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.039859138528138535,
      "avg_duration_per_task_ms": 221347.44805194804,
      "avg_output_tokens_per_task": 14514.689033189034,
      "last_run_timestamp": "2026-09-26T09:19:17.245Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Decart",
      "accuracy": 0.8324914999999999,
      "accuracy_std_dev": 0.000841499999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.023044303030303033,
      "avg_duration_per_task_ms": 640134.2070707071,
      "avg_output_tokens_per_task": 54011.681818181816,
      "last_run_timestamp": "2026-09-26T08:54:28.265Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Modal",
      "accuracy": 0.8323511666666668,
      "accuracy_std_dev": 0.013383536583139077,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.1359044436026936,
      "avg_duration_per_task_ms": 151136.74915824915,
      "avg_output_tokens_per_task": 30646.316498316497,
      "last_run_timestamp": "2026-09-26T08:54:57.237Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Parasail",
      "accuracy": 0.8319833976435935,
      "accuracy_std_dev": 0.0051619355292529005,
      "run_count": 7,
      "total_questions": 1358,
      "avg_cost_per_task_usd": 0.03598950073637703,
      "avg_duration_per_task_ms": 200380.98085419735,
      "avg_output_tokens_per_task": 13861.594256259204,
      "last_run_timestamp": "2026-09-26T09:19:46.321Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "auto-routing",
      "accuracy": 0.8318619561643835,
      "accuracy_std_dev": 0.01930397520188926,
      "run_count": 17,
      "total_questions": 3285,
      "avg_cost_per_task_usd": 0.003945210350076104,
      "avg_duration_per_task_ms": 335822.299238965,
      "avg_output_tokens_per_task": 10098.203652968037,
      "last_run_timestamp": "2026-08-29T19:56:51.477Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Morph",
      "accuracy": 0.8316497999999998,
      "accuracy_std_dev": 0.03697562642282075,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.01486108888888889,
      "avg_duration_per_task_ms": 312116.7202020202,
      "avg_output_tokens_per_task": 12505.067171717172,
      "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Together",
      "accuracy": 0.8316496666666666,
      "accuracy_std_dev": 0.020755514356966008,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.05561522937710438,
      "avg_duration_per_task_ms": 213796.11658249158,
      "avg_output_tokens_per_task": 17341.332912457914,
      "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-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Z.AI",
      "accuracy": 0.8315097680525164,
      "accuracy_std_dev": 0.043329048047164274,
      "run_count": 14,
      "total_questions": 2742,
      "avg_cost_per_task_usd": 0.16222809883296865,
      "avg_duration_per_task_ms": 711251.6560904449,
      "avg_output_tokens_per_task": 37157.696571845365,
      "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-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Google",
      "accuracy": 0.8313691904761906,
      "accuracy_std_dev": 0.0137097002420047,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.06259450793650795,
      "avg_duration_per_task_ms": 231605.2477152477,
      "avg_output_tokens_per_task": 28760.637566137568,
      "last_run_timestamp": "2026-08-30T00:09:46.578Z"
    },
    {
      "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Alibaba",
      "accuracy": 0.8309764,
      "accuracy_std_dev": 0.005061551248382267,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.24307877272727274,
      "avg_duration_per_task_ms": 341071.7717171717,
      "avg_output_tokens_per_task": 40681.778787878786,
      "last_run_timestamp": "2026-09-26T08:54:33.880Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8309565294117646,
      "accuracy_std_dev": 0.021130248647287467,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.03042767617349971,
      "avg_duration_per_task_ms": 128034.97237076648,
      "avg_output_tokens_per_task": 16834.001782531195,
      "last_run_timestamp": "2026-08-29T18:13:32.148Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "BaseTen",
      "accuracy": 0.8309340194750211,
      "accuracy_std_dev": 0.0227049955198312,
      "run_count": 6,
      "total_questions": 1181,
      "avg_cost_per_task_usd": 0.0975369652836579,
      "avg_duration_per_task_ms": 214278.25232853513,
      "avg_output_tokens_per_task": 28093.950889077052,
      "last_run_timestamp": "2026-08-29T18:58:14.477Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Modal",
      "accuracy": 0.8308604629080119,
      "accuracy_std_dev": 0.014193547884550253,
      "run_count": 7,
      "total_questions": 1348,
      "avg_cost_per_task_usd": 0.037891371661721074,
      "avg_duration_per_task_ms": 428497.83827893174,
      "avg_output_tokens_per_task": 44654.94213649852,
      "last_run_timestamp": "2026-09-26T08:56:01.204Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8308259809976247,
      "accuracy_std_dev": 0.05464180945926158,
      "run_count": 13,
      "total_questions": 2526,
      "avg_cost_per_task_usd": 0.04734239192399049,
      "avg_duration_per_task_ms": 550280.8665874901,
      "avg_output_tokens_per_task": 22668.492082343626,
      "last_run_timestamp": "2026-09-26T08:01:18.114Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Together",
      "accuracy": 0.8308079999999999,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.04318659848484849,
      "avg_duration_per_task_ms": 391952.8409090909,
      "avg_output_tokens_per_task": 15921.767676767677,
      "last_run_timestamp": "2026-05-16T10:24:57.481Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Baidu",
      "accuracy": 0.8302680551724138,
      "accuracy_std_dev": 0.010265872005588431,
      "run_count": 22,
      "total_questions": 4350,
      "avg_cost_per_task_usd": 0.004694437471264369,
      "avg_duration_per_task_ms": 306186.3236781609,
      "avg_output_tokens_per_task": 12163.184827586207,
      "last_run_timestamp": "2026-08-29T16:40:58.367Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "GMICloud",
      "accuracy": 0.8301106049723759,
      "accuracy_std_dev": 0.03034424675983057,
      "run_count": 11,
      "total_questions": 2172,
      "avg_cost_per_task_usd": 0.12463168692449354,
      "avg_duration_per_task_ms": 509594.1510128913,
      "avg_output_tokens_per_task": 35443.973756906074,
      "last_run_timestamp": "2026-09-26T08:54:25.911Z"
    },
    {
      "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": 0.8300985157116451,
      "accuracy_std_dev": 0.047147674805281085,
      "run_count": 33,
      "total_questions": 6492,
      "avg_cost_per_task_usd": 0.0361898150030807,
      "avg_duration_per_task_ms": 283344.9191312384,
      "avg_output_tokens_per_task": 18752.18977202711,
      "last_run_timestamp": "2026-08-29T19:17:08.288Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "auto-routing",
      "accuracy": 0.8298668701517706,
      "accuracy_std_dev": 0.04231782629659911,
      "run_count": 9,
      "total_questions": 1779,
      "avg_cost_per_task_usd": 0.16520087015177065,
      "avg_duration_per_task_ms": 161607.96964586846,
      "avg_output_tokens_per_task": 16300.88813940416,
      "last_run_timestamp": "2026-09-26T08:54:23.573Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AkashML",
      "accuracy": 0.8296198191945158,
      "accuracy_std_dev": 0.010647141525623993,
      "run_count": 6,
      "total_questions": 1167,
      "avg_cost_per_task_usd": 0.1271454147386461,
      "avg_duration_per_task_ms": 247054.93658954583,
      "avg_output_tokens_per_task": 29738.498714652957,
      "last_run_timestamp": "2026-09-26T08:53:15.658Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.8293609150259068,
      "accuracy_std_dev": 0.029508056688451866,
      "run_count": 15,
      "total_questions": 2895,
      "avg_cost_per_task_usd": 0.13191214404145077,
      "avg_duration_per_task_ms": 620405.4127806562,
      "avg_output_tokens_per_task": 37376.21899827288,
      "last_run_timestamp": "2026-08-29T18:14:11.455Z"
    },
    {
      "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": 0.8293536141356256,
      "accuracy_std_dev": 0.03173512046656424,
      "run_count": 16,
      "total_questions": 3141,
      "avg_cost_per_task_usd": 0.14317697262018464,
      "avg_duration_per_task_ms": 539342.0986946832,
      "avg_output_tokens_per_task": 35578.65934415791,
      "last_run_timestamp": "2026-08-29T18:09:15.344Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Parasail",
      "accuracy": 0.8293137510353228,
      "accuracy_std_dev": 0.13959916580818144,
      "run_count": 21,
      "total_questions": 4105,
      "avg_cost_per_task_usd": 0.04072066552984167,
      "avg_duration_per_task_ms": 187753.5790499391,
      "avg_output_tokens_per_task": 11251.794153471375,
      "last_run_timestamp": "2026-08-29T19:36:49.687Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Cloudflare",
      "accuracy": 0.8291486389610391,
      "accuracy_std_dev": 0.04390329909775197,
      "run_count": 6,
      "total_questions": 1155,
      "avg_cost_per_task_usd": 0.028787966233766234,
      "avg_duration_per_task_ms": 286215.91515151516,
      "avg_output_tokens_per_task": 8890.160173160173,
      "last_run_timestamp": "2026-09-26T09:45:13.813Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.8289048695652176,
      "accuracy_std_dev": 0.025607506073269167,
      "run_count": 21,
      "total_questions": 4140,
      "avg_cost_per_task_usd": 0.04958357536231883,
      "avg_duration_per_task_ms": 535038.0852657005,
      "avg_output_tokens_per_task": 25244.089371980677,
      "last_run_timestamp": "2026-08-30T00:30:19.495Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8287637142857143,
      "accuracy_std_dev": 0.08258088982275724,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.007342964646464646,
      "avg_duration_per_task_ms": 721777.8398268399,
      "avg_output_tokens_per_task": 30149.034632034633,
      "last_run_timestamp": "2026-09-26T11:30:22.248Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Novita",
      "accuracy": 0.8287408778346745,
      "accuracy_std_dev": 0.05300294804063007,
      "run_count": 21,
      "total_questions": 4101,
      "avg_cost_per_task_usd": 0.10757520897342113,
      "avg_duration_per_task_ms": 651824.1992197025,
      "avg_output_tokens_per_task": 34289.174347720065,
      "last_run_timestamp": "2026-08-29T18:11:51.472Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "StreamLake",
      "accuracy": 0.8286193999999999,
      "accuracy_std_dev": 0.012838925852266618,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.01266686868686869,
      "avg_duration_per_task_ms": 861046.6656565657,
      "avg_output_tokens_per_task": 37375.524242424246,
      "last_run_timestamp": "2026-09-26T08:09:25.366Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.8285513974854932,
      "accuracy_std_dev": 0.024874965961581598,
      "run_count": 16,
      "total_questions": 3102,
      "avg_cost_per_task_usd": 0.05205874306898775,
      "avg_duration_per_task_ms": 908984.3558994198,
      "avg_output_tokens_per_task": 29744.421018697616,
      "last_run_timestamp": "2026-08-22T17:50:04.003Z"
    },
    {
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Xiaomi",
      "accuracy": 0.8284124801026957,
      "accuracy_std_dev": 0.01497190145646912,
      "run_count": 12,
      "total_questions": 2337,
      "avg_cost_per_task_usd": 0.00971603380402225,
      "avg_duration_per_task_ms": 567277.3761232349,
      "avg_output_tokens_per_task": 34466.249037227215,
      "last_run_timestamp": "2026-09-26T08:54:38.079Z"
    },
    {
      "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": "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": 0.8269733333333333,
      "accuracy_std_dev": 0.015136600014534322,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.06164716273849608,
      "avg_duration_per_task_ms": 496054.37934904604,
      "avg_output_tokens_per_task": 48584.97418630752,
      "last_run_timestamp": "2026-09-26T16:19:15.650Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Together",
      "accuracy": 0.826964236582224,
      "accuracy_std_dev": 0.043702263548038585,
      "run_count": 12,
      "total_questions": 2329,
      "avg_cost_per_task_usd": 0.017868838127951914,
      "avg_duration_per_task_ms": 211852.5800772864,
      "avg_output_tokens_per_task": 15131.554744525547,
      "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-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "BaseTen",
      "accuracy": 0.8269358,
      "accuracy_std_dev": 0.006767606146932625,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.04578402121212121,
      "avg_duration_per_task_ms": 60291.652525252524,
      "avg_output_tokens_per_task": 11285.754545454545,
      "last_run_timestamp": "2026-09-26T09:19:03.901Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8266137120954002,
      "accuracy_std_dev": 0.009032395856780211,
      "run_count": 9,
      "total_questions": 1761,
      "avg_cost_per_task_usd": 0.005160955139125498,
      "avg_duration_per_task_ms": 482576.3299261783,
      "avg_output_tokens_per_task": 12303.996592844975,
      "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": "Wafer",
      "accuracy": 0.826599,
      "accuracy_std_dev": 0.002749143987983661,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.04722597138047138,
      "avg_duration_per_task_ms": 258616.0824915825,
      "avg_output_tokens_per_task": 15105.597643097643,
      "last_run_timestamp": "2026-09-26T09:35:33.510Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Crusoe",
      "accuracy": 0.8264543613323679,
      "accuracy_std_dev": 0.018188360323715465,
      "run_count": 7,
      "total_questions": 1381,
      "avg_cost_per_task_usd": 0.20616153874004342,
      "avg_duration_per_task_ms": 315117.1817523534,
      "avg_output_tokens_per_task": 46875.9225199131,
      "last_run_timestamp": "2026-09-12T08:53:53.147Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Novita",
      "accuracy": 0.8263186666666665,
      "accuracy_std_dev": 0.011640048205322105,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.04355452946127946,
      "avg_duration_per_task_ms": 227785.6026936027,
      "avg_output_tokens_per_task": 14401.111111111111,
      "last_run_timestamp": "2026-09-26T09:34:17.391Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.8262184258271078,
      "accuracy_std_dev": 0.01942525748194862,
      "run_count": 5,
      "total_questions": 937,
      "avg_cost_per_task_usd": 0.02446647598719317,
      "avg_duration_per_task_ms": 331857.0437566702,
      "avg_output_tokens_per_task": 48723.26467449306,
      "last_run_timestamp": "2026-09-26T08:54:40.393Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Crusoe",
      "accuracy": 0.8261094014447884,
      "accuracy_std_dev": 0.026551028658038853,
      "run_count": 5,
      "total_questions": 969,
      "avg_cost_per_task_usd": 0.2076026150670795,
      "avg_duration_per_task_ms": 857894.439628483,
      "avg_output_tokens_per_task": 47560.28689370485,
      "last_run_timestamp": "2026-08-29T18:53:23.077Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.8260196253164556,
      "accuracy_std_dev": 0.053781736077150054,
      "run_count": 12,
      "total_questions": 2370,
      "avg_cost_per_task_usd": 0.024082872573839665,
      "avg_duration_per_task_ms": 553189.7814345992,
      "avg_output_tokens_per_task": 25590.895780590716,
      "last_run_timestamp": "2026-09-26T09:17:20.976Z"
    },
    {
      "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": 0.8259301444575472,
      "accuracy_std_dev": 0.05179941930880393,
      "run_count": 26,
      "total_questions": 5088,
      "avg_cost_per_task_usd": 0.03116529402515724,
      "avg_duration_per_task_ms": 245803.80483490566,
      "avg_output_tokens_per_task": 15695.919025157233,
      "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": "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Together",
      "accuracy": 0.8249211384507535,
      "accuracy_std_dev": 0.0352196214425993,
      "run_count": 15,
      "total_questions": 2853,
      "avg_cost_per_task_usd": 0.08481681317910972,
      "avg_duration_per_task_ms": 388842.21381002455,
      "avg_output_tokens_per_task": 22797.929547844375,
      "last_run_timestamp": "2026-08-22T16:49:15.579Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "CoreWeave",
      "accuracy": 0.8249157142857142,
      "accuracy_std_dev": 0.019837994431424387,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.1225871738816739,
      "avg_duration_per_task_ms": 339042.2626262626,
      "avg_output_tokens_per_task": 50612.39971139971,
      "last_run_timestamp": "2026-09-26T08:56:23.149Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.8246351666666666,
      "accuracy_std_dev": 0.013862659393933843,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.034730324915824916,
      "avg_duration_per_task_ms": 240485.28282828283,
      "avg_output_tokens_per_task": 13780.222222222223,
      "last_run_timestamp": "2026-09-26T09:15:01.388Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Friendli",
      "accuracy": 0.8246351666666666,
      "accuracy_std_dev": 0.0064044884369392945,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.13257150673400672,
      "avg_duration_per_task_ms": 144313.25757575757,
      "avg_output_tokens_per_task": 32437.94191919192,
      "last_run_timestamp": "2026-09-26T08:54:42.422Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Mancer 2",
      "accuracy": 0.824495,
      "accuracy_std_dev": 0.00300556117888158,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.04055735353535353,
      "avg_duration_per_task_ms": 210547.6452020202,
      "avg_output_tokens_per_task": 15056.00883838384,
      "last_run_timestamp": "2026-09-26T09:19:00.848Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-5.4-mini-20260317",
      "model_slug": "openai/gpt-5.4-mini",
      "model_name": "OpenAI: GPT-5.4 Mini",
      "provider_name": "Azure",
      "accuracy": 0.82449475,
      "accuracy_std_dev": 0.010869783171135461,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.06350338131313131,
      "avg_duration_per_task_ms": 123563.64393939394,
      "avg_output_tokens_per_task": 13357.755050505051,
      "last_run_timestamp": "2026-09-01T17:56:39.303Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.8244158226324237,
      "accuracy_std_dev": 0.01458128650491802,
      "run_count": 19,
      "total_questions": 3738,
      "avg_cost_per_task_usd": 0.0051121361690743715,
      "avg_duration_per_task_ms": 427263.127340824,
      "avg_output_tokens_per_task": 11900.191011235955,
      "last_run_timestamp": "2026-08-29T18:40:24.101Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Chutes",
      "accuracy": 0.8237132857142857,
      "accuracy_std_dev": 0.024208164205574968,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.037303974025974025,
      "avg_duration_per_task_ms": 279740.5851370851,
      "avg_output_tokens_per_task": 14323.054112554113,
      "last_run_timestamp": "2026-09-26T09:20:53.278Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "Novita",
      "accuracy": 0.8234829675785206,
      "accuracy_std_dev": 0.014126665871322934,
      "run_count": 15,
      "total_questions": 2961,
      "avg_cost_per_task_usd": 0.013195947315096249,
      "avg_duration_per_task_ms": 394720.7257683215,
      "avg_output_tokens_per_task": 13095.860182370821,
      "last_run_timestamp": "2026-08-29T18:35:57.732Z"
    },
    {
      "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": 0.8233658066759387,
      "accuracy_std_dev": 0.08442661821970915,
      "run_count": 11,
      "total_questions": 2157,
      "avg_cost_per_task_usd": 0.09823874269819194,
      "avg_duration_per_task_ms": 882801.9360222531,
      "avg_output_tokens_per_task": 41455.85813630042,
      "last_run_timestamp": "2026-09-26T09:26:01.477Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DigitalOcean",
      "accuracy": 0.8232579938524589,
      "accuracy_std_dev": 0.05951885425846167,
      "run_count": 5,
      "total_questions": 976,
      "avg_cost_per_task_usd": 0.08256164446721312,
      "avg_duration_per_task_ms": 628781.7581967213,
      "avg_output_tokens_per_task": 23867.87807377049,
      "last_run_timestamp": "2026-09-26T09:35:44.547Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Chutes",
      "accuracy": 0.8230056978021977,
      "accuracy_std_dev": 0.043112993157555894,
      "run_count": 17,
      "total_questions": 3276,
      "avg_cost_per_task_usd": 0.09837996916971918,
      "avg_duration_per_task_ms": 887456.5876068376,
      "avg_output_tokens_per_task": 30118.1742979243,
      "last_run_timestamp": "2026-08-29T18:08:11.373Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Sail Research",
      "accuracy": 0.8228419167330677,
      "accuracy_std_dev": 0.11242608196694337,
      "run_count": 13,
      "total_questions": 2510,
      "avg_cost_per_task_usd": 0.007463425099601593,
      "avg_duration_per_task_ms": 612656.6079681275,
      "avg_output_tokens_per_task": 26820.300398406376,
      "last_run_timestamp": "2026-09-26T09:19:42.439Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8227271999999999,
      "accuracy_std_dev": 0.039294904353618175,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.0058356378787878796,
      "avg_duration_per_task_ms": 243021.81515151515,
      "avg_output_tokens_per_task": 13918.35101010101,
      "last_run_timestamp": "2026-08-01T20:03:34.768Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AkashML",
      "accuracy": 0.8226246495327102,
      "accuracy_std_dev": 0.027849632903150384,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.03971392640186916,
      "avg_duration_per_task_ms": 354960.40303738316,
      "avg_output_tokens_per_task": 41628.609034267916,
      "last_run_timestamp": "2026-09-26T16:24:56.918Z"
    },
    {
      "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": 0.8225956482593038,
      "accuracy_std_dev": 0.159966755691973,
      "run_count": 13,
      "total_questions": 2499,
      "avg_cost_per_task_usd": 0.08340546258503403,
      "avg_duration_per_task_ms": 451358.10764305724,
      "avg_output_tokens_per_task": 22627.635454181673,
      "last_run_timestamp": "2026-09-26T08:54:41.414Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.822297,
      "accuracy_std_dev": 0.014221076658725011,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.31472648260381597,
      "avg_duration_per_task_ms": 289101.1150392817,
      "avg_output_tokens_per_task": 20792.76374859708,
      "last_run_timestamp": "2026-09-26T08:55:59.561Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Cloudflare",
      "accuracy": 0.8221650670103093,
      "accuracy_std_dev": 0.01835186937268852,
      "run_count": 6,
      "total_questions": 1164,
      "avg_cost_per_task_usd": 0.13827411597938147,
      "avg_duration_per_task_ms": 261027.06529209623,
      "avg_output_tokens_per_task": 31195.84621993127,
      "last_run_timestamp": "2026-09-26T09:19:47.728Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Novita",
      "accuracy": 0.8221100000000001,
      "accuracy_std_dev": 0.019325780306224248,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.011398650392817058,
      "avg_duration_per_task_ms": 623049.0549943884,
      "avg_output_tokens_per_task": 33683.20538720539,
      "last_run_timestamp": "2026-09-26T09:34:52.654Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Google",
      "accuracy": 0.8219105872274142,
      "accuracy_std_dev": 0.01870885637365363,
      "run_count": 10,
      "total_questions": 1926,
      "avg_cost_per_task_usd": 0.054632894080996884,
      "avg_duration_per_task_ms": 274976.66355140187,
      "avg_output_tokens_per_task": 22789.01453790239,
      "last_run_timestamp": "2026-08-01T17:52:55.742Z"
    },
    {
      "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": 0.8217910573969281,
      "accuracy_std_dev": 0.01861988887693676,
      "run_count": 19,
      "total_questions": 3711,
      "avg_cost_per_task_usd": 0.10542012718943684,
      "avg_duration_per_task_ms": 593247.7383454595,
      "avg_output_tokens_per_task": 33480.1341956346,
      "last_run_timestamp": "2026-08-29T19:16:33.097Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "Novita",
      "accuracy": 0.821549,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.001158979797979798,
      "avg_duration_per_task_ms": 153563.0101010101,
      "avg_output_tokens_per_task": 18533.79292929293,
      "last_run_timestamp": "2026-09-23T18:11:58.526Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Ionstream",
      "accuracy": 0.8215488,
      "accuracy_std_dev": 0.011711955283384605,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.039391856565656566,
      "avg_duration_per_task_ms": 244469.52525252526,
      "avg_output_tokens_per_task": 15370.382828282829,
      "last_run_timestamp": "2026-09-26T09:19:26.857Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.82133825,
      "accuracy_std_dev": 0.012610410379424602,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.003432248737373737,
      "avg_duration_per_task_ms": 86387.43244949495,
      "avg_output_tokens_per_task": 11387.400883838383,
      "last_run_timestamp": "2026-06-13T08:37:07.102Z"
    },
    {
      "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "Xiaomi",
      "accuracy": 0.8210677142857142,
      "accuracy_std_dev": 0.013459362028123072,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0033843903318903318,
      "avg_duration_per_task_ms": 85055.07142857143,
      "avg_output_tokens_per_task": 11231.03823953824,
      "last_run_timestamp": "2026-06-13T08:05:57.280Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "AkashML",
      "accuracy": 0.8207635888324873,
      "accuracy_std_dev": 0.020270393054191847,
      "run_count": 7,
      "total_questions": 1379,
      "avg_cost_per_task_usd": 0.03637786294416244,
      "avg_duration_per_task_ms": 250330.74111675128,
      "avg_output_tokens_per_task": 14481.398839738942,
      "last_run_timestamp": "2026-09-26T08:56:59.861Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-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-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Novita",
      "accuracy": 0.8206795856832972,
      "accuracy_std_dev": 0.013489140010307805,
      "run_count": 21,
      "total_questions": 4149,
      "avg_cost_per_task_usd": 0.011921808146541337,
      "avg_duration_per_task_ms": 368546.352374066,
      "avg_output_tokens_per_task": 11799.792239093758,
      "last_run_timestamp": "2026-08-30T00:16:38.394Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Venice",
      "accuracy": 0.8206135555555556,
      "accuracy_std_dev": 0.02409886708315059,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.03784499887766555,
      "avg_duration_per_task_ms": 323218.42199775536,
      "avg_output_tokens_per_task": 39347.21212121212,
      "last_run_timestamp": "2026-09-26T16:29:49.007Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "SambaNova",
      "accuracy": 0.8201054571428571,
      "accuracy_std_dev": 0.026122159766927572,
      "run_count": 3,
      "total_questions": 315,
      "avg_cost_per_task_usd": 0.055772555555555554,
      "avg_duration_per_task_ms": 137698.42222222223,
      "avg_output_tokens_per_task": 23066.479365079365,
      "last_run_timestamp": "2026-09-12T16:00:25.695Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "StreamLake",
      "accuracy": 0.819865,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.13854771717171718,
      "avg_duration_per_task_ms": 540824.1212121212,
      "avg_output_tokens_per_task": 31240.651515151516,
      "last_run_timestamp": "2026-09-07T06:00:57.184Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8197154974692696,
      "accuracy_std_dev": 0.01804488276618475,
      "run_count": 21,
      "total_questions": 4149,
      "avg_cost_per_task_usd": 0.004644195468787659,
      "avg_duration_per_task_ms": 274482.4740901422,
      "avg_output_tokens_per_task": 11932.80212099301,
      "last_run_timestamp": "2026-08-29T19:19:45.463Z"
    },
    {
      "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": 0.8194444939024391,
      "accuracy_std_dev": 0.0983013278247835,
      "run_count": 15,
      "total_questions": 2952,
      "avg_cost_per_task_usd": 0.04524065040650407,
      "avg_duration_per_task_ms": 186780.3099593496,
      "avg_output_tokens_per_task": 21656.31910569106,
      "last_run_timestamp": "2026-09-26T09:18:49.857Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Venice",
      "accuracy": 0.8193043333333333,
      "accuracy_std_dev": 0.0031744380430068155,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.14333356734006736,
      "avg_duration_per_task_ms": 321739.1531986532,
      "avg_output_tokens_per_task": 28661.96127946128,
      "last_run_timestamp": "2026-09-26T09:35:22.612Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Z.AI",
      "accuracy": 0.818474192504259,
      "accuracy_std_dev": 0.05364142082052881,
      "run_count": 18,
      "total_questions": 3522,
      "avg_cost_per_task_usd": 0.11224730863145937,
      "avg_duration_per_task_ms": 720587.8109028961,
      "avg_output_tokens_per_task": 35871.42277115276,
      "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Baidu",
      "accuracy": 0.8183890113981765,
      "accuracy_std_dev": 0.04470928783762385,
      "run_count": 20,
      "total_questions": 3948,
      "avg_cost_per_task_usd": 0.07126378191489362,
      "avg_duration_per_task_ms": 604114.2841945288,
      "avg_output_tokens_per_task": 32383.42781155015,
      "last_run_timestamp": "2026-08-29T18:40:30.684Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Together",
      "accuracy": 0.8181816249999999,
      "accuracy_std_dev": 0.014060086761267687,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.07144717613636363,
      "avg_duration_per_task_ms": 198887.53787878787,
      "avg_output_tokens_per_task": 16429.516414141413,
      "last_run_timestamp": "2026-06-20T09:16:50.854Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "BaseTen",
      "accuracy": 0.818035862576065,
      "accuracy_std_dev": 0.01278056208505712,
      "run_count": 10,
      "total_questions": 1972,
      "avg_cost_per_task_usd": 0.0909173985801217,
      "avg_duration_per_task_ms": 165913.36257606492,
      "avg_output_tokens_per_task": 22308.299188640973,
      "last_run_timestamp": "2026-09-19T10:14:07.966Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.817971375,
      "accuracy_std_dev": 0.021218711111525495,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0422043106060606,
      "avg_duration_per_task_ms": 258582.00315656565,
      "avg_output_tokens_per_task": 14114.647727272728,
      "last_run_timestamp": "2026-09-26T09:36:02.338Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8177008571428571,
      "accuracy_std_dev": 0.044711167868022074,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.16188713708513708,
      "avg_duration_per_task_ms": 160453.3152958153,
      "avg_output_tokens_per_task": 16045.246753246753,
      "last_run_timestamp": "2026-09-26T08:55:42.287Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8173940034094783,
      "accuracy_std_dev": 0.05734385363328043,
      "run_count": 31,
      "total_questions": 5866,
      "avg_cost_per_task_usd": 0.005422403000340948,
      "avg_duration_per_task_ms": 373758.65342652577,
      "avg_output_tokens_per_task": 12923.683088987385,
      "last_run_timestamp": "2026-08-29T20:12:46.798Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "auto-routing",
      "accuracy": 0.8171395238095238,
      "accuracy_std_dev": 0.02033844061452071,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.004526424482924483,
      "avg_duration_per_task_ms": 331078.8936988937,
      "avg_output_tokens_per_task": 11555.212602212603,
      "last_run_timestamp": "2026-08-29T19:12:36.104Z"
    },
    {
      "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": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8165784761904761,
      "accuracy_std_dev": 0.01245241634466425,
      "run_count": 21,
      "total_questions": 4158,
      "avg_cost_per_task_usd": 0.3341618246753246,
      "avg_duration_per_task_ms": 300449.1738816739,
      "avg_output_tokens_per_task": 21181.114718614717,
      "last_run_timestamp": "2026-09-26T08:51:53.334Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.81596961928934,
      "accuracy_std_dev": 0.04643267204753484,
      "run_count": 21,
      "total_questions": 4137,
      "avg_cost_per_task_usd": 0.07539528402223834,
      "avg_duration_per_task_ms": 501375.46241237613,
      "avg_output_tokens_per_task": 29453.204737732656,
      "last_run_timestamp": "2026-08-29T19:11:37.135Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Chutes",
      "accuracy": 0.815885572815534,
      "accuracy_std_dev": 0.04096815993777976,
      "run_count": 12,
      "total_questions": 2369,
      "avg_cost_per_task_usd": 0.057723436893203896,
      "avg_duration_per_task_ms": 941535.38159561,
      "avg_output_tokens_per_task": 29544.111439425917,
      "last_run_timestamp": "2026-09-26T10:24:45.234Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Phala",
      "accuracy": 0.8157559809402797,
      "accuracy_std_dev": 0.02360825327613207,
      "run_count": 8,
      "total_questions": 1574,
      "avg_cost_per_task_usd": 0.021597616899618807,
      "avg_duration_per_task_ms": 295953.15184243966,
      "avg_output_tokens_per_task": 17957.35959339263,
      "last_run_timestamp": "2026-09-26T08:54:33.107Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Together",
      "accuracy": 0.8154462499999999,
      "accuracy_std_dev": 0.00789342203594232,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.024953141414141414,
      "avg_duration_per_task_ms": 185879.93371212122,
      "avg_output_tokens_per_task": 16543.260101010103,
      "last_run_timestamp": "2026-09-26T08:53:11.662Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Google",
      "accuracy": 0.8153759444444445,
      "accuracy_std_dev": 0.010779567108769665,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.36484757070707075,
      "avg_duration_per_task_ms": 359825.24242424243,
      "avg_output_tokens_per_task": 22980.845679012345,
      "last_run_timestamp": "2026-09-26T08:54:41.498Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Cloudflare",
      "accuracy": 0.8151450853503185,
      "accuracy_std_dev": 0.14825549375062988,
      "run_count": 12,
      "total_questions": 2355,
      "avg_cost_per_task_usd": 0.06763006751592358,
      "avg_duration_per_task_ms": 395557.31549893843,
      "avg_output_tokens_per_task": 18616.247983014862,
      "last_run_timestamp": "2026-09-26T08:55:14.254Z"
    },
    {
      "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": 0.8149456927157002,
      "accuracy_std_dev": 0.18265885331301351,
      "run_count": 16,
      "total_questions": 2828,
      "avg_cost_per_task_usd": 0.17484592397454032,
      "avg_duration_per_task_ms": 315586.27050919377,
      "avg_output_tokens_per_task": 31965.719589816126,
      "last_run_timestamp": "2026-09-26T08:56:04.781Z"
    },
    {
      "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "StreamLake",
      "accuracy": 0.814332751988431,
      "accuracy_std_dev": 0.01885441572018766,
      "run_count": 21,
      "total_questions": 4149,
      "avg_cost_per_task_usd": 0.06707576355748375,
      "avg_duration_per_task_ms": 462943.6165341046,
      "avg_output_tokens_per_task": 33393.012774162446,
      "last_run_timestamp": "2026-08-29T19:17:49.455Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.8142204705882352,
      "accuracy_std_dev": 0.03855851822744482,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.019141188354129528,
      "avg_duration_per_task_ms": 245806.30392156861,
      "avg_output_tokens_per_task": 17822.011883541294,
      "last_run_timestamp": "2026-08-29T19:18:38.408Z"
    },
    {
      "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": "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.8123588325266213,
      "accuracy_std_dev": 0.08477891888464963,
      "run_count": 12,
      "total_questions": 2066,
      "avg_cost_per_task_usd": 0.08345451210067764,
      "avg_duration_per_task_ms": 526015.0217812198,
      "avg_output_tokens_per_task": 26015.62391093901,
      "last_run_timestamp": "2026-09-26T09:56:28.417Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Z.AI",
      "accuracy": 0.8119979337349396,
      "accuracy_std_dev": 0.0940565298454681,
      "run_count": 14,
      "total_questions": 2656,
      "avg_cost_per_task_usd": 0.06128782756024097,
      "avg_duration_per_task_ms": 908022.2884036144,
      "avg_output_tokens_per_task": 28257.003765060243,
      "last_run_timestamp": "2026-08-29T19:12:33.941Z"
    },
    {
      "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": 0.8117484228596261,
      "accuracy_std_dev": 0.09706275314151007,
      "run_count": 12,
      "total_questions": 2301,
      "avg_cost_per_task_usd": 0.019746013472403304,
      "avg_duration_per_task_ms": 100475.48109517602,
      "avg_output_tokens_per_task": 16488.117340286833,
      "last_run_timestamp": "2026-07-18T08:31:24.271Z"
    },
    {
      "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/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Darkbloom",
      "accuracy": 0.8107742,
      "accuracy_std_dev": 0.00793923498581567,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.032759715151515156,
      "avg_duration_per_task_ms": 674236.5636363636,
      "avg_output_tokens_per_task": 15717.559595959596,
      "last_run_timestamp": "2026-09-26T09:29:37.459Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "Alibaba",
      "accuracy": 0.8099312975495916,
      "accuracy_std_dev": 0.096225883038589,
      "run_count": 13,
      "total_questions": 2571,
      "avg_cost_per_task_usd": 0.08600893154414625,
      "avg_duration_per_task_ms": 491078.3780630105,
      "avg_output_tokens_per_task": 31731.871645274212,
      "last_run_timestamp": "2026-09-26T09:22:12.564Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.8092030833333335,
      "accuracy_std_dev": 0.01675517045301904,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.33934875,
      "avg_duration_per_task_ms": 354134.20664983167,
      "avg_output_tokens_per_task": 22457.28787878788,
      "last_run_timestamp": "2026-09-26T08:56:43.900Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Nebius",
      "accuracy": 0.808081,
      "accuracy_std_dev": 0.12407749055328285,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.10305112289562289,
      "avg_duration_per_task_ms": 1499684.5976430976,
      "avg_output_tokens_per_task": 25661.05218855219,
      "last_run_timestamp": "2026-09-26T09:01:35.348Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Parasail",
      "accuracy": 0.8078395250344509,
      "accuracy_std_dev": 0.08454229752765441,
      "run_count": 13,
      "total_questions": 2177,
      "avg_cost_per_task_usd": 0.06321491272393201,
      "avg_duration_per_task_ms": 708932.5346807534,
      "avg_output_tokens_per_task": 20144.69820854387,
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Cloudflare",
      "accuracy": 0.8074874931077692,
      "accuracy_std_dev": 0.0754265571969271,
      "run_count": 17,
      "total_questions": 3192,
      "avg_cost_per_task_usd": 0.1512801278195489,
      "avg_duration_per_task_ms": 334510.98558897246,
      "avg_output_tokens_per_task": 31449.75845864662,
      "last_run_timestamp": "2026-09-26T08:52:36.362Z"
    },
    {
      "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": 0.8069288421052634,
      "accuracy_std_dev": 0.01849876504630044,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.020217205209994687,
      "avg_duration_per_task_ms": 157255.7658160553,
      "avg_output_tokens_per_task": 16045.147793726741,
      "last_run_timestamp": "2026-08-29T19:24:09.844Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.8067576807017544,
      "accuracy_std_dev": 0.04457842166074652,
      "run_count": 13,
      "total_questions": 2565,
      "avg_cost_per_task_usd": 0.07463826237816762,
      "avg_duration_per_task_ms": 680798.849122807,
      "avg_output_tokens_per_task": 28652.57037037037,
      "last_run_timestamp": "2026-09-26T10:35:48.204Z"
    },
    {
      "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": 0.8064057959183675,
      "accuracy_std_dev": 0.010831854991438695,
      "run_count": 6,
      "total_questions": 1176,
      "avg_cost_per_task_usd": 0.026680215136054425,
      "avg_duration_per_task_ms": 253897.95323129251,
      "avg_output_tokens_per_task": 11557.03656462585,
      "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": "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": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Novita",
      "accuracy": 0.8059367902364605,
      "accuracy_std_dev": 0.1514017392727079,
      "run_count": 14,
      "total_questions": 2622,
      "avg_cost_per_task_usd": 0.0040933535469107555,
      "avg_duration_per_task_ms": 403372.7581998474,
      "avg_output_tokens_per_task": 10095.213577421815,
      "last_run_timestamp": "2026-08-29T18:15:38.703Z"
    },
    {
      "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": "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": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "auto-routing",
      "accuracy": 0.8051947142857142,
      "accuracy_std_dev": 0.014445950710090603,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.022680748196248195,
      "avg_duration_per_task_ms": 254100.05483405484,
      "avg_output_tokens_per_task": 17520.435786435788,
      "last_run_timestamp": "2026-09-26T08:51:26.240Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Azure",
      "accuracy": 0.8049008888888889,
      "accuracy_std_dev": 0.023075324904150284,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.3397021683501683,
      "avg_duration_per_task_ms": 355471.79349046014,
      "avg_output_tokens_per_task": 22495.737373737375,
      "last_run_timestamp": "2026-09-26T09:00:45.748Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "AtlasCloud",
      "accuracy": 0.8048629695290858,
      "accuracy_std_dev": 0.015273014540588393,
      "run_count": 11,
      "total_questions": 2166,
      "avg_cost_per_task_usd": 0.011905135734072023,
      "avg_duration_per_task_ms": 306077.56463527237,
      "avg_output_tokens_per_task": 12391.826869806095,
      "last_run_timestamp": "2026-08-29T19:12:57.906Z"
    },
    {
      "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/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.8043396111111111,
      "accuracy_std_dev": 0.016346822909594834,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.02834438776655444,
      "avg_duration_per_task_ms": 298008.4017957351,
      "avg_output_tokens_per_task": 16109.744388327721,
      "last_run_timestamp": "2026-08-22T17:18:01.115Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Friendli",
      "accuracy": 0.8028880787003609,
      "accuracy_std_dev": 0.016209074791917563,
      "run_count": 21,
      "total_questions": 4155,
      "avg_cost_per_task_usd": 0.017011109987966302,
      "avg_duration_per_task_ms": 166436.28519855597,
      "avg_output_tokens_per_task": 11224.947051744886,
      "last_run_timestamp": "2026-08-29T18:07:47.743Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.8025491428571428,
      "accuracy_std_dev": 0.025111070964409142,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.33407534559884555,
      "avg_duration_per_task_ms": 352449.7582972583,
      "avg_output_tokens_per_task": 22169.862914862915,
      "last_run_timestamp": "2026-09-26T08:59:18.032Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Io Net",
      "accuracy": 0.8016385562742562,
      "accuracy_std_dev": 0.007745278218856984,
      "run_count": 4,
      "total_questions": 773,
      "avg_cost_per_task_usd": 0.0594386532988357,
      "avg_duration_per_task_ms": 220979.64165588617,
      "avg_output_tokens_per_task": 17329.77360931436,
      "last_run_timestamp": "2026-09-26T08:53:08.262Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "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": 0.8015572499999999,
      "accuracy_std_dev": 0.051298469831183385,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.15178521254208757,
      "avg_duration_per_task_ms": 725758.385942761,
      "avg_output_tokens_per_task": 49481.54461279461,
      "last_run_timestamp": "2026-09-26T09:47:43.340Z"
    },
    {
      "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": 0.8015315724299064,
      "accuracy_std_dev": 0.03958967650685164,
      "run_count": 13,
      "total_questions": 2568,
      "avg_cost_per_task_usd": 0.03791428894080996,
      "avg_duration_per_task_ms": 408085.0218068536,
      "avg_output_tokens_per_task": 37928.07982866043,
      "last_run_timestamp": "2026-09-26T16:31:04.519Z"
    },
    {
      "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Venice",
      "accuracy": 0.8012819999999999,
      "accuracy_std_dev": 0.08131801393957507,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.0891624254079254,
      "avg_duration_per_task_ms": 504605.8787878788,
      "avg_output_tokens_per_task": 28207.464257964257,
      "last_run_timestamp": "2026-09-26T10:00:58.971Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Parasail",
      "accuracy": 0.8004351497695853,
      "accuracy_std_dev": 0.16221306402555435,
      "run_count": 16,
      "total_questions": 2604,
      "avg_cost_per_task_usd": 0.07896243087557604,
      "avg_duration_per_task_ms": 344311.57757296466,
      "avg_output_tokens_per_task": 18416.05376344086,
      "last_run_timestamp": "2026-08-22T17:29:49.714Z"
    },
    {
      "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": 0.8001267261055633,
      "accuracy_std_dev": 0.014586786566292738,
      "run_count": 11,
      "total_questions": 2103,
      "avg_cost_per_task_usd": 0.010285525915359011,
      "avg_duration_per_task_ms": 337066.20827389444,
      "avg_output_tokens_per_task": 10713.090347123158,
      "last_run_timestamp": "2026-08-15T14:56:55.269Z"
    },
    {
      "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": 0.799421448833034,
      "accuracy_std_dev": 0.026141795886530258,
      "run_count": 17,
      "total_questions": 3342,
      "avg_cost_per_task_usd": 0.010835476361460206,
      "avg_duration_per_task_ms": 320191.3922800718,
      "avg_output_tokens_per_task": 10995.220227408738,
      "last_run_timestamp": "2026-08-29T18:21:38.666Z"
    },
    {
      "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": 0.7994009233361414,
      "accuracy_std_dev": 0.05494984321908906,
      "run_count": 18,
      "total_questions": 3561,
      "avg_cost_per_task_usd": 0.004052114574557709,
      "avg_duration_per_task_ms": 221480.79837124402,
      "avg_output_tokens_per_task": 9875.90255546195,
      "last_run_timestamp": "2026-08-29T19:20:50.265Z"
    },
    {
      "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": 0.7988753739571999,
      "accuracy_std_dev": 0.14188825709410938,
      "run_count": 14,
      "total_questions": 2757,
      "avg_cost_per_task_usd": 0.004531228509249184,
      "avg_duration_per_task_ms": 359570.6064562931,
      "avg_output_tokens_per_task": 11210.340587595212,
      "last_run_timestamp": "2026-08-29T19:18:32.746Z"
    },
    {
      "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": "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": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7979313627639155,
      "accuracy_std_dev": 0.011188259129097978,
      "run_count": 16,
      "total_questions": 3126,
      "avg_cost_per_task_usd": 0.040375727767114525,
      "avg_duration_per_task_ms": 913771.9331413948,
      "avg_output_tokens_per_task": 18420.52111324376,
      "last_run_timestamp": "2026-08-29T20:02:23.660Z"
    },
    {
      "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7972682237569061,
      "accuracy_std_dev": 0.038434880339272746,
      "run_count": 11,
      "total_questions": 2172,
      "avg_cost_per_task_usd": 0.035596821823204415,
      "avg_duration_per_task_ms": 774714.9424493554,
      "avg_output_tokens_per_task": 20523.184622467772,
      "last_run_timestamp": "2026-09-26T09:34:13.954Z"
    },
    {
      "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": 0.7964700671480144,
      "accuracy_std_dev": 0.03714125671277268,
      "run_count": 21,
      "total_questions": 4155,
      "avg_cost_per_task_usd": 0.01084919157641396,
      "avg_duration_per_task_ms": 353681.98291215405,
      "avg_output_tokens_per_task": 10705.062815884476,
      "last_run_timestamp": "2026-08-30T00:18:59.365Z"
    },
    {
      "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": "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": 0.79613792022792,
      "accuracy_std_dev": 0.03686958809587348,
      "run_count": 16,
      "total_questions": 3159,
      "avg_cost_per_task_usd": 0.03350250300728078,
      "avg_duration_per_task_ms": 409082.3953782843,
      "avg_output_tokens_per_task": 16575.017094017094,
      "last_run_timestamp": "2026-08-22T17:09:41.940Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.795641611111111,
      "accuracy_std_dev": 0.10336463207668649,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.02163882828282828,
      "avg_duration_per_task_ms": 103515.63888888889,
      "avg_output_tokens_per_task": 16579.40095398429,
      "last_run_timestamp": "2026-08-29T19:13:10.747Z"
    },
    {
      "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": 0.7956189598572703,
      "accuracy_std_dev": 0.02768081371661319,
      "run_count": 17,
      "total_questions": 3363,
      "avg_cost_per_task_usd": 0.009608434136187926,
      "avg_duration_per_task_ms": 421216.80582812964,
      "avg_output_tokens_per_task": 24272.67320844484,
      "last_run_timestamp": "2026-08-29T18:49:24.022Z"
    },
    {
      "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": 0.7953039029322547,
      "accuracy_std_dev": 0.0703931299351167,
      "run_count": 15,
      "total_questions": 2967,
      "avg_cost_per_task_usd": 0.07654278227165487,
      "avg_duration_per_task_ms": 510887.7367711493,
      "avg_output_tokens_per_task": 24439.035389282104,
      "last_run_timestamp": "2026-08-22T17:17:31.993Z"
    },
    {
      "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": "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "WandB",
      "accuracy": 0.7929293571428571,
      "accuracy_std_dev": 0.017530103445245845,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.011040959595959597,
      "avg_duration_per_task_ms": 175837.8427128427,
      "avg_output_tokens_per_task": 6562.907287157287,
      "last_run_timestamp": "2026-07-18T16:16:42.667Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7924127262693157,
      "accuracy_std_dev": 0.05800564622951288,
      "run_count": 21,
      "total_questions": 4077,
      "avg_cost_per_task_usd": 0.07450861589403973,
      "avg_duration_per_task_ms": 572032.6973264655,
      "avg_output_tokens_per_task": 33361.80819229826,
      "last_run_timestamp": "2026-08-29T18:23:09.250Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Parasail",
      "accuracy": 0.7920985116580311,
      "accuracy_std_dev": 0.11713420011607767,
      "run_count": 12,
      "total_questions": 2316,
      "avg_cost_per_task_usd": 0.10608431692573402,
      "avg_duration_per_task_ms": 736930.4978411053,
      "avg_output_tokens_per_task": 31587.977979274612,
      "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-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.7920687490720117,
      "accuracy_std_dev": 0.023696432575952595,
      "run_count": 15,
      "total_questions": 2694,
      "avg_cost_per_task_usd": 0.035895181143281364,
      "avg_duration_per_task_ms": 624123.5913140312,
      "avg_output_tokens_per_task": 19506.129547141798,
      "last_run_timestamp": "2026-08-29T18:31:11.241Z"
    },
    {
      "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": 0.7919846781170483,
      "accuracy_std_dev": 0.05508513093858188,
      "run_count": 13,
      "total_questions": 2358,
      "avg_cost_per_task_usd": 0.0414582417302799,
      "avg_duration_per_task_ms": 452143.6522476675,
      "avg_output_tokens_per_task": 21406.899491094147,
      "last_run_timestamp": "2026-09-26T09:56:00.145Z"
    },
    {
      "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-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Venice",
      "accuracy": 0.7909026798917945,
      "accuracy_std_dev": 0.03754431133546038,
      "run_count": 17,
      "total_questions": 3327,
      "avg_cost_per_task_usd": 0.06559778960024044,
      "avg_duration_per_task_ms": 355359.5933273219,
      "avg_output_tokens_per_task": 19344.585813044785,
      "last_run_timestamp": "2026-08-29T18:54:50.382Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-35b-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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Parasail",
      "accuracy": 0.7900883420677869,
      "accuracy_std_dev": 0.09870007461496674,
      "run_count": 18,
      "total_questions": 3511,
      "avg_cost_per_task_usd": 0.015616578182853886,
      "avg_duration_per_task_ms": 212237.6132156081,
      "avg_output_tokens_per_task": 15795.737396753062,
      "last_run_timestamp": "2026-08-29T20:00:43.878Z"
    },
    {
      "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": 0.7888498053927315,
      "accuracy_std_dev": 0.014122812704500139,
      "run_count": 13,
      "total_questions": 2559,
      "avg_cost_per_task_usd": 0.13022648261039466,
      "avg_duration_per_task_ms": 677813.3364595545,
      "avg_output_tokens_per_task": 32555.26768268855,
      "last_run_timestamp": "2026-08-29T20:08:40.362Z"
    },
    {
      "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": 0.7888041145038167,
      "accuracy_std_dev": 0.010988374090480538,
      "run_count": 16,
      "total_questions": 3144,
      "avg_cost_per_task_usd": 0.039076502544529254,
      "avg_duration_per_task_ms": 708080.6669847328,
      "avg_output_tokens_per_task": 17823.317748091602,
      "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": 0.7885914761215629,
      "accuracy_std_dev": 0.02234077768730239,
      "run_count": 7,
      "total_questions": 1382,
      "avg_cost_per_task_usd": 0.06106606946454414,
      "avg_duration_per_task_ms": 179070.87192474675,
      "avg_output_tokens_per_task": 19226.02532561505,
      "last_run_timestamp": "2026-09-26T09:36:55.311Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7884858931297708,
      "accuracy_std_dev": 0.009706627796897349,
      "run_count": 8,
      "total_questions": 1572,
      "avg_cost_per_task_usd": 0.01825112468193384,
      "avg_duration_per_task_ms": 147479.76017811705,
      "avg_output_tokens_per_task": 15055.673664122138,
      "last_run_timestamp": "2026-09-26T09:15:51.063Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.7884401666666667,
      "accuracy_std_dev": 0.023761946338748348,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.04315188552188553,
      "avg_duration_per_task_ms": 84212.74747474748,
      "avg_output_tokens_per_task": 26885.046296296296,
      "last_run_timestamp": "2026-09-01T17:55:54.208Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "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": "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-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7850717877461707,
      "accuracy_std_dev": 0.024805753669301024,
      "run_count": 7,
      "total_questions": 1371,
      "avg_cost_per_task_usd": 0.024575857038657913,
      "avg_duration_per_task_ms": 416592.56673960615,
      "avg_output_tokens_per_task": 11405.052516411379,
      "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": "GMICloud",
      "accuracy": 0.7847221249999999,
      "accuracy_std_dev": 0.15720335838129978,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.017239123106060607,
      "avg_duration_per_task_ms": 568625.6464646464,
      "avg_output_tokens_per_task": 27456.18876262626,
      "last_run_timestamp": "2026-09-26T08:08:15.090Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.7835429351415094,
      "accuracy_std_dev": 0.03263391726736184,
      "run_count": 13,
      "total_questions": 2544,
      "avg_cost_per_task_usd": 0.004777952437106918,
      "avg_duration_per_task_ms": 504503.3958333333,
      "avg_output_tokens_per_task": 11386.457547169812,
      "last_run_timestamp": "2026-08-01T18:07:34.003Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.7823070191082802,
      "accuracy_std_dev": 0.03047559886602813,
      "run_count": 12,
      "total_questions": 2355,
      "avg_cost_per_task_usd": 0.0051865269639065825,
      "avg_duration_per_task_ms": 669935.3341825902,
      "avg_output_tokens_per_task": 12408.887048832272,
      "last_run_timestamp": "2026-08-01T22:57:01.346Z"
    },
    {
      "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": "auto-routing",
      "accuracy": 0.7816287308868499,
      "accuracy_std_dev": 0.04679161436791379,
      "run_count": 15,
      "total_questions": 2943,
      "avg_cost_per_task_usd": 0.03975552463472646,
      "avg_duration_per_task_ms": 656449.841658172,
      "avg_output_tokens_per_task": 20722.471967380225,
      "last_run_timestamp": "2026-08-29T18:19:18.341Z"
    },
    {
      "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-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Fireworks",
      "accuracy": 0.7803030000000001,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.006998724747474747,
      "avg_duration_per_task_ms": 57307.29797979798,
      "avg_output_tokens_per_task": 4087.7727272727275,
      "last_run_timestamp": "2026-05-09T16:48:20.198Z"
    },
    {
      "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": 0.780303,
      "accuracy_std_dev": 0.08147600069959253,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.07258149494949495,
      "avg_duration_per_task_ms": 318265.46338383836,
      "avg_output_tokens_per_task": 18090.291666666668,
      "last_run_timestamp": "2026-05-23T11:07:21.864Z"
    },
    {
      "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": 0.7802894910474485,
      "accuracy_std_dev": 0.1302593050858772,
      "run_count": 13,
      "total_questions": 2234,
      "avg_cost_per_task_usd": 0.020609962399283796,
      "avg_duration_per_task_ms": 473420.5805729633,
      "avg_output_tokens_per_task": 12826.314682184422,
      "last_run_timestamp": "2026-08-01T23:27:37.975Z"
    },
    {
      "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": 0.7800773017429193,
      "accuracy_std_dev": 0.021465098000437065,
      "run_count": 14,
      "total_questions": 2754,
      "avg_cost_per_task_usd": 0.04367889179375453,
      "avg_duration_per_task_ms": 454980.2806826434,
      "avg_output_tokens_per_task": 19988.257080610023,
      "last_run_timestamp": "2026-08-29T18:26:43.144Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-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": "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": 0.7793682236135957,
      "accuracy_std_dev": 0.06789026487076204,
      "run_count": 3,
      "total_questions": 559,
      "avg_cost_per_task_usd": 0.029017345259391772,
      "avg_duration_per_task_ms": 2038522.3381037568,
      "avg_output_tokens_per_task": 57994.92486583184,
      "last_run_timestamp": "2026-09-26T09:27:27.697Z"
    },
    {
      "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": "Together",
      "accuracy": 0.7773231747713049,
      "accuracy_std_dev": 0.02619804956117758,
      "run_count": 11,
      "total_questions": 2077,
      "avg_cost_per_task_usd": 0.009061719306692344,
      "avg_duration_per_task_ms": 444770.355801637,
      "avg_output_tokens_per_task": 36121.13769860376,
      "last_run_timestamp": "2026-09-26T17:33:23.516Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Friendli",
      "accuracy": 0.7772597692307692,
      "accuracy_std_dev": 0.2020442328909321,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.056136898989899,
      "avg_duration_per_task_ms": 168843.6891996892,
      "avg_output_tokens_per_task": 17466.01320901321,
      "last_run_timestamp": "2026-07-01T03:02:14.071Z"
    },
    {
      "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": 0.774354836492891,
      "accuracy_std_dev": 0.025023485914745035,
      "run_count": 13,
      "total_questions": 2532,
      "avg_cost_per_task_usd": 0.005451107424960505,
      "avg_duration_per_task_ms": 662585.4802527647,
      "avg_output_tokens_per_task": 34288.59873617694,
      "last_run_timestamp": "2026-09-26T19:06:56.023Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Io Net",
      "accuracy": 0.774186536121673,
      "accuracy_std_dev": 0.045556886694000515,
      "run_count": 4,
      "total_questions": 789,
      "avg_cost_per_task_usd": 0.0076262953105196455,
      "avg_duration_per_task_ms": 513514.7541191382,
      "avg_output_tokens_per_task": 15020.147021546261,
      "last_run_timestamp": "2026-09-26T08:55:07.944Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Venice",
      "accuracy": 0.7733335691428571,
      "accuracy_std_dev": 0.10983726595387482,
      "run_count": 6,
      "total_questions": 875,
      "avg_cost_per_task_usd": 0.013626054857142857,
      "avg_duration_per_task_ms": 549081.1371428572,
      "avg_output_tokens_per_task": 22934.75657142857,
      "last_run_timestamp": "2026-09-26T10:43:59.946Z"
    },
    {
      "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": "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": "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": "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": 0.770754223767383,
      "accuracy_std_dev": 0.04391892511863639,
      "run_count": 12,
      "total_questions": 2373,
      "avg_cost_per_task_usd": 0.005076832279814581,
      "avg_duration_per_task_ms": 274364.3603034134,
      "avg_output_tokens_per_task": 33323.141171512856,
      "last_run_timestamp": "2026-09-26T16:14:01.905Z"
    },
    {
      "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": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Chutes",
      "accuracy": 0.7673517422506738,
      "accuracy_std_dev": 0.1696644004135443,
      "run_count": 16,
      "total_questions": 2968,
      "avg_cost_per_task_usd": 0.04616438106469003,
      "avg_duration_per_task_ms": 423953.0438005391,
      "avg_output_tokens_per_task": 18353.38847708895,
      "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-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "Azure",
      "accuracy": 0.7672560000000002,
      "accuracy_std_dev": 0.0089379450938121,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.013739436868686868,
      "avg_duration_per_task_ms": 117371.45833333333,
      "avg_output_tokens_per_task": 10352.156565656565,
      "last_run_timestamp": "2026-09-01T17:58:15.198Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "Google",
      "accuracy": 0.7672559999999999,
      "accuracy_std_dev": 0.008469894233892938,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.27037940151515155,
      "avg_duration_per_task_ms": 263755.6675084175,
      "avg_output_tokens_per_task": 17837.74579124579,
      "last_run_timestamp": "2026-09-01T13:23:02.059Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Reka",
      "accuracy": 0.7671957142857142,
      "accuracy_std_dev": 0.009679973770836427,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.017645118326118325,
      "avg_duration_per_task_ms": 105010.88528138529,
      "avg_output_tokens_per_task": 6033.81746031746,
      "last_run_timestamp": "2026-09-26T09:17:54.791Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4-sonnet-20250522",
      "model_slug": "anthropic/claude-sonnet-4",
      "model_name": "Anthropic: Claude Sonnet 4",
      "provider_name": "auto-routing",
      "accuracy": 0.7668349999999999,
      "accuracy_std_dev": 0.01935999999999999,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.2670539015151515,
      "avg_duration_per_task_ms": 177095.6616161616,
      "avg_output_tokens_per_task": 17616.689393939392,
      "last_run_timestamp": "2026-09-01T13:19:59.945Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "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": 0.76443,
      "accuracy_std_dev": 0.03329192690428117,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.013529336940836939,
      "avg_duration_per_task_ms": 416657.21428571426,
      "avg_output_tokens_per_task": 23392.275974025975,
      "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-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DeepInfra",
      "accuracy": 0.7643024133099825,
      "accuracy_std_dev": 0.031491855215101075,
      "run_count": 6,
      "total_questions": 1142,
      "avg_cost_per_task_usd": 0.04750308056042031,
      "avg_duration_per_task_ms": 265287.42907180387,
      "avg_output_tokens_per_task": 17893.702276707532,
      "last_run_timestamp": "2026-09-26T09:16:55.239Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "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": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Phala",
      "accuracy": 0.7637618470588236,
      "accuracy_std_dev": 0.12321473328330662,
      "run_count": 12,
      "total_questions": 2295,
      "avg_cost_per_task_usd": 0.0732439917211329,
      "avg_duration_per_task_ms": 793890.7620915032,
      "avg_output_tokens_per_task": 27412.28496732026,
      "last_run_timestamp": "2026-09-26T10:34:37.144Z"
    },
    {
      "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": "Novita",
      "accuracy": 0.7636952078580483,
      "accuracy_std_dev": 0.023951539715497544,
      "run_count": 12,
      "total_questions": 2367,
      "avg_cost_per_task_usd": 0.010468032530629489,
      "avg_duration_per_task_ms": 136682.62822137726,
      "avg_output_tokens_per_task": 4070.7904520490074,
      "last_run_timestamp": "2026-08-01T17:44:11.144Z"
    },
    {
      "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": "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": 0.7636362000000001,
      "accuracy_std_dev": 0.02440997912657852,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.00031175454545454543,
      "avg_duration_per_task_ms": 34241.53636363636,
      "avg_output_tokens_per_task": 2848.0313131313133,
      "last_run_timestamp": "2026-06-20T16:16:12.961Z"
    },
    {
      "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": 0.7632893237410072,
      "accuracy_std_dev": 0.11557017973361858,
      "run_count": 17,
      "total_questions": 3336,
      "avg_cost_per_task_usd": 0.0028636052158273377,
      "avg_duration_per_task_ms": 431031.65257793764,
      "avg_output_tokens_per_task": 14939.968824940048,
      "last_run_timestamp": "2026-08-29T19:07:25.263Z"
    },
    {
      "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": 0.7625483969176351,
      "accuracy_std_dev": 0.04157398706916983,
      "run_count": 20,
      "total_questions": 3958,
      "avg_cost_per_task_usd": 0.012629246336533605,
      "avg_duration_per_task_ms": 83704.65184436584,
      "avg_output_tokens_per_task": 7288.428499242042,
      "last_run_timestamp": "2026-08-22T17:12:08.919Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "StepFun",
      "accuracy": 0.760943,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.08179280303030302,
      "avg_duration_per_task_ms": 426340.89898989897,
      "avg_output_tokens_per_task": 70994.31818181818,
      "last_run_timestamp": "2026-07-18T08:26:55.806Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "AkashML",
      "accuracy": 0.760674964876033,
      "accuracy_std_dev": 0.01645419404771925,
      "run_count": 5,
      "total_questions": 968,
      "avg_cost_per_task_usd": 0.010708528925619836,
      "avg_duration_per_task_ms": 425619.78925619833,
      "avg_output_tokens_per_task": 25673.311983471074,
      "last_run_timestamp": "2026-08-29T18:56:40.728Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "GMICloud",
      "accuracy": 0.7606652169197398,
      "accuracy_std_dev": 0.04591514075032818,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.004993729573391178,
      "avg_duration_per_task_ms": 317882.5820679682,
      "avg_output_tokens_per_task": 12242.755603759942,
      "last_run_timestamp": "2026-08-29T18:46:19.350Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "OpenInference",
      "accuracy": 0.7605443265306122,
      "accuracy_std_dev": 0.0343317834601635,
      "run_count": 3,
      "total_questions": 490,
      "avg_cost_per_task_usd": 0.0037374408163265305,
      "avg_duration_per_task_ms": 668153.0571428571,
      "avg_output_tokens_per_task": 14206.583673469388,
      "last_run_timestamp": "2026-08-08T11:42:26.676Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "auto-routing",
      "accuracy": 0.757881909090909,
      "accuracy_std_dev": 0.01539297995518821,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.010656316804407715,
      "avg_duration_per_task_ms": 94392.0583103765,
      "avg_output_tokens_per_task": 3877.1974288337924,
      "last_run_timestamp": "2026-08-01T17:45:27.258Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "mistralai/mistral-small-2603",
      "model_slug": "mistralai/mistral-small-2603",
      "model_name": "Mistral: Mistral Small 4",
      "provider_name": "auto-routing",
      "accuracy": 0.7575759999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.014556641414141414,
      "avg_duration_per_task_ms": 173179.91919191918,
      "avg_output_tokens_per_task": 23980.964646464647,
      "last_run_timestamp": "2026-09-01T13:30:26.223Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7575162329411763,
      "accuracy_std_dev": 0.07851061401993352,
      "run_count": 13,
      "total_questions": 2550,
      "avg_cost_per_task_usd": 0.004892281568627452,
      "avg_duration_per_task_ms": 717348.5560784314,
      "avg_output_tokens_per_task": 32493.26745098039,
      "last_run_timestamp": "2026-09-26T18:04:35.621Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Parasail",
      "accuracy": 0.7572815631067961,
      "accuracy_std_dev": 0.10286820253600068,
      "run_count": 9,
      "total_questions": 721,
      "avg_cost_per_task_usd": 0.002934966712898752,
      "avg_duration_per_task_ms": 194322.5145631068,
      "avg_output_tokens_per_task": 13374.00693481276,
      "last_run_timestamp": "2026-09-26T16:00:24.612Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7565450000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 191,
      "avg_cost_per_task_usd": 0.020916916230366494,
      "avg_duration_per_task_ms": 124793.11518324607,
      "avg_output_tokens_per_task": 18115.47120418848,
      "last_run_timestamp": "2026-07-18T08:26:49.504Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7559323627118644,
      "accuracy_std_dev": 0.007798453236949536,
      "run_count": 3,
      "total_questions": 590,
      "avg_cost_per_task_usd": 0.002668813559322034,
      "avg_duration_per_task_ms": 91653.82203389831,
      "avg_output_tokens_per_task": 31189.77966101695,
      "last_run_timestamp": "2026-09-01T13:23:13.492Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7557626153846154,
      "accuracy_std_dev": 0.1999902666568237,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.005653325951825951,
      "avg_duration_per_task_ms": 41439.5710955711,
      "avg_output_tokens_per_task": 5907.5940170940175,
      "last_run_timestamp": "2026-07-01T02:27:21.176Z"
    },
    {
      "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.7538529877807133,
      "accuracy_std_dev": 0.025619095703080515,
      "run_count": 17,
      "total_questions": 3028,
      "avg_cost_per_task_usd": 0.004401706406869221,
      "avg_duration_per_task_ms": 679926.5132100396,
      "avg_output_tokens_per_task": 11358.332892998678,
      "last_run_timestamp": "2026-08-29T19:19:08.877Z"
    },
    {
      "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": 0.7537320340136054,
      "accuracy_std_dev": 0.07541786806041245,
      "run_count": 18,
      "total_questions": 3528,
      "avg_cost_per_task_usd": 0.006723561224489798,
      "avg_duration_per_task_ms": 628214.5221088436,
      "avg_output_tokens_per_task": 7597.763888888889,
      "last_run_timestamp": "2026-08-30T01:20:44.266Z"
    },
    {
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Phala",
      "accuracy": 0.7521143262182843,
      "accuracy_std_dev": 0.1621979257907336,
      "run_count": 14,
      "total_questions": 2483,
      "avg_cost_per_task_usd": 0.06938106121627065,
      "avg_duration_per_task_ms": 759326.9730165122,
      "avg_output_tokens_per_task": 24775.081755940395,
      "last_run_timestamp": "2026-08-29T18:25:52.753Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "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/qwen3.6-27b-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": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.75,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0067602800000000005,
      "avg_duration_per_task_ms": 1824454.58,
      "avg_output_tokens_per_task": 23925.293333333335,
      "last_run_timestamp": "2026-09-22T21:48:57.116Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": 0.7491580000000001,
      "accuracy_std_dev": 0.023808285322551064,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.0035041430976430975,
      "avg_duration_per_task_ms": 3687.0656565656564,
      "avg_output_tokens_per_task": 2784.715488215488,
      "last_run_timestamp": "2026-05-23T08:06:03.679Z"
    },
    {
      "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": 0.7489477499999999,
      "accuracy_std_dev": 0.014071086487457193,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.006045699494949496,
      "avg_duration_per_task_ms": 101503.2803030303,
      "avg_output_tokens_per_task": 2341.207702020202,
      "last_run_timestamp": "2026-07-01T05:04:09.747Z"
    },
    {
      "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": "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.7475682777777778,
      "accuracy_std_dev": 0.015187563478808556,
      "run_count": 18,
      "total_questions": 3564,
      "avg_cost_per_task_usd": 0.003526737373737373,
      "avg_duration_per_task_ms": 79157.85970819303,
      "avg_output_tokens_per_task": 5754.911335578002,
      "last_run_timestamp": "2026-08-29T18:15:48.874Z"
    },
    {
      "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": "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/qwen3-235b-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": "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": 0.746212,
      "accuracy_std_dev": 0.022409208720970065,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.005678362373737374,
      "avg_duration_per_task_ms": 183239.09722222222,
      "avg_output_tokens_per_task": 11277.801767676769,
      "last_run_timestamp": "2026-05-23T09:41:28.552Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-397b-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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "SambaNova",
      "accuracy": 0.745544660208109,
      "accuracy_std_dev": 0.22950971531350736,
      "run_count": 17,
      "total_questions": 2787,
      "avg_cost_per_task_usd": 0.044433735916756374,
      "avg_duration_per_task_ms": 63269.68963042698,
      "avg_output_tokens_per_task": 14879.802655184787,
      "last_run_timestamp": "2026-08-15T14:54:40.470Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Phala",
      "accuracy": 0.7453985513905684,
      "accuracy_std_dev": 0.1009705682494428,
      "run_count": 13,
      "total_questions": 2481,
      "avg_cost_per_task_usd": 0.1381599939540508,
      "avg_duration_per_task_ms": 992902.1555824265,
      "avg_output_tokens_per_task": 32901.31761386538,
      "last_run_timestamp": "2026-08-29T18:42:16.346Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "auto-routing",
      "accuracy": 0.7450981176470588,
      "accuracy_std_dev": 0.012808884861607752,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.01028051633986928,
      "avg_duration_per_task_ms": 78818.64052287582,
      "avg_output_tokens_per_task": 10202.047534165182,
      "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": "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": "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": 0.7431457142857144,
      "accuracy_std_dev": 0.01442199540101643,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.0028307178932178937,
      "avg_duration_per_task_ms": 131645.240981241,
      "avg_output_tokens_per_task": 4593.8621933621935,
      "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": "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": "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": 0.7425010399239544,
      "accuracy_std_dev": 0.040756824836413526,
      "run_count": 16,
      "total_questions": 3156,
      "avg_cost_per_task_usd": 0.008645365969581749,
      "avg_duration_per_task_ms": 476821.68187579216,
      "avg_output_tokens_per_task": 23199.611850443598,
      "last_run_timestamp": "2026-08-29T19:32:52.576Z"
    },
    {
      "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.7423450257654967,
      "accuracy_std_dev": 0.057576989386154395,
      "run_count": 15,
      "total_questions": 2678,
      "avg_cost_per_task_usd": 0.08512316430171771,
      "avg_duration_per_task_ms": 515919.2300224048,
      "avg_output_tokens_per_task": 24342.80283793876,
      "last_run_timestamp": "2026-08-22T17:30:52.646Z"
    },
    {
      "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": 0.7411821798483208,
      "accuracy_std_dev": 0.01874609358231574,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.003938289635247381,
      "avg_duration_per_task_ms": 186658.11412062118,
      "avg_output_tokens_per_task": 6185.643553629469,
      "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": "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": 0.7393435345268542,
      "accuracy_std_dev": 0.04818327390886166,
      "run_count": 12,
      "total_questions": 2346,
      "avg_cost_per_task_usd": 0.008884832054560953,
      "avg_duration_per_task_ms": 737670.1547314578,
      "avg_output_tokens_per_task": 22523.651321398123,
      "last_run_timestamp": "2026-08-22T10:02:15.507Z"
    },
    {
      "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": 0.7389277692307694,
      "accuracy_std_dev": 0.024649382442424383,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.009822144133644132,
      "avg_duration_per_task_ms": 14839.177544677545,
      "avg_output_tokens_per_task": 3146.2645687645686,
      "last_run_timestamp": "2026-08-01T17:45:49.025Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Phala",
      "accuracy": 0.7388467499999999,
      "accuracy_std_dev": 0.07747406520692911,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.01745328409090909,
      "avg_duration_per_task_ms": 2275674.0732323234,
      "avg_output_tokens_per_task": 17131.02398989899,
      "last_run_timestamp": "2026-08-22T18:24:57.217Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "Google",
      "accuracy": 0.7384959999999999,
      "accuracy_std_dev": 0.0076941959727230055,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.058646202861952865,
      "avg_duration_per_task_ms": 123198.74579124579,
      "avg_output_tokens_per_task": 23489.968855218856,
      "last_run_timestamp": "2026-09-01T17:53:51.881Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Google",
      "accuracy": 0.7382653844545136,
      "accuracy_std_dev": 0.017460726869285965,
      "run_count": 29,
      "total_questions": 5738,
      "avg_cost_per_task_usd": 0.005406186301847333,
      "avg_duration_per_task_ms": 69408.7223771349,
      "avg_output_tokens_per_task": 5668.429766469153,
      "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": 0.7378226666666667,
      "accuracy_std_dev": 0.013319255277813231,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.01630272525252525,
      "avg_duration_per_task_ms": 70473.10673400674,
      "avg_output_tokens_per_task": 10789.047138047137,
      "last_run_timestamp": "2026-08-29T18:28:10.968Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "GMICloud",
      "accuracy": 0.7355811297709923,
      "accuracy_std_dev": 0.10861849313307564,
      "run_count": 8,
      "total_questions": 1572,
      "avg_cost_per_task_usd": 0.00556993893129771,
      "avg_duration_per_task_ms": 485638.58396946563,
      "avg_output_tokens_per_task": 24035.00954198473,
      "last_run_timestamp": "2026-09-26T09:37:20.718Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": 0.7352940512333966,
      "accuracy_std_dev": 0.018881487682631787,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.0033387254901960785,
      "avg_duration_per_task_ms": 137055.96173308033,
      "avg_output_tokens_per_task": 5667.310562934851,
      "last_run_timestamp": "2026-08-29T18:16:39.843Z"
    },
    {
      "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": "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": 0.7333755,
      "accuracy_std_dev": 0.033868640415951154,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.005042782196969697,
      "avg_duration_per_task_ms": 94336.22032828283,
      "avg_output_tokens_per_task": 5601.336805555556,
      "last_run_timestamp": "2026-08-29T18:01:46.940Z"
    },
    {
      "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7329965999999999,
      "accuracy_std_dev": 0.010941396457491143,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.007063433333333333,
      "avg_duration_per_task_ms": 42635.63636363636,
      "avg_output_tokens_per_task": 5195.6535353535355,
      "last_run_timestamp": "2026-07-01T02:32:36.197Z"
    },
    {
      "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": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7312871538461537,
      "accuracy_std_dev": 0.012238898554301565,
      "run_count": 26,
      "total_questions": 5148,
      "avg_cost_per_task_usd": 0.20442489957264962,
      "avg_duration_per_task_ms": 242818.47999223,
      "avg_output_tokens_per_task": 38189.194832944835,
      "last_run_timestamp": "2026-09-26T08:58:38.716Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7299561400595999,
      "accuracy_std_dev": 0.20101017137136507,
      "run_count": 13,
      "total_questions": 2349,
      "avg_cost_per_task_usd": 0.004135057471264368,
      "avg_duration_per_task_ms": 932978.9770114942,
      "avg_output_tokens_per_task": 27248.059174116646,
      "last_run_timestamp": "2026-09-26T21:21:20.008Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7297043333333333,
      "accuracy_std_dev": 0.017207018593843885,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.1851553911335578,
      "avg_duration_per_task_ms": 289835.51515151514,
      "avg_output_tokens_per_task": 36869.942199775534,
      "last_run_timestamp": "2026-09-26T09:21:00.998Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "auto-routing",
      "accuracy": 0.727459888888889,
      "accuracy_std_dev": 0.010021023926885385,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.1903540600448934,
      "avg_duration_per_task_ms": 241489.60381593715,
      "avg_output_tokens_per_task": 37880.08529741863,
      "last_run_timestamp": "2026-09-26T08:54:32.344Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7272725,
      "accuracy_std_dev": 0.005050499999999958,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.058774820707070706,
      "avg_duration_per_task_ms": 121424.5,
      "avg_output_tokens_per_task": 23508.14898989899,
      "last_run_timestamp": "2026-09-01T17:49:53.883Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "DeepInfra",
      "accuracy": 0.7260439146496813,
      "accuracy_std_dev": 0.007747089236538362,
      "run_count": 12,
      "total_questions": 2355,
      "avg_cost_per_task_usd": 0.0025452929936305734,
      "avg_duration_per_task_ms": 422788.46751592355,
      "avg_output_tokens_per_task": 6748.217409766455,
      "last_run_timestamp": "2026-08-29T18:39:40.767Z"
    },
    {
      "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": 0.7258483846153846,
      "accuracy_std_dev": 0.07587921757183326,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.11719855205905207,
      "avg_duration_per_task_ms": 909372.0038850039,
      "avg_output_tokens_per_task": 29889.70202020202,
      "last_run_timestamp": "2026-08-29T19:12:26.877Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Azure",
      "accuracy": 0.7245957898383372,
      "accuracy_std_dev": 0.009250125041530539,
      "run_count": 9,
      "total_questions": 1732,
      "avg_cost_per_task_usd": 0.1865489249422633,
      "avg_duration_per_task_ms": 350199.4688221709,
      "avg_output_tokens_per_task": 37119.776558891455,
      "last_run_timestamp": "2026-09-26T08:53:09.141Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Venice",
      "accuracy": 0.7230691676829271,
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Phala",
      "accuracy": 0.7219416666666667,
      "accuracy_std_dev": 0.05789083119880804,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.011145885521885522,
      "avg_duration_per_task_ms": 86502.09595959596,
      "avg_output_tokens_per_task": 4478.992424242424,
      "last_run_timestamp": "2026-09-26T09:45:38.864Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "ModelRun",
      "accuracy": 0.7219408075949366,
      "accuracy_std_dev": 0.025598999999999983,
      "run_count": 2,
      "total_questions": 395,
      "avg_cost_per_task_usd": 0.27939689620253166,
      "avg_duration_per_task_ms": 450249.90379746835,
      "avg_output_tokens_per_task": 45941.00253164557,
      "last_run_timestamp": "2026-08-15T08:16:15.061Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Google",
      "accuracy": 0.7204764444444446,
      "accuracy_std_dev": 0.010281167113216102,
      "run_count": 27,
      "total_questions": 5346,
      "avg_cost_per_task_usd": 0.1972407962962963,
      "avg_duration_per_task_ms": 261625.21979049756,
      "avg_output_tokens_per_task": 36783.956041900485,
      "last_run_timestamp": "2026-09-26T09:21:19.221Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Parasail",
      "accuracy": 0.7184962054677487,
      "accuracy_std_dev": 0.2646528685755041,
      "run_count": 12,
      "total_questions": 2341,
      "avg_cost_per_task_usd": 0.06396244126441693,
      "avg_duration_per_task_ms": 525060.6198205895,
      "avg_output_tokens_per_task": 18294.98761213157,
      "last_run_timestamp": "2026-08-29T18:36:08.021Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7184345000000002,
      "accuracy_std_dev": 0.00575544305418794,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.006676434343434343,
      "avg_duration_per_task_ms": 1198975.5896464647,
      "avg_output_tokens_per_task": 54074.54671717172,
      "last_run_timestamp": "2026-09-26T17:35:27.801Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Nebius",
      "accuracy": 0.7183463229974161,
      "accuracy_std_dev": 0.029264511720208542,
      "run_count": 6,
      "total_questions": 1161,
      "avg_cost_per_task_usd": 0.17074779758828595,
      "avg_duration_per_task_ms": 1524939.6339362618,
      "avg_output_tokens_per_task": 41593.74849267872,
      "last_run_timestamp": "2026-08-22T16:49:30.177Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "GMICloud",
      "accuracy": 0.7173431778597786,
      "accuracy_std_dev": 0.1289599653571277,
      "run_count": 21,
      "total_questions": 4065,
      "avg_cost_per_task_usd": 0.07600008265682658,
      "avg_duration_per_task_ms": 666511.178597786,
      "avg_output_tokens_per_task": 41226.01648216482,
      "last_run_timestamp": "2026-08-29T19:10:33.460Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "inclusionai/ling-3.0-flash-20260723",
      "model_slug": "inclusionai/ling-3.0-flash",
      "model_name": "inclusionAI: Ling 3.0 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7170990662824207,
      "accuracy_std_dev": 0.10692249999999998,
      "run_count": 2,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.005610605187319885,
      "avg_duration_per_task_ms": 269295.4322766571,
      "avg_output_tokens_per_task": 27309.671469740635,
      "last_run_timestamp": "2026-09-01T13:46:27.106Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.7157249440677965,
      "accuracy_std_dev": 0.01597995721469033,
      "run_count": 18,
      "total_questions": 3540,
      "avg_cost_per_task_usd": 0.005597196045197738,
      "avg_duration_per_task_ms": 49724.024576271186,
      "avg_output_tokens_per_task": 7110.428813559322,
      "last_run_timestamp": "2026-08-29T18:12:18.859Z"
    },
    {
      "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": 0.7142573873954601,
      "accuracy_std_dev": 0.01848301628382384,
      "run_count": 17,
      "total_questions": 3348,
      "avg_cost_per_task_usd": 0.007225585424133809,
      "avg_duration_per_task_ms": 67629.55943847072,
      "avg_output_tokens_per_task": 6617.06541218638,
      "last_run_timestamp": "2026-08-29T19:10:03.342Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Phala",
      "accuracy": 0.7142555025380709,
      "accuracy_std_dev": 0.033662331581457615,
      "run_count": 4,
      "total_questions": 788,
      "avg_cost_per_task_usd": 0.0112786192893401,
      "avg_duration_per_task_ms": 562577.1395939087,
      "avg_output_tokens_per_task": 56130.946700507615,
      "last_run_timestamp": "2026-09-26T09:18:34.055Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Venice",
      "accuracy": 0.7136277774106337,
      "accuracy_std_dev": 0.03414979758617127,
      "run_count": 17,
      "total_questions": 3329,
      "avg_cost_per_task_usd": 0.0037028555121658147,
      "avg_duration_per_task_ms": 233018.78792430158,
      "avg_output_tokens_per_task": 9785.001501952538,
      "last_run_timestamp": "2026-08-29T19:25:16.673Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DeepInfra",
      "accuracy": 0.7117850996055227,
      "accuracy_std_dev": 0.05549170236951472,
      "run_count": 31,
      "total_questions": 6084,
      "avg_cost_per_task_usd": 0.005924440664036818,
      "avg_duration_per_task_ms": 329601.12508218276,
      "avg_output_tokens_per_task": 18238.3634122288,
      "last_run_timestamp": "2026-08-29T19:12:34.710Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Parasail",
      "accuracy": 0.7117845333333335,
      "accuracy_std_dev": 0.015271985026475404,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.008412858585858586,
      "avg_duration_per_task_ms": 147396.4053872054,
      "avg_output_tokens_per_task": 4380.246127946128,
      "last_run_timestamp": "2026-08-29T18:30:44.353Z"
    },
    {
      "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-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.7112795,
      "accuracy_std_dev": 0.016430227680254466,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.006384943181818182,
      "avg_duration_per_task_ms": 104150.93529040404,
      "avg_output_tokens_per_task": 4141.536616161616,
      "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-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Ionstream",
      "accuracy": 0.7108986701231368,
      "accuracy_std_dev": 0.06666024694366084,
      "run_count": 16,
      "total_questions": 3086,
      "avg_cost_per_task_usd": 0.005044747245625405,
      "avg_duration_per_task_ms": 89902.54601425794,
      "avg_output_tokens_per_task": 6253.540181464679,
      "last_run_timestamp": "2026-08-08T16:46:11.805Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-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": 0.7102692,
      "accuracy_std_dev": 0.01356127233558859,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.006768369696969696,
      "avg_duration_per_task_ms": 38096.08383838384,
      "avg_output_tokens_per_task": 4479.758080808081,
      "last_run_timestamp": "2026-06-30T19:34:58.612Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-27b-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-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.70942765,
      "accuracy_std_dev": 0.01578725575986846,
      "run_count": 20,
      "total_questions": 3960,
      "avg_cost_per_task_usd": 0.0038632861111111117,
      "avg_duration_per_task_ms": 70598.0202020202,
      "avg_output_tokens_per_task": 3586.676515151515,
      "last_run_timestamp": "2026-08-29T19:09:55.303Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-next-80b-a3b-instruct-2509",
      "model_slug": "qwen/qwen3-next-80b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Next 80B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.7088045025693731,
      "accuracy_std_dev": 0.022963109072306993,
      "run_count": 15,
      "total_questions": 2919,
      "avg_cost_per_task_usd": 0.011898455292908531,
      "avg_duration_per_task_ms": 66441.3932853717,
      "avg_output_tokens_per_task": 7873.553956834532,
      "last_run_timestamp": "2026-08-29T19:12:26.066Z"
    },
    {
      "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-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Parasail",
      "accuracy": 0.7084551708346867,
      "accuracy_std_dev": 0.05060517580742111,
      "run_count": 16,
      "total_questions": 3079,
      "avg_cost_per_task_usd": 0.00587215816823644,
      "avg_duration_per_task_ms": 124679.61123741475,
      "avg_output_tokens_per_task": 7701.0337772003895,
      "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-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": 0.7083467052631579,
      "accuracy_std_dev": 0.019685740692424424,
      "run_count": 16,
      "total_questions": 3135,
      "avg_cost_per_task_usd": 0.0076249078149920255,
      "avg_duration_per_task_ms": 77954.19457735248,
      "avg_output_tokens_per_task": 7319.329505582137,
      "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": "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": "DeepInfra",
      "accuracy": 0.7065186345004272,
      "accuracy_std_dev": 0.01590026262623776,
      "run_count": 18,
      "total_questions": 3513,
      "avg_cost_per_task_usd": 0.007040196982635925,
      "avg_duration_per_task_ms": 99879.27754056362,
      "avg_output_tokens_per_task": 6356.3692001138625,
      "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "CoreWeave",
      "accuracy": 0.7050869311010947,
      "accuracy_std_dev": 0.036824859327421404,
      "run_count": 8,
      "total_questions": 1553,
      "avg_cost_per_task_usd": 0.01090031036703155,
      "avg_duration_per_task_ms": 296639.77012234385,
      "avg_output_tokens_per_task": 47139.87057308435,
      "last_run_timestamp": "2026-09-26T08:54:38.846Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "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/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AkashML",
      "accuracy": 0.7026935333333333,
      "accuracy_std_dev": 0.2370147256769409,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.015571158585858585,
      "avg_duration_per_task_ms": 188449.88585858585,
      "avg_output_tokens_per_task": 15387.030976430977,
      "last_run_timestamp": "2026-08-15T15:24:35.389Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.7024411666666667,
      "accuracy_std_dev": 0.14218853509503576,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.013173934343434343,
      "avg_duration_per_task_ms": 786438.8409090909,
      "avg_output_tokens_per_task": 27569.127104377105,
      "last_run_timestamp": "2026-09-22T02:20:03.900Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_slug": "qwen/qwen3-vl-235b-a22b-instruct",
      "model_name": "Qwen: Qwen3 VL 235B A22B Instruct",
      "provider_name": "Venice",
      "accuracy": 0.7023595287107259,
      "accuracy_std_dev": 0.038764278307667294,
      "run_count": 14,
      "total_questions": 2769,
      "avg_cost_per_task_usd": 0.009986093535572407,
      "avg_duration_per_task_ms": 174032.60635608522,
      "avg_output_tokens_per_task": 5593.384615384615,
      "last_run_timestamp": "2026-08-29T19:24:48.378Z"
    },
    {
      "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": 0.7020202666666666,
      "accuracy_std_dev": 0.017035677274342683,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.006572137373737374,
      "avg_duration_per_task_ms": 177727.7595959596,
      "avg_output_tokens_per_task": 4611.837373737374,
      "last_run_timestamp": "2026-08-29T18:14:24.357Z"
    },
    {
      "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": "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-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": 0.6953852941176472,
      "accuracy_std_dev": 0.04336719956205002,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.0073056770647653025,
      "avg_duration_per_task_ms": 36604.64824717766,
      "avg_output_tokens_per_task": 6041.8909685086155,
      "last_run_timestamp": "2026-08-29T18:10:09.843Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6948820649606299,
      "accuracy_std_dev": 0.027729042227588305,
      "run_count": 8,
      "total_questions": 1524,
      "avg_cost_per_task_usd": 0.015564356299212597,
      "avg_duration_per_task_ms": 765563.3963254594,
      "avg_output_tokens_per_task": 77634.76837270342,
      "last_run_timestamp": "2026-09-26T16:34:21.219Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Friendli",
      "accuracy": 0.69307675,
      "accuracy_std_dev": 0.04288574538162418,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.004385503156565657,
      "avg_duration_per_task_ms": 188413.82133838383,
      "avg_output_tokens_per_task": 5338.457702020202,
      "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": "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": 0.6901446339066337,
      "accuracy_std_dev": 0.20455115646309735,
      "run_count": 13,
      "total_questions": 2442,
      "avg_cost_per_task_usd": 0.042357015970515964,
      "avg_duration_per_task_ms": 640141.3783783783,
      "avg_output_tokens_per_task": 14232.427927927927,
      "last_run_timestamp": "2026-08-22T17:45:06.645Z"
    },
    {
      "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": 0.6898169582355701,
      "accuracy_std_dev": 0.20219361577785366,
      "run_count": 11,
      "total_questions": 2131,
      "avg_cost_per_task_usd": 0.015223487095260441,
      "avg_duration_per_task_ms": 380102.5335523228,
      "avg_output_tokens_per_task": 18493.574378226185,
      "last_run_timestamp": "2026-09-26T08:53:03.343Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "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": 0.6863611111111112,
      "accuracy_std_dev": 0.022997260124245744,
      "run_count": 6,
      "total_questions": 1134,
      "avg_cost_per_task_usd": 0.010325405643738976,
      "avg_duration_per_task_ms": 361986.35802469135,
      "avg_output_tokens_per_task": 46144.19135802469,
      "last_run_timestamp": "2026-09-26T08:53:22.738Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.6858586,
      "accuracy_std_dev": 0.014069007564146077,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.006483653535353535,
      "avg_duration_per_task_ms": 266458.26868686866,
      "avg_output_tokens_per_task": 16169.268686868687,
      "last_run_timestamp": "2026-05-23T17:26:25.014Z"
    },
    {
      "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": "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": 0.6804360395869191,
      "accuracy_std_dev": 0.04000822543143897,
      "run_count": 9,
      "total_questions": 1743,
      "avg_cost_per_task_usd": 0.015301245553643147,
      "avg_duration_per_task_ms": 489170.94836488814,
      "avg_output_tokens_per_task": 18010.746987951807,
      "last_run_timestamp": "2026-08-01T17:59:38.872Z"
    },
    {
      "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": 0.6792930000000001,
      "accuracy_std_dev": 0.07828270967780816,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.024729667929292928,
      "avg_duration_per_task_ms": 202686.41414141413,
      "avg_output_tokens_per_task": 6075.666666666667,
      "last_run_timestamp": "2026-06-13T09:16:40.692Z"
    },
    {
      "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": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "DeepInfra",
      "accuracy": 0.6765503619112165,
      "accuracy_std_dev": 0.12971631760570212,
      "run_count": 17,
      "total_questions": 2951,
      "avg_cost_per_task_usd": 0.025736548288715685,
      "avg_duration_per_task_ms": 193193.36191121655,
      "avg_output_tokens_per_task": 9449.462555066078,
      "last_run_timestamp": "2026-08-29T19:20:11.540Z"
    },
    {
      "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": "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": 0.6747475999999999,
      "accuracy_std_dev": 0.013775898004848904,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0017936818181818182,
      "avg_duration_per_task_ms": 43213.43131313131,
      "avg_output_tokens_per_task": 5783.8050505050505,
      "last_run_timestamp": "2026-05-30T08:06:07.683Z"
    },
    {
      "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": 0.6742425,
      "accuracy_std_dev": 0.2011785,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.03395365151515151,
      "avg_duration_per_task_ms": 248299.7626262626,
      "avg_output_tokens_per_task": 37528.89898989899,
      "last_run_timestamp": "2026-09-26T11:57:25.011Z"
    },
    {
      "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": 0.6741976266666667,
      "accuracy_std_dev": 0.04245737107063498,
      "run_count": 14,
      "total_questions": 2700,
      "avg_cost_per_task_usd": 0.01499391444444444,
      "avg_duration_per_task_ms": 350545.94703703705,
      "avg_output_tokens_per_task": 17958.764814814815,
      "last_run_timestamp": "2026-08-01T18:07:49.346Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Venice",
      "accuracy": 0.6731558206967214,
      "accuracy_std_dev": 0.0760641946679835,
      "run_count": 15,
      "total_questions": 2928,
      "avg_cost_per_task_usd": 0.09475717725409835,
      "avg_duration_per_task_ms": 785024.8715846995,
      "avg_output_tokens_per_task": 24226.864412568306,
      "last_run_timestamp": "2026-08-22T17:24:19.901Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Cerebras",
      "accuracy": 0.6708310526315789,
      "accuracy_std_dev": 0.13484993870048123,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.043008509303561926,
      "avg_duration_per_task_ms": 55485.30568846358,
      "avg_output_tokens_per_task": 15637.817916002126,
      "last_run_timestamp": "2026-08-15T14:46:10.940Z"
    },
    {
      "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": 0.6689419784982935,
      "accuracy_std_dev": 0.1773412026363194,
      "run_count": 16,
      "total_questions": 2930,
      "avg_cost_per_task_usd": 0.0037005590443686008,
      "avg_duration_per_task_ms": 131354.5737201365,
      "avg_output_tokens_per_task": 4676.710921501706,
      "last_run_timestamp": "2026-08-29T19:47:54.565Z"
    },
    {
      "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": 0.6685414863824749,
      "accuracy_std_dev": 0.08727923360547458,
      "run_count": 19,
      "total_questions": 3378,
      "avg_cost_per_task_usd": 0.006985365304914151,
      "avg_duration_per_task_ms": 34927.65926583777,
      "avg_output_tokens_per_task": 4604.6335109532265,
      "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": 0.664983,
      "accuracy_std_dev": 0.045452036199346076,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.008639549663299663,
      "avg_duration_per_task_ms": 254582.55808080808,
      "avg_output_tokens_per_task": 42899.02777777778,
      "last_run_timestamp": "2026-09-26T09:38:02.897Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "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": "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": 0.6603752184024266,
      "accuracy_std_dev": 0.09735619534024986,
      "run_count": 15,
      "total_questions": 2967,
      "avg_cost_per_task_usd": 0.0037568466464442197,
      "avg_duration_per_task_ms": 226109.69632625548,
      "avg_output_tokens_per_task": 4143.123019885406,
      "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-4.5-air",
      "model_slug": "z-ai/glm-4.5-air",
      "model_name": "Z.ai: GLM 4.5 Air",
      "provider_name": "Z.AI",
      "accuracy": 0.6592700368271954,
      "accuracy_std_dev": 0.049300236020255434,
      "run_count": 11,
      "total_questions": 2118,
      "avg_cost_per_task_usd": 0.0190812327667611,
      "avg_duration_per_task_ms": 403168.8569405099,
      "avg_output_tokens_per_task": 18089.26912181303,
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Venice",
      "accuracy": 0.6571402410865875,
      "accuracy_std_dev": 0.017808193342575305,
      "run_count": 18,
      "total_questions": 3534,
      "avg_cost_per_task_usd": 0.023573010186757212,
      "avg_duration_per_task_ms": 539665.155065082,
      "avg_output_tokens_per_task": 12662.279852857951,
      "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-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Google",
      "accuracy": 0.6569729516129034,
      "accuracy_std_dev": 0.17122119291573554,
      "run_count": 21,
      "total_questions": 4092,
      "avg_cost_per_task_usd": 0.01374512854349951,
      "avg_duration_per_task_ms": 244855.71725317693,
      "avg_output_tokens_per_task": 9153.359237536657,
      "last_run_timestamp": "2026-08-29T19:03:14.914Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6556188125000002,
      "accuracy_std_dev": 0.01847027483288604,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.0016670344065656568,
      "avg_duration_per_task_ms": 63016.68939393939,
      "avg_output_tokens_per_task": 1545.2768308080808,
      "last_run_timestamp": "2026-08-29T20:54:16.286Z"
    },
    {
      "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": 0.655513375,
      "accuracy_std_dev": 0.1620210868528673,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.0030583200757575758,
      "avg_duration_per_task_ms": 54577.18244949495,
      "avg_output_tokens_per_task": 1983.9665404040404,
      "last_run_timestamp": "2026-07-01T02:34:47.176Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Darkbloom",
      "accuracy": 0.653757604130809,
      "accuracy_std_dev": 0.015316614138466363,
      "run_count": 3,
      "total_questions": 581,
      "avg_cost_per_task_usd": 0.01179434079173838,
      "avg_duration_per_task_ms": 1792911.3149741825,
      "avg_output_tokens_per_task": 50132.19449225473,
      "last_run_timestamp": "2026-08-29T20:13:41.890Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.6536547169042769,
      "accuracy_std_dev": 0.019981737763823795,
      "run_count": 15,
      "total_questions": 2946,
      "avg_cost_per_task_usd": 0.004769026137135098,
      "avg_duration_per_task_ms": 99040.27562797013,
      "avg_output_tokens_per_task": 8986.41310251188,
      "last_run_timestamp": "2026-09-26T08:51:07.739Z"
    },
    {
      "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": 0.6519080986717266,
      "accuracy_std_dev": 0.0160056211434452,
      "run_count": 8,
      "total_questions": 1581,
      "avg_cost_per_task_usd": 0.0014605863377609109,
      "avg_duration_per_task_ms": 61151.05566097407,
      "avg_output_tokens_per_task": 4710.759013282733,
      "last_run_timestamp": "2026-07-18T08:33:31.052Z"
    },
    {
      "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": "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": 0.6498354401756312,
      "accuracy_std_dev": 0.019005924168263848,
      "run_count": 14,
      "total_questions": 2733,
      "avg_cost_per_task_usd": 0.006061401024515184,
      "avg_duration_per_task_ms": 193666.52616172703,
      "avg_output_tokens_per_task": 9966.588730332967,
      "last_run_timestamp": "2026-09-26T08:55:02.346Z"
    },
    {
      "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": "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": "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": "AtlasCloud",
      "accuracy": 0.6468015,
      "accuracy_std_dev": 0.02305589841342123,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.0008760272727272729,
      "avg_duration_per_task_ms": 28080.331818181818,
      "avg_output_tokens_per_task": 923.8237373737373,
      "last_run_timestamp": "2026-07-01T02:41:54.492Z"
    },
    {
      "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": "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": "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": 0.6450775350966429,
      "accuracy_std_dev": 0.07933459046326091,
      "run_count": 15,
      "total_questions": 2949,
      "avg_cost_per_task_usd": 0.006724317734825365,
      "avg_duration_per_task_ms": 642083.2651746355,
      "avg_output_tokens_per_task": 19917.21973550356,
      "last_run_timestamp": "2026-08-29T19:15:54.652Z"
    },
    {
      "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": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "Io Net",
      "accuracy": 0.6425364444444446,
      "accuracy_std_dev": 0.04264131881197732,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.03682436419753086,
      "avg_duration_per_task_ms": 255794.32547699215,
      "avg_output_tokens_per_task": 12516.43265993266,
      "last_run_timestamp": "2026-06-20T09:54:07.719Z"
    },
    {
      "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": "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/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Nebius",
      "accuracy": 0.6407338178368123,
      "accuracy_std_dev": 0.01094255224010119,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.0017736578115117015,
      "avg_duration_per_task_ms": 97733.33301707779,
      "avg_output_tokens_per_task": 5732.58697027198,
      "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": "auto-routing",
      "accuracy": 0.6405006706484643,
      "accuracy_std_dev": 0.017184977086752887,
      "run_count": 18,
      "total_questions": 3516,
      "avg_cost_per_task_usd": 0.0015951729237770197,
      "avg_duration_per_task_ms": 87221.02275312855,
      "avg_output_tokens_per_task": 5581.319397042093,
      "last_run_timestamp": "2026-08-29T19:32:59.441Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "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": "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": 0.6393443457526082,
      "accuracy_std_dev": 0.041002002642534584,
      "run_count": 21,
      "total_questions": 4026,
      "avg_cost_per_task_usd": 0.01295243964232489,
      "avg_duration_per_task_ms": 568425.9622454048,
      "avg_output_tokens_per_task": 33286.69672131148,
      "last_run_timestamp": "2026-08-30T00:32:50.853Z"
    },
    {
      "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": 0.6379928049155147,
      "accuracy_std_dev": 0.03271804644949137,
      "run_count": 11,
      "total_questions": 1953,
      "avg_cost_per_task_usd": 0.01783217153097798,
      "avg_duration_per_task_ms": 896687.3640552995,
      "avg_output_tokens_per_task": 74524.547875064,
      "last_run_timestamp": "2026-09-26T16:32:03.216Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6379416006600662,
      "accuracy_std_dev": 0.01704932395167573,
      "run_count": 14,
      "total_questions": 2727,
      "avg_cost_per_task_usd": 0.0018401910524385766,
      "avg_duration_per_task_ms": 156874.9156582325,
      "avg_output_tokens_per_task": 5961.517785111844,
      "last_run_timestamp": "2026-08-29T18:40:25.942Z"
    },
    {
      "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": "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Darkbloom",
      "accuracy": 0.6359762020460358,
      "accuracy_std_dev": 0.023468612837149117,
      "run_count": 4,
      "total_questions": 782,
      "avg_cost_per_task_usd": 0.006740489769820971,
      "avg_duration_per_task_ms": 885605.4833759591,
      "avg_output_tokens_per_task": 37192.53964194373,
      "last_run_timestamp": "2026-09-26T08:26:39.364Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.6358467809621291,
      "accuracy_std_dev": 0.02889634778968133,
      "run_count": 15,
      "total_questions": 2931,
      "avg_cost_per_task_usd": 0.006419597407028318,
      "avg_duration_per_task_ms": 105406.46980552713,
      "avg_output_tokens_per_task": 8959.577959740704,
      "last_run_timestamp": "2026-09-26T09:22:12.174Z"
    },
    {
      "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": "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": 0.6333203317702227,
      "accuracy_std_dev": 0.07516813236760066,
      "run_count": 13,
      "total_questions": 2559,
      "avg_cost_per_task_usd": 0.014011412270418131,
      "avg_duration_per_task_ms": 375980.625244236,
      "avg_output_tokens_per_task": 16853.191871824933,
      "last_run_timestamp": "2026-08-01T18:25:56.196Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6327843526405451,
      "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "WandB",
      "accuracy": 0.6326842164948453,
      "accuracy_std_dev": 0.02330257323945676,
      "run_count": 9,
      "total_questions": 1746,
      "avg_cost_per_task_usd": 0.0025534261168384882,
      "avg_duration_per_task_ms": 45796.38602520046,
      "avg_output_tokens_per_task": 1364.9232531500572,
      "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": "Crusoe",
      "accuracy": 0.6321081114302126,
      "accuracy_std_dev": 0.011032621401446443,
      "run_count": 9,
      "total_questions": 1741,
      "avg_cost_per_task_usd": 0.017073134979896612,
      "avg_duration_per_task_ms": 470536.2297530155,
      "avg_output_tokens_per_task": 85368.48765077542,
      "last_run_timestamp": "2026-09-26T16:13:51.663Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.6313133333333334,
      "accuracy_std_dev": 0.010910326968315643,
      "run_count": 6,
      "total_questions": 1188,
      "avg_cost_per_task_usd": 0.0011066590909090909,
      "avg_duration_per_task_ms": 52609.71380471381,
      "avg_output_tokens_per_task": 3573.563973063973,
      "last_run_timestamp": "2026-06-30T19:57:58.812Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "CoreWeave",
      "accuracy": 0.6308680607902736,
      "accuracy_std_dev": 0.011456259937693471,
      "run_count": 5,
      "total_questions": 987,
      "avg_cost_per_task_usd": 0.0034108257345491387,
      "avg_duration_per_task_ms": 140057.4062816616,
      "avg_output_tokens_per_task": 11472.941236068895,
      "last_run_timestamp": "2026-08-29T18:20:19.448Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-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": "Nebius",
      "accuracy": 0.6286550506072877,
      "accuracy_std_dev": 0.02141189626560173,
      "run_count": 15,
      "total_questions": 2964,
      "avg_cost_per_task_usd": 0.00387920310391363,
      "avg_duration_per_task_ms": 404484.17341430497,
      "avg_output_tokens_per_task": 12837.82354925776,
      "last_run_timestamp": "2026-08-29T19:06:01.882Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.626984142857143,
      "accuracy_std_dev": 0.01220184110731506,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.003761468975468975,
      "avg_duration_per_task_ms": 15627.561327561327,
      "avg_output_tokens_per_task": 918.3910533910534,
      "last_run_timestamp": "2026-06-30T19:49:03.931Z"
    },
    {
      "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": 0.626636888888889,
      "accuracy_std_dev": 0.010990047714429298,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.00803346857463524,
      "avg_duration_per_task_ms": 245092.09932659933,
      "avg_output_tokens_per_task": 19284.18069584736,
      "last_run_timestamp": "2026-07-18T16:20:00.500Z"
    },
    {
      "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": 0.626263,
      "accuracy_std_dev": 0.021820653931539237,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.02660411447811448,
      "avg_duration_per_task_ms": 192407.0723905724,
      "avg_output_tokens_per_task": 8826.355218855218,
      "last_run_timestamp": "2026-05-23T08:06:11.911Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.6248197142857143,
      "accuracy_std_dev": 0.11942661157585356,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.6020819559884559,
      "avg_duration_per_task_ms": 624611.841991342,
      "avg_output_tokens_per_task": 40107.266955266954,
      "last_run_timestamp": "2026-09-26T08:54:32.678Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.6221339996023858,
      "accuracy_std_dev": 0.024618022699500797,
      "run_count": 13,
      "total_questions": 2515,
      "avg_cost_per_task_usd": 0.007214242942345925,
      "avg_duration_per_task_ms": 448969.5920477137,
      "avg_output_tokens_per_task": 11841.8648111332,
      "last_run_timestamp": "2026-09-26T09:22:23.002Z"
    },
    {
      "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": 0.621595600867679,
      "accuracy_std_dev": 0.01951894601537237,
      "run_count": 7,
      "total_questions": 1383,
      "avg_cost_per_task_usd": 0.008186916847433116,
      "avg_duration_per_task_ms": 220211.67172812726,
      "avg_output_tokens_per_task": 6891.177151120752,
      "last_run_timestamp": "2026-07-01T03:52:27.932Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.621418478413069,
      "accuracy_std_dev": 0.013260113971771842,
      "run_count": 13,
      "total_questions": 2571,
      "avg_cost_per_task_usd": 0.004479952158693115,
      "avg_duration_per_task_ms": 91742.81602489304,
      "avg_output_tokens_per_task": 8509.088681446909,
      "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": 0.6196523681241185,
      "accuracy_std_dev": 0.014509665172618149,
      "run_count": 11,
      "total_questions": 2127,
      "avg_cost_per_task_usd": 0.004821758345086977,
      "avg_duration_per_task_ms": 191599.74424071462,
      "avg_output_tokens_per_task": 9720.46450399624,
      "last_run_timestamp": "2026-08-01T18:16:00.374Z"
    },
    {
      "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": 0.6194785090464547,
      "accuracy_std_dev": 0.03307467351712581,
      "run_count": 11,
      "total_questions": 2045,
      "avg_cost_per_task_usd": 0.014328027872860635,
      "avg_duration_per_task_ms": 570163.035207824,
      "avg_output_tokens_per_task": 69958.57506112469,
      "last_run_timestamp": "2026-09-26T16:16:58.507Z"
    },
    {
      "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": 0.6182419367816093,
      "accuracy_std_dev": 0.020441088895802338,
      "run_count": 16,
      "total_questions": 3132,
      "avg_cost_per_task_usd": 0.0025628135376756065,
      "avg_duration_per_task_ms": 24595.393358876117,
      "avg_output_tokens_per_task": 1512.9431673052363,
      "last_run_timestamp": "2026-08-29T18:49:22.541Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6168832857142857,
      "accuracy_std_dev": 0.0274253947059201,
      "run_count": 7,
      "total_questions": 1386,
      "avg_cost_per_task_usd": 0.009174507215007215,
      "avg_duration_per_task_ms": 639672.284992785,
      "avg_output_tokens_per_task": 16259.07142857143,
      "last_run_timestamp": "2026-07-18T17:46:02.370Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6166527916666666,
      "accuracy_std_dev": 0.12145770743554425,
      "run_count": 24,
      "total_questions": 4752,
      "avg_cost_per_task_usd": 0.6553831144781144,
      "avg_duration_per_task_ms": 484995.4349747475,
      "avg_output_tokens_per_task": 40791.21506734007,
      "last_run_timestamp": "2026-09-26T09:11:13.767Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Venice",
      "accuracy": 0.616162,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0017577828282828282,
      "avg_duration_per_task_ms": 47302.35858585859,
      "avg_output_tokens_per_task": 2472.939393939394,
      "last_run_timestamp": "2026-05-16T08:06:13.049Z"
    },
    {
      "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": 0.6145900645161291,
      "accuracy_std_dev": 0.017021637149063382,
      "run_count": 16,
      "total_questions": 3162,
      "avg_cost_per_task_usd": 0.0051391173308032905,
      "avg_duration_per_task_ms": 224377.06388361796,
      "avg_output_tokens_per_task": 12215.357368753954,
      "last_run_timestamp": "2026-08-29T19:52:22.642Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.614478375,
      "accuracy_std_dev": 0.12366286539715297,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.6154596931818181,
      "avg_duration_per_task_ms": 632046.8207070708,
      "avg_output_tokens_per_task": 40881.84659090909,
      "last_run_timestamp": "2026-09-26T09:34:52.233Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6139495138993238,
      "accuracy_std_dev": 0.029024384950147974,
      "run_count": 8,
      "total_questions": 1331,
      "avg_cost_per_task_usd": 0.02793150338091661,
      "avg_duration_per_task_ms": 583500.4891059354,
      "avg_output_tokens_per_task": 29316.752817430504,
      "last_run_timestamp": "2026-09-26T16:19:07.764Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6129581867007673,
      "accuracy_std_dev": 0.0015394999999999714,
      "run_count": 2,
      "total_questions": 391,
      "avg_cost_per_task_usd": 0.0032793452685421997,
      "avg_duration_per_task_ms": 346994.51406649617,
      "avg_output_tokens_per_task": 13256.76726342711,
      "last_run_timestamp": "2026-09-01T13:34:17.711Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Google",
      "accuracy": 0.6126859920896506,
      "accuracy_std_dev": 0.1240297456369234,
      "run_count": 23,
      "total_questions": 4551,
      "avg_cost_per_task_usd": 0.6596944438584925,
      "avg_duration_per_task_ms": 597097.6884201275,
      "avg_output_tokens_per_task": 41139.12920237311,
      "last_run_timestamp": "2026-09-26T09:32:36.461Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.6122767692307693,
      "accuracy_std_dev": 0.02357255235406653,
      "run_count": 13,
      "total_questions": 2574,
      "avg_cost_per_task_usd": 0.004552111888111887,
      "avg_duration_per_task_ms": 191391.34887334888,
      "avg_output_tokens_per_task": 9066.714452214452,
      "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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Alibaba",
      "accuracy": 0.6121935714285716,
      "accuracy_std_dev": 0.019793131697327618,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.00792913455988456,
      "avg_duration_per_task_ms": 24216.782467532466,
      "avg_output_tokens_per_task": 1511.1587301587301,
      "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": "Venice",
      "accuracy": 0.6119012322274883,
      "accuracy_std_dev": 0.019905936820709,
      "run_count": 13,
      "total_questions": 2532,
      "avg_cost_per_task_usd": 0.0026039996050552924,
      "avg_duration_per_task_ms": 47372.351105845184,
      "avg_output_tokens_per_task": 1655.2235387045814,
      "last_run_timestamp": "2026-08-29T19:00:24.068Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3",
      "model_slug": "deepseek/deepseek-chat",
      "model_name": "DeepSeek: DeepSeek V3",
      "provider_name": "auto-routing",
      "accuracy": 0.609428,
      "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": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Azure",
      "accuracy": 0.6087965,
      "accuracy_std_dev": 0.12831918473770787,
      "run_count": 8,
      "total_questions": 1584,
      "avg_cost_per_task_usd": 0.6162571098484848,
      "avg_duration_per_task_ms": 633013.7184343435,
      "avg_output_tokens_per_task": 41027.7380050505,
      "last_run_timestamp": "2026-09-26T09:34:56.525Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "DeepInfra",
      "accuracy": 0.6086993033591731,
      "accuracy_std_dev": 0.019442174128425027,
      "run_count": 10,
      "total_questions": 1935,
      "avg_cost_per_task_usd": 0.001751839793281654,
      "avg_duration_per_task_ms": 39301.46408268734,
      "avg_output_tokens_per_task": 1636.5943152454781,
      "last_run_timestamp": "2026-08-29T18:37:17.171Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Together",
      "accuracy": 0.608586,
      "accuracy_std_dev": 0.0025249999999999995,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0020087575757575756,
      "avg_duration_per_task_ms": 10469.585858585859,
      "avg_output_tokens_per_task": 733.570707070707,
      "last_run_timestamp": "2026-05-23T08:09:18.247Z"
    },
    {
      "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": "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "auto-routing",
      "accuracy": 0.6058966877256319,
      "accuracy_std_dev": 0.030100999151127752,
      "run_count": 17,
      "total_questions": 3324,
      "avg_cost_per_task_usd": 0.0021504960890493382,
      "avg_duration_per_task_ms": 36972.96871239471,
      "avg_output_tokens_per_task": 1435.0361010830325,
      "last_run_timestamp": "2026-08-29T18:47:46.000Z"
    },
    {
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Venice",
      "accuracy": 0.6022078955765048,
      "accuracy_std_dev": 0.015661301991180567,
      "run_count": 21,
      "total_questions": 4137,
      "avg_cost_per_task_usd": 0.04455390645395214,
      "avg_duration_per_task_ms": 535138.5900410926,
      "avg_output_tokens_per_task": 16948.74981870921,
      "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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Google",
      "accuracy": 0.6020365852036575,
      "accuracy_std_dev": 0.04030630297634977,
      "run_count": 13,
      "total_questions": 2406,
      "avg_cost_per_task_usd": 0.0031222069825436404,
      "avg_duration_per_task_ms": 36515.22194513716,
      "avg_output_tokens_per_task": 1660.556940980881,
      "last_run_timestamp": "2026-08-29T18:43:09.218Z"
    },
    {
      "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": "qwen/qwen3-14b-04-28",
      "model_slug": "qwen/qwen3-14b",
      "model_name": "Qwen: Qwen3 14B",
      "provider_name": "auto-routing",
      "accuracy": 0.594276,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.002775611111111111,
      "avg_duration_per_task_ms": 509237.26767676766,
      "avg_output_tokens_per_task": 10972.9898989899,
      "last_run_timestamp": "2026-09-01T13:31:26.819Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.5922561000000001,
      "accuracy_std_dev": 0.11721268776753647,
      "run_count": 10,
      "total_questions": 1980,
      "avg_cost_per_task_usd": 0.6146375166666666,
      "avg_duration_per_task_ms": 629521.5186868687,
      "avg_output_tokens_per_task": 40864.73838383838,
      "last_run_timestamp": "2026-09-26T09:23:13.151Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "google/gemini-2.5-flash-lite",
      "model_slug": "google/gemini-2.5-flash-lite",
      "model_name": "Google: Gemini 2.5 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.5900675,
      "accuracy_std_dev": 0.007575500000000013,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.017035757575757576,
      "avg_duration_per_task_ms": 123031.02272727272,
      "avg_output_tokens_per_task": 42470.74747474748,
      "last_run_timestamp": "2026-09-01T18:10:49.087Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "CoreWeave",
      "accuracy": 0.5872781331360947,
      "accuracy_std_dev": 0.004687499999999956,
      "run_count": 2,
      "total_questions": 338,
      "avg_cost_per_task_usd": 0.09507673668639052,
      "avg_duration_per_task_ms": 191708.50591715975,
      "avg_output_tokens_per_task": 21455.86094674556,
      "last_run_timestamp": "2026-08-01T18:26:20.700Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "DeepInfra",
      "accuracy": 0.5871213125,
      "accuracy_std_dev": 0.03216602090078666,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.003158161616161616,
      "avg_duration_per_task_ms": 217609.65056818182,
      "avg_output_tokens_per_task": 11118.01957070707,
      "last_run_timestamp": "2026-08-29T19:16:45.297Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "Io Net",
      "accuracy": 0.577862,
      "accuracy_std_dev": 0.03661599999999998,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.006390335858585858,
      "avg_duration_per_task_ms": 667369.1111111111,
      "avg_output_tokens_per_task": 32604.280303030304,
      "last_run_timestamp": "2026-09-26T08:13:27.134Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Venice",
      "accuracy": 0.5773608170445662,
      "accuracy_std_dev": 0.036921629989749905,
      "run_count": 20,
      "total_questions": 3837,
      "avg_cost_per_task_usd": 0.05988801198853271,
      "avg_duration_per_task_ms": 462586.3312483711,
      "avg_output_tokens_per_task": 19012.92468074016,
      "last_run_timestamp": "2026-08-29T19:11:06.196Z"
    },
    {
      "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": "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": 0.5704424582524273,
      "accuracy_std_dev": 0.11994568871681048,
      "run_count": 16,
      "total_questions": 3090,
      "avg_cost_per_task_usd": 0.0021708019417475727,
      "avg_duration_per_task_ms": 65226.70291262136,
      "avg_output_tokens_per_task": 2007.0271844660194,
      "last_run_timestamp": "2026-08-29T20:57:58.443Z"
    },
    {
      "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": 0.5644344238110643,
      "accuracy_std_dev": 0.11775681330114128,
      "run_count": 16,
      "total_questions": 3091,
      "avg_cost_per_task_usd": 0.0663304202523455,
      "avg_duration_per_task_ms": 713891.3856357166,
      "avg_output_tokens_per_task": 21649.021675833064,
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Phala",
      "accuracy": 0.5536774896401309,
      "accuracy_std_dev": 0.11298267708866899,
      "run_count": 14,
      "total_questions": 2751,
      "avg_cost_per_task_usd": 0.03626482333696838,
      "avg_duration_per_task_ms": 497739.62849872775,
      "avg_output_tokens_per_task": 13499.503089785532,
      "last_run_timestamp": "2026-07-11T16:02:27.787Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-14b-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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Darkbloom",
      "accuracy": 0.5514706253196932,
      "accuracy_std_dev": 0.20217034155010072,
      "run_count": 8,
      "total_questions": 1564,
      "avg_cost_per_task_usd": 0.02285590537084399,
      "avg_duration_per_task_ms": 745462.75,
      "avg_output_tokens_per_task": 32123.599104859335,
      "last_run_timestamp": "2026-09-26T16:19:31.961Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "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": 0.5503437185539289,
      "accuracy_std_dev": 0.037362835232183034,
      "run_count": 19,
      "total_questions": 3347,
      "avg_cost_per_task_usd": 0.031049354048401548,
      "avg_duration_per_task_ms": 403541.892739767,
      "avg_output_tokens_per_task": 14846.570959067822,
      "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Phala",
      "accuracy": 0.5483589008379888,
      "accuracy_std_dev": 0.173972612705257,
      "run_count": 15,
      "total_questions": 2864,
      "avg_cost_per_task_usd": 0.03186034916201117,
      "avg_duration_per_task_ms": 397835.3034217877,
      "avg_output_tokens_per_task": 11537.356843575419,
      "last_run_timestamp": "2026-07-18T16:15:33.587Z"
    },
    {
      "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": 0.5475589166666667,
      "accuracy_std_dev": 0.10881973957303452,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.00421212079124579,
      "avg_duration_per_task_ms": 247814.2074915825,
      "avg_output_tokens_per_task": 11271.605218855218,
      "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": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "OpenAI",
      "accuracy": 0.542088,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.023943484848484846,
      "avg_duration_per_task_ms": 10793.646464646465,
      "avg_output_tokens_per_task": 1334.611111111111,
      "last_run_timestamp": "2026-08-01T08:53:34.329Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Novita",
      "accuracy": 0.53661625,
      "accuracy_std_dev": 0.018685337946836782,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.004247876262626262,
      "avg_duration_per_task_ms": 201941.1426767677,
      "avg_output_tokens_per_task": 9385.565656565657,
      "last_run_timestamp": "2026-05-23T11:11:04.972Z"
    },
    {
      "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": 0.5341293636363637,
      "accuracy_std_dev": 0.015229376769154397,
      "run_count": 11,
      "total_questions": 2178,
      "avg_cost_per_task_usd": 0.003491447199265381,
      "avg_duration_per_task_ms": 77240.00826446281,
      "avg_output_tokens_per_task": 7540.567033976125,
      "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": "qwen/qwen3-vl-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.5323234,
      "accuracy_std_dev": 0.01905887848326865,
      "run_count": 5,
      "total_questions": 990,
      "avg_cost_per_task_usd": 0.0038402525252525252,
      "avg_duration_per_task_ms": 124965.20303030303,
      "avg_output_tokens_per_task": 7619.874747474748,
      "last_run_timestamp": "2026-07-01T03:39:59.085Z"
    },
    {
      "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": 0.5315655,
      "accuracy_std_dev": 0.027285771058374,
      "run_count": 4,
      "total_questions": 792,
      "avg_cost_per_task_usd": 0.004162521464646465,
      "avg_duration_per_task_ms": 147513.7563131313,
      "avg_output_tokens_per_task": 8281.712121212122,
      "last_run_timestamp": "2026-06-20T17:38:01.923Z"
    },
    {
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "DeepInfra",
      "accuracy": 0.5303269819620254,
      "accuracy_std_dev": 0.1650574482984897,
      "run_count": 16,
      "total_questions": 3160,
      "avg_cost_per_task_usd": 0.0020356806962025316,
      "avg_duration_per_task_ms": 82497.10664556963,
      "avg_output_tokens_per_task": 2227.1496835443036,
      "last_run_timestamp": "2026-08-29T20:34:10.117Z"
    },
    {
      "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": "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": "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": 0.521297721396731,
      "accuracy_std_dev": 0.03986710324587105,
      "run_count": 22,
      "total_questions": 4038,
      "avg_cost_per_task_usd": 0.009685821198613175,
      "avg_duration_per_task_ms": 422937.02204061416,
      "avg_output_tokens_per_task": 24523.735264982664,
      "last_run_timestamp": "2026-08-30T00:11:29.373Z"
    },
    {
      "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": "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": 0.5137279018733274,
      "accuracy_std_dev": 0.06412042341716619,
      "run_count": 19,
      "total_questions": 3363,
      "avg_cost_per_task_usd": 0.06096582634552484,
      "avg_duration_per_task_ms": 510270.59768064227,
      "avg_output_tokens_per_task": 17365.03865596194,
      "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": "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": 0.5122718770053475,
      "accuracy_std_dev": 0.11548223967755335,
      "run_count": 14,
      "total_questions": 2431,
      "avg_cost_per_task_usd": 0.008256612093788564,
      "avg_duration_per_task_ms": 484397.99177293293,
      "avg_output_tokens_per_task": 41115.3813245578,
      "last_run_timestamp": "2026-09-26T16:17:38.509Z"
    },
    {
      "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": 0.5117335484247374,
      "accuracy_std_dev": 0.013140266237519679,
      "run_count": 13,
      "total_questions": 2571,
      "avg_cost_per_task_usd": 0.004239138467522365,
      "avg_duration_per_task_ms": 118277.81796966161,
      "avg_output_tokens_per_task": 8576.858031894204,
      "last_run_timestamp": "2026-08-01T18:08:24.732Z"
    },
    {
      "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": "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": "Groq",
      "accuracy": 0.4963358823529412,
      "accuracy_std_dev": 0.019652006075963045,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.0017086723113487819,
      "avg_duration_per_task_ms": 8000.373737373738,
      "avg_output_tokens_per_task": 1693.5855614973261,
      "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": "AkashML",
      "accuracy": 0.4951475294117647,
      "accuracy_std_dev": 0.023405011434501268,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.0007195903149138444,
      "avg_duration_per_task_ms": 45332.08585858586,
      "avg_output_tokens_per_task": 1560.3024361259656,
      "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": "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": "Inceptron",
      "accuracy": 0.489899,
      "accuracy_std_dev": 0.020202000000000025,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.0003088434343434343,
      "avg_duration_per_task_ms": 18505.795454545456,
      "avg_output_tokens_per_task": 721.5959595959596,
      "last_run_timestamp": "2026-05-16T08:12:45.858Z"
    },
    {
      "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": 0.48945589473684215,
      "accuracy_std_dev": 0.024456503773659585,
      "run_count": 19,
      "total_questions": 3762,
      "avg_cost_per_task_usd": 0.0008684388623072833,
      "avg_duration_per_task_ms": 44982.24348750665,
      "avg_output_tokens_per_task": 1513.1935140882508,
      "last_run_timestamp": "2026-08-29T18:24:23.944Z"
    },
    {
      "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": 0.48919281499107675,
      "accuracy_std_dev": 0.02018458247839152,
      "run_count": 17,
      "total_questions": 3362,
      "avg_cost_per_task_usd": 0.0008363173706127306,
      "avg_duration_per_task_ms": 58785.93456276026,
      "avg_output_tokens_per_task": 1903.9723378941108,
      "last_run_timestamp": "2026-08-29T19:00:36.673Z"
    },
    {
      "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": 0.4885771112490687,
      "accuracy_std_dev": 0.03899715161939785,
      "run_count": 21,
      "total_questions": 4027,
      "avg_cost_per_task_usd": 0.007579310404767817,
      "avg_duration_per_task_ms": 295718.4092376459,
      "avg_output_tokens_per_task": 16477.777253538614,
      "last_run_timestamp": "2026-08-30T00:23:23.003Z"
    },
    {
      "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": "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": "WandB",
      "accuracy": 0.48653191666666656,
      "accuracy_std_dev": 0.02063010630308019,
      "run_count": 12,
      "total_questions": 2376,
      "avg_cost_per_task_usd": 0.0013203434343434341,
      "avg_duration_per_task_ms": 19607.073232323233,
      "avg_output_tokens_per_task": 1411.3367003367002,
      "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": "Together",
      "accuracy": 0.48590056249999997,
      "accuracy_std_dev": 0.024073916665575915,
      "run_count": 16,
      "total_questions": 3168,
      "avg_cost_per_task_usd": 0.0023842935606060607,
      "avg_duration_per_task_ms": 42122.70233585859,
      "avg_output_tokens_per_task": 1688.7894570707072,
      "last_run_timestamp": "2026-08-29T18:45:20.686Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "z-ai/glm-4.7-flash-20260119",
      "model_slug": "z-ai/glm-4.7-flash",
      "model_name": "Z.ai: GLM 4.7 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.48383308151910986,
      "accuracy_std_dev": 0.0343561201621871,
      "run_count": 21,
      "total_questions": 4134,
      "avg_cost_per_task_usd": 0.006950208998548622,
      "avg_duration_per_task_ms": 299773.75689404935,
      "avg_output_tokens_per_task": 17335.253991291727,
      "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": "Nebius",
      "accuracy": 0.47486757142857144,
      "accuracy_std_dev": 0.011496839122834002,
      "run_count": 14,
      "total_questions": 2772,
      "avg_cost_per_task_usd": 0.000819739538239538,
      "avg_duration_per_task_ms": 122852.08513708513,
      "avg_output_tokens_per_task": 1852.6435786435786,
      "last_run_timestamp": "2026-08-29T20:09:29.346Z"
    },
    {
      "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": 0.4748159255751015,
      "accuracy_std_dev": 0.1131125589884548,
      "run_count": 12,
      "total_questions": 2217,
      "avg_cost_per_task_usd": 0.01081031709517366,
      "avg_duration_per_task_ms": 287884.16599007667,
      "avg_output_tokens_per_task": 14729.669373026612,
      "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": "DeepInfra",
      "accuracy": 0.4747473529411765,
      "accuracy_std_dev": 0.01954779462494329,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.0006343000594177065,
      "avg_duration_per_task_ms": 120340.93464052287,
      "avg_output_tokens_per_task": 1809.9459298871063,
      "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": "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": "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": 0.3889878823529412,
      "accuracy_std_dev": 0.2561627234269568,
      "run_count": 17,
      "total_questions": 3366,
      "avg_cost_per_task_usd": 0.0011359013666072488,
      "avg_duration_per_task_ms": 33787.26500297088,
      "avg_output_tokens_per_task": 967.2875816993464,
      "last_run_timestamp": "2026-08-29T18:43:09.114Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "tencent/hy-mt2-30b-a3b-20260521",
      "model_slug": "tencent/hy-mt2-30b-a3b",
      "model_name": "Tencent: Hy-MT2-30B-A3B",
      "provider_name": "Reka",
      "accuracy": 0.388889,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.00023147979797979797,
      "avg_duration_per_task_ms": 8604.353535353535,
      "avg_output_tokens_per_task": 710.540404040404,
      "last_run_timestamp": "2026-09-24T06:00:03.613Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.3860826666666667,
      "accuracy_std_dev": 0.005722799567422292,
      "run_count": 3,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.00015977104377104374,
      "avg_duration_per_task_ms": 6254.590909090909,
      "avg_output_tokens_per_task": 953.9612794612794,
      "last_run_timestamp": "2026-09-01T13:54:02.609Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.364571888888889,
      "accuracy_std_dev": 0.40360819680929966,
      "run_count": 9,
      "total_questions": 1782,
      "avg_cost_per_task_usd": 0.003561915263748597,
      "avg_duration_per_task_ms": 170742.85016835018,
      "avg_output_tokens_per_task": 3450.5415263748596,
      "last_run_timestamp": "2026-08-15T14:54:58.071Z"
    },
    {
      "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": "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": "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": 0.30662180000000006,
      "accuracy_std_dev": 0.048718266987787376,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.006156473063973065,
      "avg_duration_per_task_ms": 310379.08855218854,
      "avg_output_tokens_per_task": 13940.810774410775,
      "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": "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": "sao10k/l3-lunaris-8b",
      "model_slug": "sao10k/l3-lunaris-8b",
      "model_name": "Sao10K: Llama 3 8B Lunaris",
      "provider_name": "DeepInfra",
      "accuracy": 0.29040400000000005,
      "accuracy_std_dev": 0.004209000000000018,
      "run_count": 2,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.000060893939393939394,
      "avg_duration_per_task_ms": 9271.39393939394,
      "avg_output_tokens_per_task": 573.0479797979798,
      "last_run_timestamp": "2026-09-01T13:00:51.348Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.28953874164345406,
      "accuracy_std_dev": 0.024162932213900572,
      "run_count": 22,
      "total_questions": 4308,
      "avg_cost_per_task_usd": 0.00016581569173630456,
      "avg_duration_per_task_ms": 163891.54758588673,
      "avg_output_tokens_per_task": 3896.552460538533,
      "last_run_timestamp": "2026-08-29T16:57:41.646Z"
    },
    {
      "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": 0.28136913333333335,
      "accuracy_std_dev": 0.025692936660663933,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.0007517552188552191,
      "avg_duration_per_task_ms": 48215.582154882155,
      "avg_output_tokens_per_task": 9778.388552188551,
      "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": "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": 0.2692480666666667,
      "accuracy_std_dev": 0.018385093311218802,
      "run_count": 15,
      "total_questions": 2970,
      "avg_cost_per_task_usd": 0.0008740006734006734,
      "avg_duration_per_task_ms": 22071.978787878787,
      "avg_output_tokens_per_task": 10913.142760942761,
      "last_run_timestamp": "2026-08-29T18:45:59.793Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "sao10k/l3-lunaris-8b",
      "model_slug": "sao10k/l3-lunaris-8b",
      "model_name": "Sao10K: Llama 3 8B Lunaris",
      "provider_name": "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": "WandB",
      "accuracy": 0.25380578104993595,
      "accuracy_std_dev": 0.022641417314753803,
      "run_count": 12,
      "total_questions": 2343,
      "avg_cost_per_task_usd": 0.0007069568928723856,
      "avg_duration_per_task_ms": 20649.668373879642,
      "avg_output_tokens_per_task": 2769.859581732821,
      "last_run_timestamp": "2026-07-18T08:26:34.893Z"
    },
    {
      "benchmark": "gpqa_diamond",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/gpqa-diamond",
      "model_permaslug": "meta-llama/llama-3.1-8b-instruct",
      "model_slug": "meta-llama/llama-3.1-8b-instruct",
      "model_name": "Meta: Llama 3.1 8B Instruct",
      "provider_name": "Cloudflare",
      "accuracy": 0.25143756923076926,
      "accuracy_std_dev": 0.08124746081830182,
      "run_count": 11,
      "total_questions": 2145,
      "avg_cost_per_task_usd": 0.0014683734265734264,
      "avg_duration_per_task_ms": 503997.0027972028,
      "avg_output_tokens_per_task": 5763.862937062937,
      "last_run_timestamp": "2026-08-29T17:16:14.290Z"
    },
    {
      "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": "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": "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": "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": 0.0038196250000000005,
      "accuracy_std_dev": 0.0032291292324711936,
      "run_count": 5,
      "total_questions": 960,
      "avg_cost_per_task_usd": 0.003103454166666667,
      "avg_duration_per_task_ms": 152207.65520833334,
      "avg_output_tokens_per_task": 2930.8010416666666,
      "last_run_timestamp": "2026-08-22T17:29:06.480Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "BaseTen",
      "accuracy": 0.822917,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.0087319375,
      "avg_duration_per_task_ms": 31283.875,
      "avg_output_tokens_per_task": 9903.5625,
      "last_run_timestamp": "2026-09-26T08:05:12.798Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google AI Studio",
      "accuracy": 0.806667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.20567228,
      "avg_duration_per_task_ms": 69455.26,
      "avg_output_tokens_per_task": 15737.28,
      "last_run_timestamp": "2026-08-15T00:15:11.753Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.7-flash-20260813",
      "model_slug": "google/gemini-3.7-flash",
      "model_name": "Google: Gemini 3.7 Flash",
      "provider_name": "Google",
      "accuracy": 0.805556,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.07709047916666667,
      "avg_duration_per_task_ms": 121300.85416666667,
      "avg_output_tokens_per_task": 14694.6875,
      "last_run_timestamp": "2026-08-15T00:15:15.845Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Azure",
      "accuracy": 0.804166625,
      "accuracy_std_dev": 0.021065488167483215,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.8715094075,
      "avg_duration_per_task_ms": 137048.4925,
      "avg_output_tokens_per_task": 5786.43,
      "last_run_timestamp": "2026-09-26T08:21:11.883Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.8,
      "accuracy_std_dev": 0.033110611481316166,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.7322859599999999,
      "avg_duration_per_task_ms": 146639.07333333333,
      "avg_output_tokens_per_task": 5472.34,
      "last_run_timestamp": "2026-08-15T08:09:14.229Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Fireworks",
      "accuracy": 0.8,
      "accuracy_std_dev": 0.019999999999999962,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.030081680000000003,
      "avg_duration_per_task_ms": 53399.74,
      "avg_output_tokens_per_task": 3282.63,
      "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": "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": "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": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "auto-routing",
      "accuracy": 0.795833375,
      "accuracy_std_dev": 0.02846769733811245,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.9313334725,
      "avg_duration_per_task_ms": 130292.9175,
      "avg_output_tokens_per_task": 5748.73,
      "last_run_timestamp": "2026-09-26T08:18:08.440Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "auto-routing",
      "accuracy": 0.7933333333333333,
      "accuracy_std_dev": 0.014401440259764123,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.6872690333333334,
      "avg_duration_per_task_ms": 114143.74666666667,
      "avg_output_tokens_per_task": 6925.7266666666665,
      "last_run_timestamp": "2026-09-26T08:22:07.944Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7911111666666667,
      "accuracy_std_dev": 0.019116304205369344,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.8711063166666666,
      "avg_duration_per_task_ms": 151211.67333333334,
      "avg_output_tokens_per_task": 5696.93,
      "last_run_timestamp": "2026-09-26T08:09:55.578Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Anthropic",
      "accuracy": 0.790833375,
      "accuracy_std_dev": 0.01630862714897777,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.8205446474999999,
      "avg_duration_per_task_ms": 138181.955,
      "avg_output_tokens_per_task": 5731.4725,
      "last_run_timestamp": "2026-09-26T08:05:12.739Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Crusoe",
      "accuracy": 0.788889,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08521802666666665,
      "avg_duration_per_task_ms": 39816.48,
      "avg_output_tokens_per_task": 4684.933333333333,
      "last_run_timestamp": "2026-09-26T08:13:10.629Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Anthropic",
      "accuracy": 0.78833325,
      "accuracy_std_dev": 0.009860203404975974,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.8196274600000001,
      "avg_duration_per_task_ms": 183041.66,
      "avg_output_tokens_per_task": 7098.595,
      "last_run_timestamp": "2026-09-26T08:15:33.263Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-5-fable-20260609",
      "model_slug": "anthropic/claude-fable-5",
      "model_name": "Anthropic: Claude Fable 5",
      "provider_name": "Google",
      "accuracy": 0.7877778666666666,
      "accuracy_std_dev": 0.02049995242552745,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.9492467306666665,
      "avg_duration_per_task_ms": 98954.73466666667,
      "avg_output_tokens_per_task": 5602.965333333334,
      "last_run_timestamp": "2026-09-26T08:12:28.628Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7875,
      "accuracy_std_dev": 0.024593980747329203,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.48769549000000006,
      "avg_duration_per_task_ms": 129773.455,
      "avg_output_tokens_per_task": 7130.5275,
      "last_run_timestamp": "2026-09-26T08:22:26.398Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Wafer",
      "accuracy": 0.7870992438478748,
      "accuracy_std_dev": 0.008959550928450943,
      "run_count": 9,
      "total_questions": 447,
      "avg_cost_per_task_usd": 0.11607393959731543,
      "avg_duration_per_task_ms": 143483.8120805369,
      "avg_output_tokens_per_task": 11288.324384787473,
      "last_run_timestamp": "2026-09-26T08:27:30.005Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "Google AI Studio",
      "accuracy": 0.7866665,
      "accuracy_std_dev": 0.026666499999999982,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.12536731,
      "avg_duration_per_task_ms": 118841.15,
      "avg_output_tokens_per_task": 19578.58,
      "last_run_timestamp": "2026-09-01T17:30:58.594Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Crusoe",
      "accuracy": 0.7859532642140468,
      "accuracy_std_dev": 0.022060806735148097,
      "run_count": 6,
      "total_questions": 299,
      "avg_cost_per_task_usd": 0.028781180602006688,
      "avg_duration_per_task_ms": 346700.6989966555,
      "avg_output_tokens_per_task": 11874.802675585284,
      "last_run_timestamp": "2026-08-29T17:06:21.366Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "BaseTen",
      "accuracy": 0.7857143728222997,
      "accuracy_std_dev": 0.022636034639976627,
      "run_count": 6,
      "total_questions": 287,
      "avg_cost_per_task_usd": 0.11757857142857142,
      "avg_duration_per_task_ms": 73705.68641114983,
      "avg_output_tokens_per_task": 8241.627177700348,
      "last_run_timestamp": "2026-08-29T17:15:26.910Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DeepInfra",
      "accuracy": 0.784471297188755,
      "accuracy_std_dev": 0.02769278737577711,
      "run_count": 5,
      "total_questions": 249,
      "avg_cost_per_task_usd": 0.09832240562248996,
      "avg_duration_per_task_ms": 406639.9036144578,
      "avg_output_tokens_per_task": 17335.586345381525,
      "last_run_timestamp": "2026-09-26T08:23:09.250Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Morph",
      "accuracy": 0.784,
      "accuracy_std_dev": 0.04319121652836371,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.082875132,
      "avg_duration_per_task_ms": 213022.804,
      "avg_output_tokens_per_task": 11102.478666666666,
      "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": "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": 0.7831884347826087,
      "accuracy_std_dev": 0.03796104117304211,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.10458759999999999,
      "avg_duration_per_task_ms": 387959.7686956522,
      "avg_output_tokens_per_task": 16737.08,
      "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": "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": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Azure",
      "accuracy": 0.78041675,
      "accuracy_std_dev": 0.02409111033789228,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.4685391825,
      "avg_duration_per_task_ms": 111064.53375,
      "avg_output_tokens_per_task": 7216.75875,
      "last_run_timestamp": "2026-09-26T08:13:10.617Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Google",
      "accuracy": 0.78000025,
      "accuracy_std_dev": 0.008164965810553043,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.7813667750000001,
      "avg_duration_per_task_ms": 113768.35,
      "avg_output_tokens_per_task": 6852.055,
      "last_run_timestamp": "2026-09-26T08:13:59.150Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Together",
      "accuracy": 0.78,
      "accuracy_std_dev": 0.025071326821120347,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.11856992714285713,
      "avg_duration_per_task_ms": 98828.59428571428,
      "avg_output_tokens_per_task": 15445.625714285714,
      "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": "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": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7788889166666667,
      "accuracy_std_dev": 0.02298676618897032,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.4834922383333334,
      "avg_duration_per_task_ms": 127667.31666666667,
      "avg_output_tokens_per_task": 7153.7075,
      "last_run_timestamp": "2026-09-26T08:12:13.435Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Fireworks",
      "accuracy": 0.7783335,
      "accuracy_std_dev": 0.028037782049406136,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.099593165,
      "avg_duration_per_task_ms": 81195.23,
      "avg_output_tokens_per_task": 4770.2,
      "last_run_timestamp": "2026-09-26T08:13:27.135Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-fable-5.1-20260831",
      "model_slug": "anthropic/claude-fable-5.1",
      "model_name": "Anthropic: Claude Fable 5.1",
      "provider_name": "Azure",
      "accuracy": 0.77833325,
      "accuracy_std_dev": 0.026404708125398767,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.807941035,
      "avg_duration_per_task_ms": 183353.72,
      "avg_output_tokens_per_task": 6938.485,
      "last_run_timestamp": "2026-09-26T08:11:40.896Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Google",
      "accuracy": 0.77813332,
      "accuracy_std_dev": 0.02505255531991896,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.5623465103999999,
      "avg_duration_per_task_ms": 118279.7304,
      "avg_output_tokens_per_task": 7235.6968,
      "last_run_timestamp": "2026-09-26T08:16:10.311Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ambient",
      "accuracy": 0.777778,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.012714583333333333,
      "avg_duration_per_task_ms": 516568.7291666667,
      "avg_output_tokens_per_task": 14230.208333333334,
      "last_run_timestamp": "2026-08-03T17:20:47.207Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "BaseTen",
      "accuracy": 0.7770961820040899,
      "accuracy_std_dev": 0.01908827114120082,
      "run_count": 10,
      "total_questions": 489,
      "avg_cost_per_task_usd": 0.07789318609406953,
      "avg_duration_per_task_ms": 97210.93251533742,
      "avg_output_tokens_per_task": 4258.01226993865,
      "last_run_timestamp": "2026-09-26T08:11:43.280Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7770371111111111,
      "accuracy_std_dev": 0.02224700963397825,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.4357714733333334,
      "avg_duration_per_task_ms": 133682.32888888888,
      "avg_output_tokens_per_task": 7264.266666666666,
      "last_run_timestamp": "2026-09-26T08:17:01.715Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepSeek",
      "accuracy": 0.7766665,
      "accuracy_std_dev": 0.019148513142887433,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.008114253333333333,
      "avg_duration_per_task_ms": 195047.99,
      "avg_output_tokens_per_task": 21425.716666666667,
      "last_run_timestamp": "2026-08-08T08:30:38.880Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Novita",
      "accuracy": 0.776,
      "accuracy_std_dev": 0.022939553421982742,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.007782092000000001,
      "avg_duration_per_task_ms": 211248.992,
      "avg_output_tokens_per_task": 4235.592,
      "last_run_timestamp": "2026-09-26T08:27:53.093Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7755555000000001,
      "accuracy_std_dev": 0.0236225357991756,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.4783927666666667,
      "avg_duration_per_task_ms": 128906.20333333334,
      "avg_output_tokens_per_task": 7843.846666666666,
      "last_run_timestamp": "2026-09-26T08:22:36.885Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "StreamLake",
      "accuracy": 0.77472225,
      "accuracy_std_dev": 0.029058032346567567,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.029999545833333332,
      "avg_duration_per_task_ms": 132057.45166666666,
      "avg_output_tokens_per_task": 5965.595833333334,
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Cloudflare",
      "accuracy": 0.7746666,
      "accuracy_std_dev": 0.02902873978387625,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.15693796000000002,
      "avg_duration_per_task_ms": 605970.636,
      "avg_output_tokens_per_task": 12427.38,
      "last_run_timestamp": "2026-09-26T08:23:03.870Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Z.AI",
      "accuracy": 0.7746666,
      "accuracy_std_dev": 0.022861927596770998,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.080879576,
      "avg_duration_per_task_ms": 126701.14,
      "avg_output_tokens_per_task": 4510.112,
      "last_run_timestamp": "2026-09-26T08:11:43.279Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Alibaba",
      "accuracy": 0.7746666,
      "accuracy_std_dev": 0.020395973461445754,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.120308296,
      "avg_duration_per_task_ms": 80105.144,
      "avg_output_tokens_per_task": 4494.34,
      "last_run_timestamp": "2026-09-26T08:09:25.342Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Cloudflare",
      "accuracy": 0.7746664000000001,
      "accuracy_std_dev": 0.016546172434735443,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.09735419199999999,
      "avg_duration_per_task_ms": 130067.476,
      "avg_output_tokens_per_task": 4561.928,
      "last_run_timestamp": "2026-09-26T08:10:40.765Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Anthropic",
      "accuracy": 0.7733332727272727,
      "accuracy_std_dev": 0.018422746153456893,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.4337744636363636,
      "avg_duration_per_task_ms": 96869.80363636363,
      "avg_output_tokens_per_task": 5661.309090909091,
      "last_run_timestamp": "2026-09-26T08:11:40.823Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "DeepSeek",
      "accuracy": 0.7733331666666667,
      "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": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "Novita",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06535734,
      "avg_duration_per_task_ms": 368327.08,
      "avg_output_tokens_per_task": 13147.48,
      "last_run_timestamp": "2026-09-23T15:00:14.137Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "StreamLake",
      "accuracy": 0.773333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.057549260000000005,
      "avg_duration_per_task_ms": 140710.34,
      "avg_output_tokens_per_task": 4743.94,
      "last_run_timestamp": "2026-09-11T04:09:34.110Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Alibaba",
      "accuracy": 0.7726666,
      "accuracy_std_dev": 0.03641728405359191,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.113975304,
      "avg_duration_per_task_ms": 177276.2,
      "avg_output_tokens_per_task": 9278.056,
      "last_run_timestamp": "2026-08-29T16:15:12.658Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5-20260723",
      "model_slug": "anthropic/claude-opus-5",
      "model_name": "Anthropic: Claude Opus 5",
      "provider_name": "Anthropic",
      "accuracy": 0.7725,
      "accuracy_std_dev": 0.030810865688909155,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.47031593,
      "avg_duration_per_task_ms": 117109.07,
      "avg_output_tokens_per_task": 7210.535,
      "last_run_timestamp": "2026-09-26T08:12:30.655Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7725,
      "accuracy_std_dev": 0.03302967521033168,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.22805288999999998,
      "avg_duration_per_task_ms": 411518.45,
      "avg_output_tokens_per_task": 13622.02,
      "last_run_timestamp": "2026-09-26T08:05:12.789Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Parasail",
      "accuracy": 0.7722223333333333,
      "accuracy_std_dev": 0.03890480468042762,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07840497666666667,
      "avg_duration_per_task_ms": 265427.81666666665,
      "avg_output_tokens_per_task": 11547.233333333334,
      "last_run_timestamp": "2026-09-26T08:13:10.636Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.77166675,
      "accuracy_std_dev": 0.023033966466666143,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.5286243174999999,
      "avg_duration_per_task_ms": 174531.3525,
      "avg_output_tokens_per_task": 10507.29,
      "last_run_timestamp": "2026-09-26T08:05:12.739Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "auto-routing",
      "accuracy": 0.7716665,
      "accuracy_std_dev": 0.033124370015393,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.15296470833333334,
      "avg_duration_per_task_ms": 558733.8216666667,
      "avg_output_tokens_per_task": 36167.525,
      "last_run_timestamp": "2026-09-26T08:01:18.093Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Chutes",
      "accuracy": 0.7716390570902395,
      "accuracy_std_dev": 0.03607105625734658,
      "run_count": 11,
      "total_questions": 543,
      "avg_cost_per_task_usd": 0.07918082872928178,
      "avg_duration_per_task_ms": 988188.6095764273,
      "avg_output_tokens_per_task": 25083.388581952117,
      "last_run_timestamp": "2026-09-26T08:24:45.038Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7715151818181818,
      "accuracy_std_dev": 0.029417231634840356,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.5523802672727273,
      "avg_duration_per_task_ms": 183979.0109090909,
      "avg_output_tokens_per_task": 10526.685454545455,
      "last_run_timestamp": "2026-09-26T08:05:12.761Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7710438282828282,
      "accuracy_std_dev": 0.01667057037866429,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.1615302070707071,
      "avg_duration_per_task_ms": 338222.0707070707,
      "avg_output_tokens_per_task": 20443.252525252527,
      "last_run_timestamp": "2026-09-26T08:27:30.395Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Parasail",
      "accuracy": 0.7706935637583894,
      "accuracy_std_dev": 0.013644816256244217,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.09587820805369127,
      "avg_duration_per_task_ms": 288021.6644295302,
      "avg_output_tokens_per_task": 13114.979865771813,
      "last_run_timestamp": "2026-08-29T08:48:12.957Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "auto-routing",
      "accuracy": 0.7706668,
      "accuracy_std_dev": 0.023702035544653124,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.074997112,
      "avg_duration_per_task_ms": 263018.86,
      "avg_output_tokens_per_task": 12780.056,
      "last_run_timestamp": "2026-09-26T08:10:10.216Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Friendli",
      "accuracy": 0.7706668,
      "accuracy_std_dev": 0.01818435958069462,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.071001692,
      "avg_duration_per_task_ms": 34525.74,
      "avg_output_tokens_per_task": 4762.176,
      "last_run_timestamp": "2026-09-26T08:16:16.877Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Anthropic",
      "accuracy": 0.770303,
      "accuracy_std_dev": 0.018282610525752703,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.4931648818181818,
      "avg_duration_per_task_ms": 129914.10363636364,
      "avg_output_tokens_per_task": 8068.774545454546,
      "last_run_timestamp": "2026-09-26T08:08:19.694Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Azure",
      "accuracy": 0.7702221333333333,
      "accuracy_std_dev": 0.02970127552786618,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.5012201893333333,
      "avg_duration_per_task_ms": 129604.79333333333,
      "avg_output_tokens_per_task": 7921.828,
      "last_run_timestamp": "2026-09-26T08:23:02.726Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.03341663757396906,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.4393913750000001,
      "avg_duration_per_task_ms": 81234.3425,
      "avg_output_tokens_per_task": 5692.71875,
      "last_run_timestamp": "2026-09-26T08:16:17.700Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Parasail",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.007453485392754197,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.081881785,
      "avg_duration_per_task_ms": 99238.51,
      "avg_output_tokens_per_task": 4510.805,
      "last_run_timestamp": "2026-09-26T08:25:16.354Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Together",
      "accuracy": 0.77,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.073975635,
      "avg_duration_per_task_ms": 45899.64,
      "avg_output_tokens_per_task": 4662.84,
      "last_run_timestamp": "2026-09-26T08:21:33.013Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "auto-routing",
      "accuracy": 0.7688888333333334,
      "accuracy_std_dev": 0.014229173651535601,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.10475028333333332,
      "avg_duration_per_task_ms": 333223.94666666666,
      "avg_output_tokens_per_task": 20225.06333333333,
      "last_run_timestamp": "2026-09-26T08:21:56.720Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Io Net",
      "accuracy": 0.7688886666666668,
      "accuracy_std_dev": 0.011331263428624758,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0562409,
      "avg_duration_per_task_ms": 124558.21333333333,
      "avg_output_tokens_per_task": 4508.253333333333,
      "last_run_timestamp": "2026-09-26T08:24:47.398Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-pro-20260709",
      "model_slug": "openai/gpt-5.6-sol-pro",
      "model_name": "OpenAI: GPT-5.6 Sol Pro",
      "provider_name": "Azure",
      "accuracy": 0.7687072040816327,
      "accuracy_std_dev": 0.0156945,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 1.7723489081632653,
      "avg_duration_per_task_ms": 313060.2448979592,
      "avg_output_tokens_per_task": 23077.85714285714,
      "last_run_timestamp": "2026-08-01T08:06:07.759Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Wafer",
      "accuracy": 0.7683335,
      "accuracy_std_dev": 0.010929191289843909,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.09395701500000002,
      "avg_duration_per_task_ms": 284778.19,
      "avg_output_tokens_per_task": 21938.115,
      "last_run_timestamp": "2026-09-26T08:23:03.305Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Azure",
      "accuracy": 0.76833325,
      "accuracy_std_dev": 0.02048026828284971,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.4290680700000001,
      "avg_duration_per_task_ms": 101484.8275,
      "avg_output_tokens_per_task": 5562.66,
      "last_run_timestamp": "2026-09-26T08:11:43.281Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Decart",
      "accuracy": 0.76833325,
      "accuracy_std_dev": 0.02074978581064154,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.062534375,
      "avg_duration_per_task_ms": 51740.385,
      "avg_output_tokens_per_task": 4651.035,
      "last_run_timestamp": "2026-09-26T08:14:03.992Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "SiliconFlow",
      "accuracy": 0.76799996,
      "accuracy_std_dev": 0.03859765703923492,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.0899618072,
      "avg_duration_per_task_ms": 144055.6648,
      "avg_output_tokens_per_task": 5923.9224,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7678431764705882,
      "accuracy_std_dev": 0.037450497276578024,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.13259897294117645,
      "avg_duration_per_task_ms": 176390.52823529413,
      "avg_output_tokens_per_task": 6495.158823529412,
      "last_run_timestamp": "2026-08-29T16:12:45.134Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Morph",
      "accuracy": 0.7671155218855219,
      "accuracy_std_dev": 0.020370521752577246,
      "run_count": 6,
      "total_questions": 297,
      "avg_cost_per_task_usd": 0.04689751515151515,
      "avg_duration_per_task_ms": 143803.91919191918,
      "avg_output_tokens_per_task": 4429.346801346801,
      "last_run_timestamp": "2026-09-26T08:24:40.664Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Alibaba",
      "accuracy": 0.7666666666666667,
      "accuracy_std_dev": 0.020366930350175255,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.1362619733333333,
      "avg_duration_per_task_ms": 318797.34,
      "avg_output_tokens_per_task": 19441.493333333332,
      "last_run_timestamp": "2026-09-26T08:05:12.735Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Makora",
      "accuracy": 0.7666666666666667,
      "accuracy_std_dev": 0.005443174461857933,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12935825333333334,
      "avg_duration_per_task_ms": 58122.44,
      "avg_output_tokens_per_task": 4752.84,
      "last_run_timestamp": "2026-09-19T08:15:57.323Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "xiaomi/mimo-v2.6-flash-20260921",
      "model_slug": "xiaomi/mimo-v2.6-flash",
      "model_name": "Xiaomi: MiMo-V2.6-Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7659570000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 47,
      "avg_cost_per_task_usd": 0.006887468085106384,
      "avg_duration_per_task_ms": 773525.340425532,
      "avg_output_tokens_per_task": 9139.936170212766,
      "last_run_timestamp": "2026-09-22T21:00:04.649Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.765217347826087,
      "accuracy_std_dev": 0.020893702830421756,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.5270942373913046,
      "avg_duration_per_task_ms": 123051.67565217392,
      "avg_output_tokens_per_task": 8091.90347826087,
      "last_run_timestamp": "2026-09-26T08:26:39.210Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "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": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.764166625,
      "accuracy_std_dev": 0.030081035231676705,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.56641688625,
      "avg_duration_per_task_ms": 181084.34625,
      "avg_output_tokens_per_task": 10579.67875,
      "last_run_timestamp": "2026-09-26T08:14:04.078Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7633332500000001,
      "accuracy_std_dev": 0.017743016363840187,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.018957181666666663,
      "avg_duration_per_task_ms": 239454.035,
      "avg_output_tokens_per_task": 17272.2,
      "last_run_timestamp": "2026-09-26T08:11:43.368Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "InferenceNet",
      "accuracy": 0.7633332500000001,
      "accuracy_std_dev": 0.005773358354328977,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.039311875,
      "avg_duration_per_task_ms": 88152.73,
      "avg_output_tokens_per_task": 4482.07,
      "last_run_timestamp": "2026-09-26T08:07:18.978Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Azure",
      "accuracy": 0.7633331666666667,
      "accuracy_std_dev": 0.02631724543853748,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.5436579066666667,
      "avg_duration_per_task_ms": 181008.31666666668,
      "avg_output_tokens_per_task": 7997.5633333333335,
      "last_run_timestamp": "2026-09-01T13:02:58.914Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "StreamLake",
      "accuracy": 0.7626666,
      "accuracy_std_dev": 0.016110821761784848,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.09137509199999999,
      "avg_duration_per_task_ms": 332031.4,
      "avg_output_tokens_per_task": 19933.684,
      "last_run_timestamp": "2026-09-26T08:21:56.175Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7626666,
      "accuracy_std_dev": 0.03227647784749755,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.146569104,
      "avg_duration_per_task_ms": 373419.58,
      "avg_output_tokens_per_task": 24831.152,
      "last_run_timestamp": "2026-08-29T08:26:04.467Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "PrimeIntellect",
      "accuracy": 0.7626262929292928,
      "accuracy_std_dev": 0.02553140404849484,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.10055969696969695,
      "avg_duration_per_task_ms": 35430.38888888889,
      "avg_output_tokens_per_task": 3961.580808080808,
      "last_run_timestamp": "2026-09-26T08:30:51.947Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Novita",
      "accuracy": 0.7622223333333333,
      "accuracy_std_dev": 0.025141417307348132,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.08369306,
      "avg_duration_per_task_ms": 261571.12,
      "avg_output_tokens_per_task": 19541.486666666668,
      "last_run_timestamp": "2026-09-26T08:16:17.045Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Baidu",
      "accuracy": 0.7622223333333333,
      "accuracy_std_dev": 0.013698571612479268,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06797058666666667,
      "avg_duration_per_task_ms": 364277.74,
      "avg_output_tokens_per_task": 20919.446666666667,
      "last_run_timestamp": "2026-09-26T08:05:25.554Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7622223333333333,
      "accuracy_std_dev": 0.006285707880227596,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.1522076066666667,
      "avg_duration_per_task_ms": 298029.9066666667,
      "avg_output_tokens_per_task": 19805.7,
      "last_run_timestamp": "2026-09-26T08:05:12.730Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "AtlasCloud",
      "accuracy": 0.7622223333333333,
      "accuracy_std_dev": 0.013698571612479268,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.09941037999999999,
      "avg_duration_per_task_ms": 137232.76666666666,
      "avg_output_tokens_per_task": 5643.006666666667,
      "last_run_timestamp": "2026-09-26T08:22:37.050Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "GMICloud",
      "accuracy": 0.76222225,
      "accuracy_std_dev": 0.015234723173970049,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.016681795000000003,
      "avg_duration_per_task_ms": 202557.275,
      "avg_output_tokens_per_task": 19520.191666666666,
      "last_run_timestamp": "2026-09-26T08:22:26.866Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "auto-routing",
      "accuracy": 0.7622222083333333,
      "accuracy_std_dev": 0.03818408546434319,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.03669861666666666,
      "avg_duration_per_task_ms": 129387.95,
      "avg_output_tokens_per_task": 6015.355,
      "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": "Sail Research",
      "accuracy": 0.7619999000000001,
      "accuracy_std_dev": 0.014313938818159046,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.008990316,
      "avg_duration_per_task_ms": 356282.446,
      "avg_output_tokens_per_task": 14140.9,
      "last_run_timestamp": "2026-09-26T08:22:54.418Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7619047142857143,
      "accuracy_std_dev": 0.02487158128532752,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03577183142857143,
      "avg_duration_per_task_ms": 462004.73142857145,
      "avg_output_tokens_per_task": 13726.342857142858,
      "last_run_timestamp": "2026-09-26T08:07:18.969Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Inceptron",
      "accuracy": 0.76166675,
      "accuracy_std_dev": 0.028431218171008805,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.051254930000000004,
      "avg_duration_per_task_ms": 163368.405,
      "avg_output_tokens_per_task": 4534.255,
      "last_run_timestamp": "2026-09-26T08:24:09.390Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "auto-routing",
      "accuracy": 0.76166675,
      "accuracy_std_dev": 0.025110926131616508,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.06906778,
      "avg_duration_per_task_ms": 159996.095,
      "avg_output_tokens_per_task": 6808.63,
      "last_run_timestamp": "2026-09-26T08:22:24.923Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "auto-routing",
      "accuracy": 0.761666625,
      "accuracy_std_dev": 0.020749836011023693,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.4161910325,
      "avg_duration_per_task_ms": 98455.6425,
      "avg_output_tokens_per_task": 5629.3925,
      "last_run_timestamp": "2026-09-26T08:10:40.762Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7614156780821918,
      "accuracy_std_dev": 0.0228848139730987,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.024887965753424657,
      "avg_duration_per_task_ms": 458119.4315068493,
      "avg_output_tokens_per_task": 13217.239726027397,
      "last_run_timestamp": "2026-08-29T16:59:36.100Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.7612120909090909,
      "accuracy_std_dev": 0.02825828791202695,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.5180262836363636,
      "avg_duration_per_task_ms": 222126.0781818182,
      "avg_output_tokens_per_task": 10227.196363636363,
      "last_run_timestamp": "2026-09-26T08:22:25.160Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cohere",
      "accuracy": 0.7609424343434344,
      "accuracy_std_dev": 0.012517,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.02651870707070707,
      "avg_duration_per_task_ms": 84229.25252525252,
      "avg_output_tokens_per_task": 17462.434343434343,
      "last_run_timestamp": "2026-09-26T08:10:56.841Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "auto-routing",
      "accuracy": 0.760833625,
      "accuracy_std_dev": 0.018390020316312206,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.4986317424999999,
      "avg_duration_per_task_ms": 132839.435,
      "avg_output_tokens_per_task": 8114.2475,
      "last_run_timestamp": "2026-09-26T08:16:12.134Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Parasail",
      "accuracy": 0.7605556666666666,
      "accuracy_std_dev": 0.013391060222734018,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.12239992333333333,
      "avg_duration_per_task_ms": 458072.06,
      "avg_output_tokens_per_task": 22242.516666666666,
      "last_run_timestamp": "2026-09-26T08:22:24.688Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "BaseTen",
      "accuracy": 0.7604167291666667,
      "accuracy_std_dev": 0.017826499999999967,
      "run_count": 2,
      "total_questions": 96,
      "avg_cost_per_task_usd": 0.2660930625,
      "avg_duration_per_task_ms": 146365.54166666666,
      "avg_output_tokens_per_task": 6163.291666666667,
      "last_run_timestamp": "2026-09-26T08:05:12.774Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Sail Research",
      "accuracy": 0.7603703333333334,
      "accuracy_std_dev": 0.028822460188771784,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.056350817777777784,
      "avg_duration_per_task_ms": 98825.95333333334,
      "avg_output_tokens_per_task": 4242.228888888889,
      "last_run_timestamp": "2026-09-26T08:32:07.081Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Alibaba",
      "accuracy": 0.7602777916666666,
      "accuracy_std_dev": 0.03482648069713999,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.11371839333333329,
      "avg_duration_per_task_ms": 294570.1925,
      "avg_output_tokens_per_task": 17313.15166666667,
      "last_run_timestamp": "2026-08-29T16:25:47.585Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-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/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Chutes",
      "accuracy": 0.7600423065539111,
      "accuracy_std_dev": 0.030704280834836183,
      "run_count": 19,
      "total_questions": 946,
      "avg_cost_per_task_usd": 0.08771633932346722,
      "avg_duration_per_task_ms": 364680.0031712474,
      "avg_output_tokens_per_task": 16580.900634249472,
      "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-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01747962,
      "avg_duration_per_task_ms": 589736.28,
      "avg_output_tokens_per_task": 13675.16,
      "last_run_timestamp": "2026-08-01T08:06:01.204Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "DeepInfra",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.034318637633216124,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.03080999,
      "avg_duration_per_task_ms": 262760.8,
      "avg_output_tokens_per_task": 18663.095,
      "last_run_timestamp": "2026-09-26T08:09:25.342Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3-flash-preview-20251217",
      "model_slug": "google/gemini-3-flash-preview",
      "model_name": "Google: Gemini 3 Flash Preview",
      "provider_name": "Google",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.033333,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.11355728,
      "avg_duration_per_task_ms": 147828.6,
      "avg_output_tokens_per_task": 17720.73,
      "last_run_timestamp": "2026-09-01T17:30:58.271Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "Venice",
      "accuracy": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.19103682,
      "avg_duration_per_task_ms": 233230.24,
      "avg_output_tokens_per_task": 15236.66,
      "last_run_timestamp": "2026-08-04T03:12:00.236Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "AkashML",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.020727419665747104,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.0731778,
      "avg_duration_per_task_ms": 258357.41,
      "avg_output_tokens_per_task": 11852.263333333334,
      "last_run_timestamp": "2026-09-26T08:07:46.003Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Chutes",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.038682027601113875,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07101471333333334,
      "avg_duration_per_task_ms": 318099.35333333333,
      "avg_output_tokens_per_task": 12493.926666666666,
      "last_run_timestamp": "2026-09-26T08:05:12.731Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "x-ai/grok-4.6-20260810",
      "model_slug": "x-ai/grok-4.6",
      "model_name": "SpaceXAI: Grok 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.76,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.35068044,
      "avg_duration_per_task_ms": 248200.34,
      "avg_output_tokens_per_task": 11571.02,
      "last_run_timestamp": "2026-09-01T12:41:31.788Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Makora",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07038507,
      "avg_duration_per_task_ms": 256302.49,
      "avg_output_tokens_per_task": 10045.27,
      "last_run_timestamp": "2026-08-26T00:00:05.638Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Relace",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.027619090000000002,
      "avg_duration_per_task_ms": 42870.03,
      "avg_output_tokens_per_task": 4628.08,
      "last_run_timestamp": "2026-09-26T09:28:27.720Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Baidu",
      "accuracy": 0.76,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04223252,
      "avg_duration_per_task_ms": 69942.84,
      "avg_output_tokens_per_task": 4621.51,
      "last_run_timestamp": "2026-09-26T08:05:12.737Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.027371610829992666,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.009037317142857142,
      "avg_duration_per_task_ms": 270209.39714285714,
      "avg_output_tokens_per_task": 6316.0257142857145,
      "last_run_timestamp": "2026-09-26T08:08:19.666Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.76,
      "accuracy_std_dev": 0.026666999999999996,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.014824000000000002,
      "avg_duration_per_task_ms": 335920.77,
      "avg_output_tokens_per_task": 4448.51,
      "last_run_timestamp": "2026-09-12T08:11:26.740Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7599999583333333,
      "accuracy_std_dev": 0.04023080402655736,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.05313657,
      "avg_duration_per_task_ms": 148959.73916666667,
      "avg_output_tokens_per_task": 5980.9625,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Novita",
      "accuracy": 0.7594871538461538,
      "accuracy_std_dev": 0.03003836972132133,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.05324420692307693,
      "avg_duration_per_task_ms": 141495.42,
      "avg_output_tokens_per_task": 5929.143846153846,
      "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7592915452961673,
      "accuracy_std_dev": 0.05377860847392117,
      "run_count": 23,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.15847917944250875,
      "avg_duration_per_task_ms": 312268.7691637631,
      "avg_output_tokens_per_task": 17697.43205574913,
      "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-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "StreamLake",
      "accuracy": 0.759166625,
      "accuracy_std_dev": 0.02319920826201565,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0076381949999999995,
      "avg_duration_per_task_ms": 261034.725,
      "avg_output_tokens_per_task": 15531.6525,
      "last_run_timestamp": "2026-09-26T08:24:41.493Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Google",
      "accuracy": 0.758888875,
      "accuracy_std_dev": 0.0209643399476995,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.6800024841666666,
      "avg_duration_per_task_ms": 74777.63666666667,
      "avg_output_tokens_per_task": 5635.594166666667,
      "last_run_timestamp": "2026-09-26T08:23:22.613Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.7-opus-20260416",
      "model_slug": "anthropic/claude-opus-4.7",
      "model_name": "Anthropic: Claude Opus 4.7",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7588888333333333,
      "accuracy_std_dev": 0.03578052612067383,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.4306518199999999,
      "avg_duration_per_task_ms": 94862.28333333334,
      "avg_output_tokens_per_task": 5554.29,
      "last_run_timestamp": "2026-09-26T08:25:11.484Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.8-opus-20260528",
      "model_slug": "anthropic/claude-opus-4.8",
      "model_name": "Anthropic: Claude Opus 4.8",
      "provider_name": "Google",
      "accuracy": 0.7588406086956522,
      "accuracy_std_dev": 0.021414040391536566,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.59698552,
      "avg_duration_per_task_ms": 123337.09304347826,
      "avg_output_tokens_per_task": 8161.509565217391,
      "last_run_timestamp": "2026-09-26T08:22:57.486Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "Google",
      "accuracy": 0.7586668,
      "accuracy_std_dev": 0.01654634971708262,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.27091952399999997,
      "avg_duration_per_task_ms": 154225.472,
      "avg_output_tokens_per_task": 23085.08,
      "last_run_timestamp": "2026-09-01T17:32:40.487Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Reka",
      "accuracy": 0.7586668,
      "accuracy_std_dev": 0.012927091279943853,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.08354813200000001,
      "avg_duration_per_task_ms": 248803.664,
      "avg_output_tokens_per_task": 12064.656,
      "last_run_timestamp": "2026-09-26T08:08:19.674Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Mistral",
      "accuracy": 0.7585185555555556,
      "accuracy_std_dev": 0.01743499783711622,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.07257207555555555,
      "avg_duration_per_task_ms": 58870.41333333333,
      "avg_output_tokens_per_task": 4565.811111111111,
      "last_run_timestamp": "2026-09-26T08:26:14.308Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Azure",
      "accuracy": 0.7583332500000001,
      "accuracy_std_dev": 0.015184156584660873,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.163021,
      "avg_duration_per_task_ms": 227043.59,
      "avg_output_tokens_per_task": 11333.075,
      "last_run_timestamp": "2026-08-29T16:55:33.128Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.7583332500000001,
      "accuracy_std_dev": 0.022298623526744884,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.025212460000000003,
      "avg_duration_per_task_ms": 129050.295,
      "avg_output_tokens_per_task": 16100.67,
      "last_run_timestamp": "2026-09-26T08:22:17.628Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Crusoe",
      "accuracy": 0.7583331666666667,
      "accuracy_std_dev": 0.013844473427047428,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.09591996333333334,
      "avg_duration_per_task_ms": 214673.9,
      "avg_output_tokens_per_task": 9689.54,
      "last_run_timestamp": "2026-08-29T16:31:41.554Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260423",
      "model_slug": "deepseek/deepseek-v4-pro",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0423",
      "provider_name": "Ionstream",
      "accuracy": 0.758,
      "accuracy_std_dev": 0.03709447398198281,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.079319272,
      "avg_duration_per_task_ms": 620750.224,
      "avg_output_tokens_per_task": 10895.356,
      "last_run_timestamp": "2026-08-29T17:19:42.622Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7579964141414142,
      "accuracy_std_dev": 0.023533015514113357,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.20713992929292932,
      "avg_duration_per_task_ms": 664979.7727272727,
      "avg_output_tokens_per_task": 53099.67929292929,
      "last_run_timestamp": "2026-09-26T08:05:12.736Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Azure",
      "accuracy": 0.7575,
      "accuracy_std_dev": 0.023318483371565994,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.34416930749999997,
      "avg_duration_per_task_ms": 221562.1725,
      "avg_output_tokens_per_task": 11914.215,
      "last_run_timestamp": "2026-09-26T08:22:18.965Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Phala",
      "accuracy": 0.7573335999999999,
      "accuracy_std_dev": 0.017178865569064807,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.151421216,
      "avg_duration_per_task_ms": 315167.644,
      "avg_output_tokens_per_task": 20253,
      "last_run_timestamp": "2026-09-26T08:10:25.658Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Modal",
      "accuracy": 0.7573333999999999,
      "accuracy_std_dev": 0.031439640936880967,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.081021904,
      "avg_duration_per_task_ms": 42035.076,
      "avg_output_tokens_per_task": 4841.128,
      "last_run_timestamp": "2026-09-26T08:26:16.831Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Makora",
      "accuracy": 0.756944,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.2442584375,
      "avg_duration_per_task_ms": 645873.5208333334,
      "avg_output_tokens_per_task": 12350.729166666666,
      "last_run_timestamp": "2026-09-26T08:23:04.076Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "Google",
      "accuracy": 0.7568027346938776,
      "accuracy_std_dev": 0.038809665680729605,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.3020268469387756,
      "avg_duration_per_task_ms": 145185.1581632653,
      "avg_output_tokens_per_task": 20600.117346938776,
      "last_run_timestamp": "2026-09-01T17:31:47.860Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Parasail",
      "accuracy": 0.7567114127516779,
      "accuracy_std_dev": 0.02249334300085622,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.020425387583892618,
      "avg_duration_per_task_ms": 519535.90939597314,
      "avg_output_tokens_per_task": 17284.703020134228,
      "last_run_timestamp": "2026-09-26T08:07:18.967Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7566669999999999,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01828589,
      "avg_duration_per_task_ms": 240722.54,
      "avg_output_tokens_per_task": 15147.8,
      "last_run_timestamp": "2026-09-19T08:02:39.158Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.6-pro-20260921",
      "model_slug": "xiaomi/mimo-v2.6-pro",
      "model_name": "Xiaomi: MiMo-V2.6-Pro",
      "provider_name": "GMICloud",
      "accuracy": 0.7566669999999999,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.05416404,
      "avg_duration_per_task_ms": 1008004.72,
      "avg_output_tokens_per_task": 38151.6,
      "last_run_timestamp": "2026-09-26T00:00:04.127Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "InferenceNet",
      "accuracy": 0.7566667499999999,
      "accuracy_std_dev": 0.013743584364622665,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.33711718500000004,
      "avg_duration_per_task_ms": 415294.135,
      "avg_output_tokens_per_task": 10214.705,
      "last_run_timestamp": "2026-09-26T08:13:27.132Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7566665,
      "accuracy_std_dev": 0.02687425451214104,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.54838445,
      "avg_duration_per_task_ms": 178864.39333333334,
      "avg_output_tokens_per_task": 10795.433333333332,
      "last_run_timestamp": "2026-09-26T08:26:47.832Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Novita",
      "accuracy": 0.756,
      "accuracy_std_dev": 0.012365076271499492,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.0694329,
      "avg_duration_per_task_ms": 122849.852,
      "avg_output_tokens_per_task": 4849.072,
      "last_run_timestamp": "2026-09-26T08:26:14.187Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Z.AI",
      "accuracy": 0.755919868852459,
      "accuracy_std_dev": 0.04303119426679743,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.054472353369763214,
      "avg_duration_per_task_ms": 164605.7941712204,
      "avg_output_tokens_per_task": 6348.341530054645,
      "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": "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": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7555556666666666,
      "accuracy_std_dev": 0.012570472951413688,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05382074666666667,
      "avg_duration_per_task_ms": 90716.92666666667,
      "avg_output_tokens_per_task": 14804.206666666667,
      "last_run_timestamp": "2026-09-26T08:22:43.308Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Alibaba",
      "accuracy": 0.7555556666666666,
      "accuracy_std_dev": 0.03823241466202323,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0327312,
      "avg_duration_per_task_ms": 153585.12,
      "avg_output_tokens_per_task": 15537.24,
      "last_run_timestamp": "2026-09-26T08:17:02.789Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Wafer",
      "accuracy": 0.7555556666666666,
      "accuracy_std_dev": 0.011331263428624758,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05866996,
      "avg_duration_per_task_ms": 356562.5266666667,
      "avg_output_tokens_per_task": 13311.113333333333,
      "last_run_timestamp": "2026-09-26T08:12:56.927Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DekaLLM",
      "accuracy": 0.7555556666666666,
      "accuracy_std_dev": 0.012570472951413688,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.06505471333333333,
      "avg_duration_per_task_ms": 157291.16,
      "avg_output_tokens_per_task": 14450.066666666668,
      "last_run_timestamp": "2026-09-26T08:05:12.747Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7553623043478261,
      "accuracy_std_dev": 0.032979826821380405,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.19905160782608697,
      "avg_duration_per_task_ms": 122704.35391304348,
      "avg_output_tokens_per_task": 7790.118260869565,
      "last_run_timestamp": "2026-09-26T08:11:40.828Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Phala",
      "accuracy": 0.7552381428571429,
      "accuracy_std_dev": 0.009897295726302106,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.03504835714285714,
      "avg_duration_per_task_ms": 271800.3542857143,
      "avg_output_tokens_per_task": 13667.517142857143,
      "last_run_timestamp": "2026-09-26T08:10:40.746Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Novita",
      "accuracy": 0.7551515454545454,
      "accuracy_std_dev": 0.023968159933786534,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.024316309090909087,
      "avg_duration_per_task_ms": 165689.23272727273,
      "avg_output_tokens_per_task": 18155.507272727275,
      "last_run_timestamp": "2026-09-26T08:08:49.576Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.755,
      "accuracy_std_dev": 0.04773577934491767,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.04855596,
      "avg_duration_per_task_ms": 135012.90833333333,
      "avg_output_tokens_per_task": 6055.553333333333,
      "last_run_timestamp": "2026-08-29T16:58:32.929Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Phala",
      "accuracy": 0.755,
      "accuracy_std_dev": 0.017240198505237704,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.055826839999999996,
      "avg_duration_per_task_ms": 122326.96,
      "avg_output_tokens_per_task": 4308.39,
      "last_run_timestamp": "2026-09-26T08:26:39.685Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7546664000000001,
      "accuracy_std_dev": 0.04509740371950474,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.17806932799999997,
      "avg_duration_per_task_ms": 249010.144,
      "avg_output_tokens_per_task": 9041.408,
      "last_run_timestamp": "2026-09-26T08:26:47.380Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7544445,
      "accuracy_std_dev": 0.025939233358948773,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.19473402666666667,
      "avg_duration_per_task_ms": 127608.55,
      "avg_output_tokens_per_task": 7689.33,
      "last_run_timestamp": "2026-09-26T08:22:38.150Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7543103448275862,
      "accuracy_std_dev": 0.016399321098441732,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.04285599137931035,
      "avg_duration_per_task_ms": 524163.9051724138,
      "avg_output_tokens_per_task": 22101.97988505747,
      "last_run_timestamp": "2026-09-26T16:00:25.681Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Ionstream",
      "accuracy": 0.7542953608247424,
      "accuracy_std_dev": 0.02652500000000002,
      "run_count": 2,
      "total_questions": 97,
      "avg_cost_per_task_usd": 0.021299783505154636,
      "avg_duration_per_task_ms": 633495.5257731959,
      "avg_output_tokens_per_task": 26267.113402061856,
      "last_run_timestamp": "2026-08-08T08:30:53.738Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "auto-routing",
      "accuracy": 0.754166625,
      "accuracy_std_dev": 0.02895150279319496,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.1985393725,
      "avg_duration_per_task_ms": 132141.125,
      "avg_output_tokens_per_task": 7833.895,
      "last_run_timestamp": "2026-09-26T08:08:19.638Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Together",
      "accuracy": 0.754166625,
      "accuracy_std_dev": 0.0239066485989646,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.02336587,
      "avg_duration_per_task_ms": 124888.855,
      "avg_output_tokens_per_task": 7646.1525,
      "last_run_timestamp": "2026-09-26T16:00:25.687Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Cloudflare",
      "accuracy": 0.7539393636363636,
      "accuracy_std_dev": 0.01999089692049736,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.024043370909090908,
      "avg_duration_per_task_ms": 299110.4909090909,
      "avg_output_tokens_per_task": 17382.84909090909,
      "last_run_timestamp": "2026-09-26T08:12:56.930Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "Novita",
      "accuracy": 0.7538888333333333,
      "accuracy_std_dev": 0.033934516643169634,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.1837781666666667,
      "avg_duration_per_task_ms": 552485.2983333333,
      "avg_output_tokens_per_task": 26570.453333333335,
      "last_run_timestamp": "2026-09-26T08:13:27.139Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Cloudflare",
      "accuracy": 0.7538536705202312,
      "accuracy_std_dev": 0.022320967608497434,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.20567977745664742,
      "avg_duration_per_task_ms": 493851.28612716764,
      "avg_output_tokens_per_task": 20258.49710982659,
      "last_run_timestamp": "2026-09-26T08:08:00.526Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Novita",
      "accuracy": 0.7533668080808081,
      "accuracy_std_dev": 0.00967513458239728,
      "run_count": 4,
      "total_questions": 198,
      "avg_cost_per_task_usd": 0.0209760505050505,
      "avg_duration_per_task_ms": 99025.76767676767,
      "avg_output_tokens_per_task": 14135.353535353535,
      "last_run_timestamp": "2026-09-26T08:05:09.043Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7533335,
      "accuracy_std_dev": 0.026666500000000037,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.5044717400000001,
      "avg_duration_per_task_ms": 153665.59,
      "avg_output_tokens_per_task": 8278.2,
      "last_run_timestamp": "2026-09-01T12:41:31.807Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "auto-routing",
      "accuracy": 0.7533335,
      "accuracy_std_dev": 0.014142371326973414,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.19823152,
      "avg_duration_per_task_ms": 359205.05,
      "avg_output_tokens_per_task": 12583.985,
      "last_run_timestamp": "2026-09-26T08:10:43.550Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Reka",
      "accuracy": 0.7533333999999999,
      "accuracy_std_dev": 0.01738441192102859,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.019357872,
      "avg_duration_per_task_ms": 107397.768,
      "avg_output_tokens_per_task": 16129.496,
      "last_run_timestamp": "2026-09-26T08:24:48.615Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "NextBit",
      "accuracy": 0.7533333999999999,
      "accuracy_std_dev": 0.0193218356643462,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.037210308,
      "avg_duration_per_task_ms": 202748.952,
      "avg_output_tokens_per_task": 17567.136,
      "last_run_timestamp": "2026-09-26T08:23:54.132Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Morph",
      "accuracy": 0.7533333333333333,
      "accuracy_std_dev": 0.014401903168987377,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.021082753333333332,
      "avg_duration_per_task_ms": 915986.2133333334,
      "avg_output_tokens_per_task": 13268.026666666667,
      "last_run_timestamp": "2026-08-15T08:15:52.453Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-pro-preview-20260219",
      "model_slug": "google/gemini-3.1-pro-preview",
      "model_name": "Google: Gemini 3.1 Pro Preview",
      "provider_name": "auto-routing",
      "accuracy": 0.753333,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.36418883999999996,
      "avg_duration_per_task_ms": 138377.24,
      "avg_output_tokens_per_task": 12769.81,
      "last_run_timestamp": "2026-09-01T17:31:47.883Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.1-20251113",
      "model_slug": "openai/gpt-5.1",
      "model_name": "OpenAI: GPT-5.1",
      "provider_name": "OpenAI",
      "accuracy": 0.753333,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.41880896,
      "avg_duration_per_task_ms": 447303.41,
      "avg_output_tokens_per_task": 36551.01,
      "last_run_timestamp": "2026-09-01T12:59:50.177Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.6857741799999999,
      "avg_duration_per_task_ms": 105619.36,
      "avg_output_tokens_per_task": 7597.08,
      "last_run_timestamp": "2026-09-01T13:02:16.941Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0196027,
      "avg_duration_per_task_ms": 76385.02,
      "avg_output_tokens_per_task": 9950,
      "last_run_timestamp": "2026-07-18T08:04:02.230Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "CoreWeave",
      "accuracy": 0.753333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.10139794,
      "avg_duration_per_task_ms": 119629.34,
      "avg_output_tokens_per_task": 9506.02,
      "last_run_timestamp": "2026-08-01T17:46:04.356Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Wafer",
      "accuracy": 0.7525925555555556,
      "accuracy_std_dev": 0.02463451135257706,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021416855555555555,
      "avg_duration_per_task_ms": 123014.89111111111,
      "avg_output_tokens_per_task": 15302.471111111112,
      "last_run_timestamp": "2026-09-26T08:21:23.410Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Anthropic",
      "accuracy": 0.752499875,
      "accuracy_std_dev": 0.04128993437400179,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.19582053000000002,
      "avg_duration_per_task_ms": 106281.8925,
      "avg_output_tokens_per_task": 7612.88,
      "last_run_timestamp": "2026-09-26T08:23:00.199Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Azure",
      "accuracy": 0.7524445333333333,
      "accuracy_std_dev": 0.03730686310384309,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.19597836800000001,
      "avg_duration_per_task_ms": 104969.73733333334,
      "avg_output_tokens_per_task": 7646.229333333334,
      "last_run_timestamp": "2026-09-26T08:12:13.439Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Alibaba",
      "accuracy": 0.752,
      "accuracy_std_dev": 0.015434478002187193,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.036405735999999994,
      "avg_duration_per_task_ms": 202702.796,
      "avg_output_tokens_per_task": 18338.232,
      "last_run_timestamp": "2026-09-26T08:01:18.110Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7516667499999999,
      "accuracy_std_dev": 0.009860203404975986,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07850260499999999,
      "avg_duration_per_task_ms": 143951.215,
      "avg_output_tokens_per_task": 4587.75,
      "last_run_timestamp": "2026-09-26T08:22:24.392Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Venice",
      "accuracy": 0.7516667499999999,
      "accuracy_std_dev": 0.0172401743375031,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.14700563,
      "avg_duration_per_task_ms": 224982.975,
      "avg_output_tokens_per_task": 8550.835,
      "last_run_timestamp": "2026-09-26T08:08:19.643Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Google",
      "accuracy": 0.751666625,
      "accuracy_std_dev": 0.02598077013358873,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.54754274,
      "avg_duration_per_task_ms": 191361.1975,
      "avg_output_tokens_per_task": 10587.185,
      "last_run_timestamp": "2026-09-26T08:16:15.812Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "Azure",
      "accuracy": 0.7516665,
      "accuracy_std_dev": 0.03378863221188454,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.025036024999999996,
      "avg_duration_per_task_ms": 332301.835,
      "avg_output_tokens_per_task": 13337.575,
      "last_run_timestamp": "2026-08-29T17:19:05.764Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Modal",
      "accuracy": 0.7516665,
      "accuracy_std_dev": 0.03484088434942489,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.017483123333333336,
      "avg_duration_per_task_ms": 78588.8,
      "avg_output_tokens_per_task": 4208.683333333333,
      "last_run_timestamp": "2026-09-26T08:25:00.207Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AkashML",
      "accuracy": 0.7511110833333333,
      "accuracy_std_dev": 0.016629606947942455,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.051246248333333334,
      "avg_duration_per_task_ms": 276941.145,
      "avg_output_tokens_per_task": 29769.108333333334,
      "last_run_timestamp": "2026-09-26T16:00:20.167Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Novita",
      "accuracy": 0.7506668,
      "accuracy_std_dev": 0.027840878020637227,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.08836137599999999,
      "avg_duration_per_task_ms": 290555.676,
      "avg_output_tokens_per_task": 12011.224,
      "last_run_timestamp": "2026-09-26T08:10:43.592Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Z.AI",
      "accuracy": 0.7506666,
      "accuracy_std_dev": 0.029090015088342618,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.008165608,
      "avg_duration_per_task_ms": 191715.892,
      "avg_output_tokens_per_task": 4404.68,
      "last_run_timestamp": "2026-09-26T08:23:40.133Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "Alibaba",
      "accuracy": 0.7502796040268457,
      "accuracy_std_dev": 0.04400657815221114,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.11341485906040269,
      "avg_duration_per_task_ms": 467292.23825503356,
      "avg_output_tokens_per_task": 23614.795302013423,
      "last_run_timestamp": "2026-09-26T08:23:03.773Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Io Net",
      "accuracy": 0.7500002567567566,
      "accuracy_std_dev": 0.007287030686241284,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.009324743243243245,
      "avg_duration_per_task_ms": 215748.13513513515,
      "avg_output_tokens_per_task": 3777.3783783783783,
      "last_run_timestamp": "2026-09-26T08:06:43.931Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "amazon/nova-micro-v1",
      "model_slug": "amazon/nova-micro-v1",
      "model_name": "Amazon: Nova Micro 1.0",
      "provider_name": "Anthropic",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 1.2253275,
      "avg_duration_per_task_ms": 98359.26,
      "avg_output_tokens_per_task": 5434.86,
      "last_run_timestamp": "2026-09-01T13:03:25.552Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-opus-20251124",
      "model_slug": "anthropic/claude-opus-4.5",
      "model_name": "Anthropic: Claude Opus 4.5",
      "provider_name": "Azure",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.027688836500474354,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.5338496725,
      "avg_duration_per_task_ms": 189310.245,
      "avg_output_tokens_per_task": 10232.6975,
      "last_run_timestamp": "2026-09-26T08:09:25.333Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "BaseTen",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.28014248,
      "avg_duration_per_task_ms": 223020.56,
      "avg_output_tokens_per_task": 18715.38,
      "last_run_timestamp": "2026-09-12T08:12:13.731Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Z.AI",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.02839314819146025,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.035507067777777776,
      "avg_duration_per_task_ms": 259194.36444444445,
      "avg_output_tokens_per_task": 5673.533333333334,
      "last_run_timestamp": "2026-08-29T16:10:33.029Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.75,
      "accuracy_std_dev": 0.020275820303997587,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.0059813,
      "avg_duration_per_task_ms": 146386.56,
      "avg_output_tokens_per_task": 4218.225,
      "last_run_timestamp": "2026-09-26T08:22:40.372Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-sonnet-5-20260630",
      "model_slug": "anthropic/claude-sonnet-5",
      "model_name": "Anthropic: Claude Sonnet 5",
      "provider_name": "Google",
      "accuracy": 0.7499999583333333,
      "accuracy_std_dev": 0.0368179040090995,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.2266885883333333,
      "avg_duration_per_task_ms": 100830.78416666666,
      "avg_output_tokens_per_task": 7512.6875,
      "last_run_timestamp": "2026-09-26T08:24:10.457Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "CoreWeave",
      "accuracy": 0.7496296666666666,
      "accuracy_std_dev": 0.01780858133971749,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021206060000000002,
      "avg_duration_per_task_ms": 122947.83777777778,
      "avg_output_tokens_per_task": 16271.093333333334,
      "last_run_timestamp": "2026-09-26T08:08:19.683Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Alibaba",
      "accuracy": 0.7493332,
      "accuracy_std_dev": 0.01162353169049754,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.172436168,
      "avg_duration_per_task_ms": 308861.672,
      "avg_output_tokens_per_task": 10652.584,
      "last_run_timestamp": "2026-09-26T08:22:58.738Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "DeepInfra",
      "accuracy": 0.748889,
      "accuracy_std_dev": 0.01257094435593444,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.05290565333333333,
      "avg_duration_per_task_ms": 112293.75333333333,
      "avg_output_tokens_per_task": 4801.486666666667,
      "last_run_timestamp": "2026-09-26T08:17:02.746Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.748889,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.01121336,
      "avg_duration_per_task_ms": 201242.51333333334,
      "avg_output_tokens_per_task": 4454.833333333333,
      "last_run_timestamp": "2026-09-26T08:24:26.075Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "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": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.748333375,
      "accuracy_std_dev": 0.027233626580651628,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.488940975,
      "avg_duration_per_task_ms": 194699.49,
      "avg_output_tokens_per_task": 10000.255,
      "last_run_timestamp": "2026-09-26T08:11:43.276Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Modal",
      "accuracy": 0.7483332500000001,
      "accuracy_std_dev": 0.012802094015726486,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.17867437000000003,
      "avg_duration_per_task_ms": 156388.555,
      "avg_output_tokens_per_task": 12772.235,
      "last_run_timestamp": "2026-09-26T08:22:52.562Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Together",
      "accuracy": 0.7481202957393484,
      "accuracy_std_dev": 0.01771080940943976,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.01360957142857143,
      "avg_duration_per_task_ms": 218299.9649122807,
      "avg_output_tokens_per_task": 16333.967418546366,
      "last_run_timestamp": "2026-09-26T08:20:57.091Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.748116,
      "accuracy_std_dev": 0.022837863998193896,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.37004660695652175,
      "avg_duration_per_task_ms": 167935.71652173914,
      "avg_output_tokens_per_task": 11884.72,
      "last_run_timestamp": "2026-09-26T08:22:24.476Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "CoreWeave",
      "accuracy": 0.7480002,
      "accuracy_std_dev": 0.014236085605249795,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.14549154800000003,
      "avg_duration_per_task_ms": 139795.724,
      "avg_output_tokens_per_task": 19960.988,
      "last_run_timestamp": "2026-09-26T08:09:25.342Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Cloudflare",
      "accuracy": 0.748,
      "accuracy_std_dev": 0.015434218839967281,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.011429584,
      "avg_duration_per_task_ms": 130555.472,
      "avg_output_tokens_per_task": 4185.616,
      "last_run_timestamp": "2026-09-26T08:24:45.322Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Venice",
      "accuracy": 0.7479998000000001,
      "accuracy_std_dev": 0.023247453386554,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.025147304000000002,
      "avg_duration_per_task_ms": 339787.244,
      "avg_output_tokens_per_task": 4397.436,
      "last_run_timestamp": "2026-09-26T08:13:28.807Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Google",
      "accuracy": 0.747916625,
      "accuracy_std_dev": 0.029459786032986997,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.56374661625,
      "avg_duration_per_task_ms": 203094.54,
      "avg_output_tokens_per_task": 9850.0775,
      "last_run_timestamp": "2026-09-26T08:23:07.274Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Anthropic",
      "accuracy": 0.7478787272727273,
      "accuracy_std_dev": 0.029242020409898036,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.33650350909090904,
      "avg_duration_per_task_ms": 209916.4818181818,
      "avg_output_tokens_per_task": 11777.303636363637,
      "last_run_timestamp": "2026-09-26T08:09:55.575Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DeepInfra",
      "accuracy": 0.7477703913043479,
      "accuracy_std_dev": 0.01921375665647317,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.0067899197324414705,
      "avg_duration_per_task_ms": 182508.58026755851,
      "avg_output_tokens_per_task": 12210.97491638796,
      "last_run_timestamp": "2026-09-26T08:06:10.867Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Venice",
      "accuracy": 0.747500125,
      "accuracy_std_dev": 0.03323948977661021,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0216457225,
      "avg_duration_per_task_ms": 167096.8075,
      "avg_output_tokens_per_task": 13839.2125,
      "last_run_timestamp": "2026-09-26T08:10:10.199Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Baidu",
      "accuracy": 0.7475,
      "accuracy_std_dev": 0.028565091361660317,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.012148102500000002,
      "avg_duration_per_task_ms": 152199.3025,
      "avg_output_tokens_per_task": 16425.225,
      "last_run_timestamp": "2026-09-26T08:04:53.554Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Friendli",
      "accuracy": 0.7471111333333333,
      "accuracy_std_dev": 0.1405273225224033,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.04440916933333333,
      "avg_duration_per_task_ms": 42335.25866666667,
      "avg_output_tokens_per_task": 3598.3533333333335,
      "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/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.7469239563758389,
      "accuracy_std_dev": 0.04294748693624446,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.1647119630872483,
      "avg_duration_per_task_ms": 412023.6426174497,
      "avg_output_tokens_per_task": 24209.687919463086,
      "last_run_timestamp": "2026-09-26T08:12:00.291Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Wafer",
      "accuracy": 0.7466668333333333,
      "accuracy_std_dev": 0.02143021404790186,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.24169485,
      "avg_duration_per_task_ms": 277144.9666666667,
      "avg_output_tokens_per_task": 10442,
      "last_run_timestamp": "2026-09-26T08:09:25.340Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7466667499999999,
      "accuracy_std_dev": 0.03349941956493424,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.54667351,
      "avg_duration_per_task_ms": 183079.6925,
      "avg_output_tokens_per_task": 10525.565,
      "last_run_timestamp": "2026-09-26T08:07:18.989Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "NextBit",
      "accuracy": 0.7466667499999999,
      "accuracy_std_dev": 0.014907082584714572,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.12188539999999999,
      "avg_duration_per_task_ms": 449678.28,
      "avg_output_tokens_per_task": 19415.325,
      "last_run_timestamp": "2026-09-26T08:23:24.023Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.7466667499999999,
      "accuracy_std_dev": 0.02494415986132826,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.0097815,
      "avg_duration_per_task_ms": 206039.625,
      "avg_output_tokens_per_task": 4442.965,
      "last_run_timestamp": "2026-09-26T08:26:29.786Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Venice",
      "accuracy": 0.7466666666666667,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.2640489466666667,
      "avg_duration_per_task_ms": 261052.74666666667,
      "avg_output_tokens_per_task": 16845.666666666668,
      "last_run_timestamp": "2026-09-26T08:22:21.004Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-20260519",
      "model_slug": "google/gemini-3.5-flash",
      "model_name": "Google: Gemini 3.5 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.42183384,
      "avg_duration_per_task_ms": 170456.6,
      "avg_output_tokens_per_task": 28298.04,
      "last_run_timestamp": "2026-09-01T17:30:52.780Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Fireworks",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.011547101609639251,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04422623666666667,
      "avg_duration_per_task_ms": 266653.04,
      "avg_output_tokens_per_task": 13360.296666666667,
      "last_run_timestamp": "2026-09-19T08:01:36.427Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "GMICloud",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.024366269999999995,
      "avg_duration_per_task_ms": 118191.69,
      "avg_output_tokens_per_task": 9023.81,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Reka",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04198429,
      "avg_duration_per_task_ms": 312370.3,
      "avg_output_tokens_per_task": 24895.75,
      "last_run_timestamp": "2026-09-26T16:00:24.612Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.03717445,
      "avg_duration_per_task_ms": 350241.71,
      "avg_output_tokens_per_task": 12501.07,
      "last_run_timestamp": "2026-09-26T08:10:43.553Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Sail Research",
      "accuracy": 0.7466665,
      "accuracy_std_dev": 0.019999888889274717,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.06283585333333332,
      "avg_duration_per_task_ms": 361191.69,
      "avg_output_tokens_per_task": 8869.28,
      "last_run_timestamp": "2026-08-29T08:25:32.049Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Novita",
      "accuracy": 0.7466664000000001,
      "accuracy_std_dev": 0.03126950946593184,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.172299032,
      "avg_duration_per_task_ms": 325786.904,
      "avg_output_tokens_per_task": 11139.324,
      "last_run_timestamp": "2026-09-26T08:08:19.641Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Parasail",
      "accuracy": 0.7463991265432098,
      "accuracy_std_dev": 0.02001099180406232,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.03682306790123457,
      "avg_duration_per_task_ms": 266463.73302469135,
      "avg_output_tokens_per_task": 23459.353395061727,
      "last_run_timestamp": "2026-09-26T16:00:20.166Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "Together",
      "accuracy": 0.7457142857142857,
      "accuracy_std_dev": 0.01866272796407924,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.041890954285714294,
      "avg_duration_per_task_ms": 192572.89428571428,
      "avg_output_tokens_per_task": 12803.885714285714,
      "last_run_timestamp": "2026-09-26T08:05:12.818Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7457142857142857,
      "accuracy_std_dev": 0.048328140467653255,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.08099248857142857,
      "avg_duration_per_task_ms": 179157.66714285716,
      "avg_output_tokens_per_task": 7300.001428571429,
      "last_run_timestamp": "2026-09-26T08:09:25.337Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "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": 0.7455555833333333,
      "accuracy_std_dev": 0.043276290803892344,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.11665106833333336,
      "avg_duration_per_task_ms": 700763.9033333333,
      "avg_output_tokens_per_task": 27544.08,
      "last_run_timestamp": "2026-09-26T08:01:16.576Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Alibaba",
      "accuracy": 0.7455555,
      "accuracy_std_dev": 0.027599586239350292,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.11502652333333335,
      "avg_duration_per_task_ms": 272510.09,
      "avg_output_tokens_per_task": 12128.78,
      "last_run_timestamp": "2026-09-26T08:22:52.642Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7454546363636363,
      "accuracy_std_dev": 0.026180138836609183,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.35241780545454543,
      "avg_duration_per_task_ms": 231797.27818181817,
      "avg_output_tokens_per_task": 10791.974545454545,
      "last_run_timestamp": "2026-09-26T08:23:01.142Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7447618571428571,
      "accuracy_std_dev": 0.035405124090119165,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.08685644000000001,
      "avg_duration_per_task_ms": 256492.48857142858,
      "avg_output_tokens_per_task": 12140.245714285715,
      "last_run_timestamp": "2026-09-26T08:23:17.900Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Sail Research",
      "accuracy": 0.7441947955056181,
      "accuracy_std_dev": 0.013929720069329346,
      "run_count": 9,
      "total_questions": 445,
      "avg_cost_per_task_usd": 0.10463494606741572,
      "avg_duration_per_task_ms": 943614.8921348314,
      "avg_output_tokens_per_task": 21227.357303370787,
      "last_run_timestamp": "2026-09-26T08:24:43.106Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "auto-routing",
      "accuracy": 0.74416675,
      "accuracy_std_dev": 0.021586450870569266,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.32272964499999995,
      "avg_duration_per_task_ms": 199123.9275,
      "avg_output_tokens_per_task": 11446.9275,
      "last_run_timestamp": "2026-09-26T08:10:43.577Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Inceptron",
      "accuracy": 0.7441077474747474,
      "accuracy_std_dev": 0.03625850000000003,
      "run_count": 2,
      "total_questions": 99,
      "avg_cost_per_task_usd": 0.012172767676767676,
      "avg_duration_per_task_ms": 161557.24242424243,
      "avg_output_tokens_per_task": 3853.3636363636365,
      "last_run_timestamp": "2026-09-26T08:10:43.660Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Darkbloom",
      "accuracy": 0.7440306848137535,
      "accuracy_std_dev": 0.019647453058182128,
      "run_count": 7,
      "total_questions": 349,
      "avg_cost_per_task_usd": 0.028864432664756447,
      "avg_duration_per_task_ms": 650673.9828080229,
      "avg_output_tokens_per_task": 21865.07736389685,
      "last_run_timestamp": "2026-09-26T16:00:24.615Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Relace",
      "accuracy": 0.744,
      "accuracy_std_dev": 0.019137165704461043,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.006426796,
      "avg_duration_per_task_ms": 131863.972,
      "avg_output_tokens_per_task": 4262.824,
      "last_run_timestamp": "2026-09-26T08:07:18.986Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Ambient",
      "accuracy": 0.7439998000000001,
      "accuracy_std_dev": 0.025508095447524134,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.010447856000000002,
      "avg_duration_per_task_ms": 427694.796,
      "avg_output_tokens_per_task": 12591.384,
      "last_run_timestamp": "2026-08-29T09:31:13.940Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7433330000000001,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.34578612,
      "avg_duration_per_task_ms": 121381.44,
      "avg_output_tokens_per_task": 6958.96,
      "last_run_timestamp": "2026-09-26T08:08:19.642Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.7430555,
      "accuracy_std_dev": 0.035484932031074834,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.068885335,
      "avg_duration_per_task_ms": 169463.79083333333,
      "avg_output_tokens_per_task": 5588.91,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Together",
      "accuracy": 0.7428571428571429,
      "accuracy_std_dev": 0.042666241494480235,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.09274233999999999,
      "avg_duration_per_task_ms": 75968.94571428571,
      "avg_output_tokens_per_task": 4836.024285714286,
      "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": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.7427776666666667,
      "accuracy_std_dev": 0.036434422742175517,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3671939166666667,
      "avg_duration_per_task_ms": 137596.83,
      "avg_output_tokens_per_task": 5212.336666666667,
      "last_run_timestamp": "2026-09-01T13:00:30.911Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AkashML",
      "accuracy": 0.7426668,
      "accuracy_std_dev": 0.03690245640821219,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.011645404,
      "avg_duration_per_task_ms": 585840.916,
      "avg_output_tokens_per_task": 18794.172,
      "last_run_timestamp": "2026-08-29T08:43:07.210Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "GMICloud",
      "accuracy": 0.7426666000000001,
      "accuracy_std_dev": 0.01236473120775379,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.07381416,
      "avg_duration_per_task_ms": 152125.172,
      "avg_output_tokens_per_task": 4690.728,
      "last_run_timestamp": "2026-09-26T08:27:29.717Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Azure",
      "accuracy": 0.7425,
      "accuracy_std_dev": 0.029988405167664416,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.5165171675,
      "avg_duration_per_task_ms": 228220.77,
      "avg_output_tokens_per_task": 10000.7275,
      "last_run_timestamp": "2026-09-26T08:11:12.012Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.742499875,
      "accuracy_std_dev": 0.021327474782762628,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0092532975,
      "avg_duration_per_task_ms": 294444.1,
      "avg_output_tokens_per_task": 4179.7575,
      "last_run_timestamp": "2026-09-26T08:24:45.222Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.7423371321839081,
      "accuracy_std_dev": 0.08048438687076966,
      "run_count": 21,
      "total_questions": 1044,
      "avg_cost_per_task_usd": 0.051994346743295033,
      "avg_duration_per_task_ms": 148657.38218390805,
      "avg_output_tokens_per_task": 19702.44827586207,
      "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": "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": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7422221666666666,
      "accuracy_std_dev": 0.009937943254964244,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.34229832999999993,
      "avg_duration_per_task_ms": 115494.89,
      "avg_output_tokens_per_task": 7015.326666666667,
      "last_run_timestamp": "2026-09-26T08:16:17.866Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7422221666666666,
      "accuracy_std_dev": 0.029731460662720387,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.34941253333333333,
      "avg_duration_per_task_ms": 221865.17666666667,
      "avg_output_tokens_per_task": 10845.116666666667,
      "last_run_timestamp": "2026-09-26T08:05:12.736Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Azure",
      "accuracy": 0.7422221666666666,
      "accuracy_std_dev": 0.01791636261835036,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3279842533333333,
      "avg_duration_per_task_ms": 196889.34333333332,
      "avg_output_tokens_per_task": 10978.62,
      "last_run_timestamp": "2026-09-01T17:31:47.053Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Together",
      "accuracy": 0.7416666666666667,
      "accuracy_std_dev": 0.0371060392419108,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.15347501333333333,
      "avg_duration_per_task_ms": 347171.24333333335,
      "avg_output_tokens_per_task": 21806.65,
      "last_run_timestamp": "2026-09-26T08:23:03.682Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "NextBit",
      "accuracy": 0.7414969591836733,
      "accuracy_std_dev": 0.023315275936604284,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.01141882142857143,
      "avg_duration_per_task_ms": 212540.27551020408,
      "avg_output_tokens_per_task": 4189.09693877551,
      "last_run_timestamp": "2026-09-26T08:26:46.238Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Sail Research",
      "accuracy": 0.7411113333333333,
      "accuracy_std_dev": 0.028588182830595485,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.018602686666666667,
      "avg_duration_per_task_ms": 242873.4,
      "avg_output_tokens_per_task": 16988.533333333333,
      "last_run_timestamp": "2026-09-26T08:21:38.495Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.7411111666666667,
      "accuracy_std_dev": 0.02732958194836176,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.34793083666666663,
      "avg_duration_per_task_ms": 202714.10666666666,
      "avg_output_tokens_per_task": 11886.45,
      "last_run_timestamp": "2026-09-26T08:09:25.366Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta/muse-glimmer-30b-20260810",
      "model_slug": "meta/muse-glimmer-30b",
      "model_name": "Meta: Muse Glimmer 30B",
      "provider_name": "auto-routing",
      "accuracy": 0.7411110000000001,
      "accuracy_std_dev": 0.033425712732366185,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03801252333333333,
      "avg_duration_per_task_ms": 199405.32333333333,
      "avg_output_tokens_per_task": 13535.723333333333,
      "last_run_timestamp": "2026-09-26T08:09:22.651Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Cerebras",
      "accuracy": 0.7409571552587647,
      "accuracy_std_dev": 0.034434709574561184,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.24114451085141902,
      "avg_duration_per_task_ms": 25708.291318864776,
      "avg_output_tokens_per_task": 5412.737061769616,
      "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": "anthropic/claude-4.6-sonnet-20260217",
      "model_slug": "anthropic/claude-sonnet-4.6",
      "model_name": "Anthropic: Claude Sonnet 4.6",
      "provider_name": "Google",
      "accuracy": 0.740909090909091,
      "accuracy_std_dev": 0.034424213616038424,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.3756752018181817,
      "avg_duration_per_task_ms": 205007.96545454545,
      "avg_output_tokens_per_task": 11919.866363636364,
      "last_run_timestamp": "2026-09-26T08:26:37.377Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Morph",
      "accuracy": 0.7406779661016949,
      "accuracy_std_dev": 0.0192449897289312,
      "run_count": 6,
      "total_questions": 295,
      "avg_cost_per_task_usd": 0.02004586440677966,
      "avg_duration_per_task_ms": 1117850.5423728814,
      "avg_output_tokens_per_task": 22657.922033898307,
      "last_run_timestamp": "2026-09-12T08:11:08.573Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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-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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Phala",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.02380486645100116,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.09328288000000001,
      "avg_duration_per_task_ms": 826540.2925,
      "avg_output_tokens_per_task": 38601.0775,
      "last_run_timestamp": "2026-09-26T16:00:20.164Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "Novita",
      "accuracy": 0.74,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02561014,
      "avg_duration_per_task_ms": 147210.62,
      "avg_output_tokens_per_task": 10787.18,
      "last_run_timestamp": "2026-07-18T08:04:42.215Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Reka",
      "accuracy": 0.74,
      "accuracy_std_dev": 0.01115542718142162,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.014306968,
      "avg_duration_per_task_ms": 188774.736,
      "avg_output_tokens_per_task": 4205.028,
      "last_run_timestamp": "2026-09-26T08:25:40.376Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Wafer",
      "accuracy": 0.7396135507246376,
      "accuracy_std_dev": 0.02509467552355785,
      "run_count": 7,
      "total_questions": 345,
      "avg_cost_per_task_usd": 0.0175928115942029,
      "avg_duration_per_task_ms": 309043.7826086957,
      "avg_output_tokens_per_task": 14060.257971014493,
      "last_run_timestamp": "2026-09-26T08:21:25.884Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7394443333333334,
      "accuracy_std_dev": 0.04158517957825306,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.09852801333333332,
      "avg_duration_per_task_ms": 155104.675,
      "avg_output_tokens_per_task": 8559.225,
      "last_run_timestamp": "2026-09-26T08:22:32.817Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.7391975308641975,
      "accuracy_std_dev": 0.03518784605833174,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.03759480401234569,
      "avg_duration_per_task_ms": 156986.57098765433,
      "avg_output_tokens_per_task": 7739.7793209876545,
      "last_run_timestamp": "2026-09-26T08:12:13.455Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Friendli",
      "accuracy": 0.739160162601626,
      "accuracy_std_dev": 0.016384230901693223,
      "run_count": 5,
      "total_questions": 246,
      "avg_cost_per_task_usd": 0.02238267073170732,
      "avg_duration_per_task_ms": 118042.10162601626,
      "avg_output_tokens_per_task": 3823.4349593495936,
      "last_run_timestamp": "2026-09-26T08:17:01.644Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Morph",
      "accuracy": 0.7391112,
      "accuracy_std_dev": 0.02063675876746798,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.5042409586666666,
      "avg_duration_per_task_ms": 734521.768,
      "avg_output_tokens_per_task": 9377.424,
      "last_run_timestamp": "2026-09-26T08:10:56.862Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "Novita",
      "accuracy": 0.7390475000000001,
      "accuracy_std_dev": 0.022797592068180002,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.022234177142857144,
      "avg_duration_per_task_ms": 351589.08714285714,
      "avg_output_tokens_per_task": 11306.812857142857,
      "last_run_timestamp": "2026-09-26T08:08:19.641Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "CoreWeave",
      "accuracy": 0.738889,
      "accuracy_std_dev": 0.017812306747489716,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.022993023333333338,
      "avg_duration_per_task_ms": 282964.01666666666,
      "avg_output_tokens_per_task": 13071.093333333334,
      "last_run_timestamp": "2026-08-29T16:33:29.042Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "GMICloud",
      "accuracy": 0.7386666000000001,
      "accuracy_std_dev": 0.02777674167068556,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.008449704,
      "avg_duration_per_task_ms": 262399.948,
      "avg_output_tokens_per_task": 5465.996,
      "last_run_timestamp": "2026-09-26T08:30:17.131Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "auto-routing",
      "accuracy": 0.7383335,
      "accuracy_std_dev": 0.007264793338423339,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.029026915000000004,
      "avg_duration_per_task_ms": 190934.755,
      "avg_output_tokens_per_task": 18429.705,
      "last_run_timestamp": "2026-09-26T08:10:43.574Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "Reka",
      "accuracy": 0.7383335,
      "accuracy_std_dev": 0.01092919128984391,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.07321903,
      "avg_duration_per_task_ms": 107189.01,
      "avg_output_tokens_per_task": 4495.765,
      "last_run_timestamp": "2026-09-26T08:22:24.359Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.7383335,
      "accuracy_std_dev": 0.01092919128984391,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.023440455,
      "avg_duration_per_task_ms": 211636.695,
      "avg_output_tokens_per_task": 4181.36,
      "last_run_timestamp": "2026-09-26T08:23:00.240Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Phala",
      "accuracy": 0.73833325,
      "accuracy_std_dev": 0.03345812889175814,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0298002275,
      "avg_duration_per_task_ms": 280974.21,
      "avg_output_tokens_per_task": 17047.3125,
      "last_run_timestamp": "2026-09-26T08:28:05.592Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-20260816",
      "model_slug": "z-ai/glm-5.3",
      "model_name": "Z.ai: GLM 5.3",
      "provider_name": "AkashML",
      "accuracy": 0.7382865983935742,
      "accuracy_std_dev": 0.027461257286584682,
      "run_count": 5,
      "total_questions": 249,
      "avg_cost_per_task_usd": 0.07498479919678715,
      "avg_duration_per_task_ms": 91080.29718875502,
      "avg_output_tokens_per_task": 4367.955823293173,
      "last_run_timestamp": "2026-09-26T08:11:26.495Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "GMICloud",
      "accuracy": 0.7380952857142857,
      "accuracy_std_dev": 0.034862897001639706,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.10877247142857141,
      "avg_duration_per_task_ms": 324815.4542857143,
      "avg_output_tokens_per_task": 18549.505714285715,
      "last_run_timestamp": "2026-09-26T08:13:27.148Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.7377122627345845,
      "accuracy_std_dev": 0.03780777470286003,
      "run_count": 15,
      "total_questions": 746,
      "avg_cost_per_task_usd": 0.027386981233243967,
      "avg_duration_per_task_ms": 394334.3485254692,
      "avg_output_tokens_per_task": 12507.815013404826,
      "last_run_timestamp": "2026-09-26T08:22:26.272Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Venice",
      "accuracy": 0.737695687919463,
      "accuracy_std_dev": 0.024690397243074747,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.12595012919463086,
      "avg_duration_per_task_ms": 492774.60067114094,
      "avg_output_tokens_per_task": 24465.090604026846,
      "last_run_timestamp": "2026-09-26T08:23:04.265Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.7376066307692307,
      "accuracy_std_dev": 0.012768517269048922,
      "run_count": 4,
      "total_questions": 195,
      "avg_cost_per_task_usd": 0.012177446153846153,
      "avg_duration_per_task_ms": 65592.95897435897,
      "avg_output_tokens_per_task": 3712.5128205128203,
      "last_run_timestamp": "2026-09-26T08:16:16.340Z"
    },
    {
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Parasail",
      "accuracy": 0.7375155805243446,
      "accuracy_std_dev": 0.026231812733573265,
      "run_count": 11,
      "total_questions": 534,
      "avg_cost_per_task_usd": 0.017135816479400746,
      "avg_duration_per_task_ms": 93129.53558052435,
      "avg_output_tokens_per_task": 5183.471910112359,
      "last_run_timestamp": "2026-09-26T16:00:25.658Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "AtlasCloud",
      "accuracy": 0.73680004,
      "accuracy_std_dev": 0.029461548394447975,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.033284238400000006,
      "avg_duration_per_task_ms": 120902.7176,
      "avg_output_tokens_per_task": 5596.316,
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Fireworks",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.17216302,
      "avg_duration_per_task_ms": 234612.69,
      "avg_output_tokens_per_task": 17080.8,
      "last_run_timestamp": "2026-09-19T08:12:02.180Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.01666650000000003,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02745565,
      "avg_duration_per_task_ms": 321964.94,
      "avg_output_tokens_per_task": 12731.03,
      "last_run_timestamp": "2026-09-26T08:24:06.205Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-31b-it-20260402",
      "model_slug": "google/gemma-4-31b-it",
      "model_name": "Google: Gemma 4 31B",
      "provider_name": "Phala",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.020005850000000002,
      "avg_duration_per_task_ms": 329585.74,
      "avg_output_tokens_per_task": 11689.47,
      "last_run_timestamp": "2026-08-29T16:43:42.382Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.5-20260423",
      "model_slug": "openai/gpt-5.5",
      "model_name": "OpenAI: GPT-5.5",
      "provider_name": "OpenAI",
      "accuracy": 0.7366665,
      "accuracy_std_dev": 0.05666649999999995,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.5401649000000001,
      "avg_duration_per_task_ms": 161301.85,
      "avg_output_tokens_per_task": 8843.07,
      "last_run_timestamp": "2026-09-01T13:02:17.066Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "Xiaomi",
      "accuracy": 0.7365591666666667,
      "accuracy_std_dev": 0.037933644459181036,
      "run_count": 15,
      "total_questions": 744,
      "avg_cost_per_task_usd": 0.020751344086021507,
      "avg_duration_per_task_ms": 306670.2688172043,
      "avg_output_tokens_per_task": 12001.696236559139,
      "last_run_timestamp": "2026-09-26T08:22:59.882Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Parasail",
      "accuracy": 0.7363080617180984,
      "accuracy_std_dev": 0.1346197222374049,
      "run_count": 24,
      "total_questions": 1199,
      "avg_cost_per_task_usd": 0.09551945788156796,
      "avg_duration_per_task_ms": 256553.70475396165,
      "avg_output_tokens_per_task": 13626.88156797331,
      "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": "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": "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": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.7355556666666666,
      "accuracy_std_dev": 0.012570472951413688,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.020994060000000002,
      "avg_duration_per_task_ms": 93700.16666666667,
      "avg_output_tokens_per_task": 4193.253333333333,
      "last_run_timestamp": "2026-09-26T08:26:42.017Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.5-flash-lite-20260721",
      "model_slug": "google/gemini-3.5-flash-lite",
      "model_name": "Google: Gemini 3.5 Flash Lite",
      "provider_name": "Google AI Studio",
      "accuracy": 0.7355553333333333,
      "accuracy_std_dev": 0.003142853940113798,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.13248254666666667,
      "avg_duration_per_task_ms": 94960.21333333333,
      "avg_output_tokens_per_task": 24879.886666666665,
      "last_run_timestamp": "2026-09-01T17:31:46.230Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Phala",
      "accuracy": 0.7355553333333333,
      "accuracy_std_dev": 0.013698571612479268,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.011050406666666665,
      "avg_duration_per_task_ms": 621248.44,
      "avg_output_tokens_per_task": 4366.6466666666665,
      "last_run_timestamp": "2026-09-26T08:07:18.968Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "GMICloud",
      "accuracy": 0.7350002499999999,
      "accuracy_std_dev": 0.008660205926391138,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.03314667,
      "avg_duration_per_task_ms": 134389.755,
      "avg_output_tokens_per_task": 16623.845,
      "last_run_timestamp": "2026-09-26T08:05:12.731Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "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": 0.7341239911045219,
      "accuracy_std_dev": 0.02480538350740109,
      "run_count": 27,
      "total_questions": 1349,
      "avg_cost_per_task_usd": 0.10910542920681986,
      "avg_duration_per_task_ms": 159024.67679762788,
      "avg_output_tokens_per_task": 8159.174944403262,
      "last_run_timestamp": "2026-09-26T08:08:19.667Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "NextBit",
      "accuracy": 0.7341176470588235,
      "accuracy_std_dev": 0.06332332647233949,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.07183540352941177,
      "avg_duration_per_task_ms": 480891.6094117647,
      "avg_output_tokens_per_task": 22660.84,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Inceptron",
      "accuracy": 0.7339180701754385,
      "accuracy_std_dev": 0.02156426040350099,
      "run_count": 8,
      "total_questions": 399,
      "avg_cost_per_task_usd": 0.009527433583959901,
      "avg_duration_per_task_ms": 516946.6215538847,
      "avg_output_tokens_per_task": 13232.288220551378,
      "last_run_timestamp": "2026-09-26T08:07:18.969Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Anthropic",
      "accuracy": 0.7333335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.34495834000000003,
      "avg_duration_per_task_ms": 101333.53,
      "avg_output_tokens_per_task": 6956.68,
      "last_run_timestamp": "2026-09-26T08:22:24.747Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "auto-routing",
      "accuracy": 0.7333335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.09979016999999998,
      "avg_duration_per_task_ms": 95640.78,
      "avg_output_tokens_per_task": 4171.89,
      "last_run_timestamp": "2026-09-01T12:57:13.483Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "auto-routing",
      "accuracy": 0.733333375,
      "accuracy_std_dev": 0.05196139592749965,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0075224875,
      "avg_duration_per_task_ms": 325334.9425,
      "avg_output_tokens_per_task": 15635.5625,
      "last_run_timestamp": "2026-09-26T08:07:31.852Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7333333333333333,
      "accuracy_std_dev": 0.031972210155418145,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.019457343333333335,
      "avg_duration_per_task_ms": 71958.22,
      "avg_output_tokens_per_task": 3419.1466666666665,
      "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Mancer 2",
      "accuracy": 0.73333325,
      "accuracy_std_dev": 0.004713927363409425,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.093701365,
      "avg_duration_per_task_ms": 409506.12,
      "avg_output_tokens_per_task": 12562.605,
      "last_run_timestamp": "2026-09-26T08:23:24.418Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Together",
      "accuracy": 0.7333332,
      "accuracy_std_dev": 0.01633001326882499,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.016486864,
      "avg_duration_per_task_ms": 85790.632,
      "avg_output_tokens_per_task": 4187.072,
      "last_run_timestamp": "2026-09-26T08:05:55.055Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Google",
      "accuracy": 0.7333331666666667,
      "accuracy_std_dev": 0.012171581962597242,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.3451316866666666,
      "avg_duration_per_task_ms": 86668.75666666667,
      "avg_output_tokens_per_task": 6980.716666666666,
      "last_run_timestamp": "2026-09-26T08:17:01.560Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.34216192,
      "avg_duration_per_task_ms": 118971.98,
      "avg_output_tokens_per_task": 7077.54,
      "last_run_timestamp": "2026-09-26T08:22:25.197Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "NextBit",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.01186947,
      "avg_duration_per_task_ms": 176514.74,
      "avg_output_tokens_per_task": 14201.61,
      "last_run_timestamp": "2026-08-29T16:50:07.898Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7333330000000001,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.058005,
      "avg_duration_per_task_ms": 191243.62,
      "avg_output_tokens_per_task": 30812.32,
      "last_run_timestamp": "2026-09-12T08:09:12.993Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Ionstream",
      "accuracy": 0.7329933877551019,
      "accuracy_std_dev": 0.02075268657909138,
      "run_count": 4,
      "total_questions": 196,
      "avg_cost_per_task_usd": 0.05900951530612245,
      "avg_duration_per_task_ms": 460991.887755102,
      "avg_output_tokens_per_task": 12569.994897959185,
      "last_run_timestamp": "2026-09-26T08:23:24.611Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "ModelRun",
      "accuracy": 0.7328571428571429,
      "accuracy_std_dev": 0.02503509736024079,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.06745117142857142,
      "avg_duration_per_task_ms": 76562.92857142857,
      "avg_output_tokens_per_task": 8496.594285714285,
      "last_run_timestamp": "2026-09-26T08:08:19.642Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "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": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.73166675,
      "accuracy_std_dev": 0.03210208110368392,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.00461813,
      "avg_duration_per_task_ms": 254807.5,
      "avg_output_tokens_per_task": 4301.06,
      "last_run_timestamp": "2026-09-26T08:25:43.420Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "ModelRun",
      "accuracy": 0.7314287142857142,
      "accuracy_std_dev": 0.02301532369478017,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.07236491428571429,
      "avg_duration_per_task_ms": 89556.96,
      "avg_output_tokens_per_task": 10594.671428571428,
      "last_run_timestamp": "2026-09-26T16:00:24.624Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Modal",
      "accuracy": 0.7314287142857142,
      "accuracy_std_dev": 0.009897460680314293,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.39489586285714284,
      "avg_duration_per_task_ms": 142184.54857142858,
      "avg_output_tokens_per_task": 11046.07142857143,
      "last_run_timestamp": "2026-09-26T08:21:22.720Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "DeepInfra",
      "accuracy": 0.7311110000000001,
      "accuracy_std_dev": 0.008314764498569207,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012396426666666667,
      "avg_duration_per_task_ms": 337807.9,
      "avg_output_tokens_per_task": 15691.053333333333,
      "last_run_timestamp": "2026-09-26T08:22:46.339Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Phala",
      "accuracy": 0.7311110000000001,
      "accuracy_std_dev": 0.013698950130089036,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.025903600000000002,
      "avg_duration_per_task_ms": 117172.05333333333,
      "avg_output_tokens_per_task": 15158.646666666667,
      "last_run_timestamp": "2026-09-26T08:07:18.966Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7311110000000001,
      "accuracy_std_dev": 0.035416115465514666,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.3332058333333333,
      "avg_duration_per_task_ms": 117249.60666666667,
      "avg_output_tokens_per_task": 11086.966666666667,
      "last_run_timestamp": "2026-09-01T17:31:47.546Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.7310255769230769,
      "accuracy_std_dev": 0.04276519757499896,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.03939189384615385,
      "avg_duration_per_task_ms": 165400.37076923077,
      "avg_output_tokens_per_task": 5925.508461538461,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DigitalOcean",
      "accuracy": 0.730909,
      "accuracy_std_dev": 0.028180123769003633,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.06749359090909092,
      "avg_duration_per_task_ms": 291533.3,
      "avg_output_tokens_per_task": 9954.196363636363,
      "last_run_timestamp": "2026-09-26T08:23:16.657Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AkashML",
      "accuracy": 0.7305882352941176,
      "accuracy_std_dev": 0.10471396142743807,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.06950796235294117,
      "avg_duration_per_task_ms": 400129.06588235294,
      "avg_output_tokens_per_task": 38127.568235294115,
      "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": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7305263684210526,
      "accuracy_std_dev": 0.033390586105116714,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.35496904315789474,
      "avg_duration_per_task_ms": 206890.78210526315,
      "avg_output_tokens_per_task": 10297.44,
      "last_run_timestamp": "2026-09-26T08:17:01.551Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Relace",
      "accuracy": 0.7304761428571429,
      "accuracy_std_dev": 0.040293813213262725,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.006170451428571428,
      "avg_duration_per_task_ms": 328862.59714285715,
      "avg_output_tokens_per_task": 15818.588571428572,
      "last_run_timestamp": "2026-09-26T08:13:27.139Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "Z.AI",
      "accuracy": 0.7302564615384616,
      "accuracy_std_dev": 0.04434636251775576,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.09595046461538462,
      "avg_duration_per_task_ms": 158127.4846153846,
      "avg_output_tokens_per_task": 8436.901538461538,
      "last_run_timestamp": "2026-09-26T08:22:58.695Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "anthropic/claude-4.6-opus-20260205",
      "model_slug": "anthropic/claude-opus-4.6",
      "model_name": "Anthropic: Claude Opus 4.6",
      "provider_name": "Claude Platform on AWS",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.030489986214056124,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.49638783,
      "avg_duration_per_task_ms": 192253.77333333335,
      "avg_output_tokens_per_task": 9911.976666666667,
      "last_run_timestamp": "2026-09-26T08:22:57.969Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Google",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.0247205998233255,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.36644244,
      "avg_duration_per_task_ms": 230988.18,
      "avg_output_tokens_per_task": 10810.57375,
      "last_run_timestamp": "2026-09-26T08:22:54.473Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.6-flash-20260721",
      "model_slug": "google/gemini-3.6-flash",
      "model_name": "Google: Gemini 3.6 Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.0033330000000000304,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.23651355,
      "avg_duration_per_task_ms": 142691.29,
      "avg_output_tokens_per_task": 22898.27,
      "last_run_timestamp": "2026-09-01T17:31:43.526Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Parasail",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.026874047787782167,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.40094411999999996,
      "avg_duration_per_task_ms": 212371.995,
      "avg_output_tokens_per_task": 8595.31,
      "last_run_timestamp": "2026-09-26T08:23:04.164Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-20260305",
      "model_slug": "openai/gpt-5.4",
      "model_name": "OpenAI: GPT-5.4",
      "provider_name": "OpenAI",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.016666999999999987,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.30719247000000005,
      "avg_duration_per_task_ms": 162813.34,
      "avg_output_tokens_per_task": 10978.12,
      "last_run_timestamp": "2026-09-01T17:32:02.405Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Together",
      "accuracy": 0.73,
      "accuracy_std_dev": 0.01374388753228139,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.243089465,
      "avg_duration_per_task_ms": 228014.985,
      "avg_output_tokens_per_task": 12589.82,
      "last_run_timestamp": "2026-09-26T08:11:43.368Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Novita",
      "accuracy": 0.72986668,
      "accuracy_std_dev": 0.038093252876298195,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.037141388,
      "avg_duration_per_task_ms": 176289.5272,
      "avg_output_tokens_per_task": 6180.716,
      "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": "qwen/qwen3.5-122b-a10b-20260224",
      "model_slug": "qwen/qwen3.5-122b-a10b",
      "model_name": "Qwen: Qwen3.5-122B-A10B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7296182105263157,
      "accuracy_std_dev": 0.10571754149171865,
      "run_count": 13,
      "total_questions": 646,
      "avg_cost_per_task_usd": 0.13478412538699688,
      "avg_duration_per_task_ms": 559761.580495356,
      "avg_output_tokens_per_task": 26074.630030959754,
      "last_run_timestamp": "2026-09-26T08:05:12.733Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Alibaba",
      "accuracy": 0.7295868703170029,
      "accuracy_std_dev": 0.031298136757478015,
      "run_count": 7,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.07363949279538905,
      "avg_duration_per_task_ms": 211055.38616714696,
      "avg_output_tokens_per_task": 8076.533141210374,
      "last_run_timestamp": "2026-09-26T08:14:46.950Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Phala",
      "accuracy": 0.729167,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.044939625000000004,
      "avg_duration_per_task_ms": 321431,
      "avg_output_tokens_per_task": 27916.854166666668,
      "last_run_timestamp": "2026-07-11T08:03:26.385Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Venice",
      "accuracy": 0.72916675,
      "accuracy_std_dev": 0.019346895388860176,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.059703584999999997,
      "avg_duration_per_task_ms": 1171969.47,
      "avg_output_tokens_per_task": 26770.295,
      "last_run_timestamp": "2026-09-26T08:10:43.579Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Modal",
      "accuracy": 0.7288886666666667,
      "accuracy_std_dev": 0.02572412999932599,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.041228259999999996,
      "avg_duration_per_task_ms": 125095.52,
      "avg_output_tokens_per_task": 23895.02666666667,
      "last_run_timestamp": "2026-09-26T08:16:16.958Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7285714285714285,
      "accuracy_std_dev": 0.03504452237962912,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.04105087428571428,
      "avg_duration_per_task_ms": 218582.77142857143,
      "avg_output_tokens_per_task": 20493.322857142855,
      "last_run_timestamp": "2026-09-26T08:07:18.966Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "anthropic/claude-opus-5.5-20260921",
      "model_slug": "anthropic/claude-opus-5.5",
      "model_name": "Anthropic: Claude Opus 5.5",
      "provider_name": "Azure",
      "accuracy": 0.7283335,
      "accuracy_std_dev": 0.02279116544299565,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.345777245,
      "avg_duration_per_task_ms": 85393.745,
      "avg_output_tokens_per_task": 6868.74,
      "last_run_timestamp": "2026-09-26T08:25:13.689Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "GMICloud",
      "accuracy": 0.728187812080537,
      "accuracy_std_dev": 0.031197835362943585,
      "run_count": 12,
      "total_questions": 596,
      "avg_cost_per_task_usd": 0.09276904697986577,
      "avg_duration_per_task_ms": 219176.42953020133,
      "avg_output_tokens_per_task": 8491.71812080537,
      "last_run_timestamp": "2026-09-26T08:23:18.761Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Xiaomi",
      "accuracy": 0.7280700247678018,
      "accuracy_std_dev": 0.022449281062274982,
      "run_count": 13,
      "total_questions": 646,
      "avg_cost_per_task_usd": 0.007610120743034056,
      "avg_duration_per_task_ms": 323992.3544891641,
      "avg_output_tokens_per_task": 14009.664086687306,
      "last_run_timestamp": "2026-09-26T08:05:12.821Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "StreamLake",
      "accuracy": 0.7276787589285714,
      "accuracy_std_dev": 0.023663327303270934,
      "run_count": 9,
      "total_questions": 448,
      "avg_cost_per_task_usd": 0.019942993303571428,
      "avg_duration_per_task_ms": 481307.24776785716,
      "avg_output_tokens_per_task": 14518.11830357143,
      "last_run_timestamp": "2026-08-29T16:31:42.782Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Sail Research",
      "accuracy": 0.7275,
      "accuracy_std_dev": 0.011273272229038026,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.25819418250000004,
      "avg_duration_per_task_ms": 292629.7475,
      "avg_output_tokens_per_task": 10249.3025,
      "last_run_timestamp": "2026-09-26T08:16:17.361Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Novita",
      "accuracy": 0.7272255559006211,
      "accuracy_std_dev": 0.03191875859800738,
      "run_count": 13,
      "total_questions": 644,
      "avg_cost_per_task_usd": 0.04242405434782608,
      "avg_duration_per_task_ms": 215017.30279503105,
      "avg_output_tokens_per_task": 8513.728260869566,
      "last_run_timestamp": "2026-09-26T08:25:16.178Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7272222083333333,
      "accuracy_std_dev": 0.033550248154794085,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.03448022916666667,
      "avg_duration_per_task_ms": 350584.57416666666,
      "avg_output_tokens_per_task": 8108.9575,
      "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "StreamLake",
      "accuracy": 0.7272108979591837,
      "accuracy_std_dev": 0.019828664901097113,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.020857016326530614,
      "avg_duration_per_task_ms": 240921.06530612244,
      "avg_output_tokens_per_task": 10264.273469387756,
      "last_run_timestamp": "2026-09-26T08:13:28.724Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Mara",
      "accuracy": 0.7271692397260274,
      "accuracy_std_dev": 0.01097209078627325,
      "run_count": 3,
      "total_questions": 146,
      "avg_cost_per_task_usd": 0.12186630821917807,
      "avg_duration_per_task_ms": 191102.53424657535,
      "avg_output_tokens_per_task": 9680.431506849314,
      "last_run_timestamp": "2026-09-26T16:00:20.148Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7270795428051002,
      "accuracy_std_dev": 0.09660785679619506,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.0763265227686703,
      "avg_duration_per_task_ms": 209098.53916211292,
      "avg_output_tokens_per_task": 22814.218579234974,
      "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": "AtlasCloud",
      "accuracy": 0.7269444166666666,
      "accuracy_std_dev": 0.03653912012363903,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.03204830166666667,
      "avg_duration_per_task_ms": 247789.1025,
      "avg_output_tokens_per_task": 8437.178333333333,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "SiliconFlow",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.029598939999999997,
      "avg_duration_per_task_ms": 159012.9,
      "avg_output_tokens_per_task": 16200.32,
      "last_run_timestamp": "2026-09-26T08:25:06.306Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.2-20251211",
      "model_slug": "openai/gpt-5.2",
      "model_name": "OpenAI: GPT-5.2",
      "provider_name": "OpenAI",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.27023132,
      "avg_duration_per_task_ms": 228964.68,
      "avg_output_tokens_per_task": 14058.24,
      "last_run_timestamp": "2026-08-01T08:05:59.194Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "tencent/hy4-preview-20260827",
      "model_slug": "tencent/hy4-preview",
      "model_name": "Tencent: Hy4 preview",
      "provider_name": "SiliconFlow",
      "accuracy": 0.726667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06664494,
      "avg_duration_per_task_ms": 229624.28,
      "avg_output_tokens_per_task": 13080.68,
      "last_run_timestamp": "2026-09-23T06:00:08.551Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Relace",
      "accuracy": 0.72666675,
      "accuracy_std_dev": 0.01414201777284626,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.19420406499999998,
      "avg_duration_per_task_ms": 374699.79,
      "avg_output_tokens_per_task": 9514.245,
      "last_run_timestamp": "2026-09-26T08:17:01.582Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Venice",
      "accuracy": 0.7266666363636364,
      "accuracy_std_dev": 0.025897973870460233,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.028845547272727274,
      "avg_duration_per_task_ms": 194541.59636363637,
      "avg_output_tokens_per_task": 9348.309090909092,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "DeepInfra",
      "accuracy": 0.7266665,
      "accuracy_std_dev": 0.020000166668055548,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.46161753000000005,
      "avg_duration_per_task_ms": 792082.91,
      "avg_output_tokens_per_task": 9376.13,
      "last_run_timestamp": "2026-09-26T08:24:06.679Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.7266665,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.20674873,
      "avg_duration_per_task_ms": 122950.24,
      "avg_output_tokens_per_task": 5001.27,
      "last_run_timestamp": "2026-09-01T13:00:51.721Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Morph",
      "accuracy": 0.7266665,
      "accuracy_std_dev": 0.022607791184677924,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.022968040000000002,
      "avg_duration_per_task_ms": 247273.5,
      "avg_output_tokens_per_task": 4817.96,
      "last_run_timestamp": "2026-09-26T08:11:43.349Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Baidu",
      "accuracy": 0.7261538461538461,
      "accuracy_std_dev": 0.03891341591641658,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.04461130461538462,
      "avg_duration_per_task_ms": 117488.42923076924,
      "avg_output_tokens_per_task": 8622.856923076923,
      "last_run_timestamp": "2026-09-26T08:17:01.729Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Novita",
      "accuracy": 0.7260000999999999,
      "accuracy_std_dev": 0.02958217406293864,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.009236828,
      "avg_duration_per_task_ms": 354448.192,
      "avg_output_tokens_per_task": 13566.026,
      "last_run_timestamp": "2026-09-26T08:23:24.666Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Novita",
      "accuracy": 0.725238,
      "accuracy_std_dev": 0.02095781531348834,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.07056013857142857,
      "avg_duration_per_task_ms": 238159.57714285713,
      "avg_output_tokens_per_task": 8221.92,
      "last_run_timestamp": "2026-09-26T08:23:24.553Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7251281538461538,
      "accuracy_std_dev": 0.02890059864923259,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.0717470676923077,
      "avg_duration_per_task_ms": 210149.61076923076,
      "avg_output_tokens_per_task": 7836.80923076923,
      "last_run_timestamp": "2026-09-26T08:23:24.640Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "InferenceNet",
      "accuracy": 0.725,
      "accuracy_std_dev": 0.025000000000000022,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00440039,
      "avg_duration_per_task_ms": 1062673.02,
      "avg_output_tokens_per_task": 13355.81,
      "last_run_timestamp": "2026-09-26T08:07:18.990Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Cloudflare",
      "accuracy": 0.725,
      "accuracy_std_dev": 0.04555216789572149,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0219137175,
      "avg_duration_per_task_ms": 61697.205,
      "avg_output_tokens_per_task": 2955.5725,
      "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": "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": "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DeepInfra",
      "accuracy": 0.7246489750390015,
      "accuracy_std_dev": 0.032220821203342284,
      "run_count": 13,
      "total_questions": 641,
      "avg_cost_per_task_usd": 0.09729576131045242,
      "avg_duration_per_task_ms": 301974.6053042122,
      "avg_output_tokens_per_task": 13658.625585023401,
      "last_run_timestamp": "2026-09-26T08:12:00.292Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "CoreWeave",
      "accuracy": 0.7244446666666666,
      "accuracy_std_dev": 0.003142853940113798,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.02107129333333333,
      "avg_duration_per_task_ms": 190752.33333333334,
      "avg_output_tokens_per_task": 8303.633333333333,
      "last_run_timestamp": "2026-09-26T16:00:20.149Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Venice",
      "accuracy": 0.724127,
      "accuracy_std_dev": 0.08098533957157808,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.06660713428571428,
      "avg_duration_per_task_ms": 276096.2704761905,
      "avg_output_tokens_per_task": 27588.389523809525,
      "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-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": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Together",
      "accuracy": 0.724,
      "accuracy_std_dev": 0.040792156108742275,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.013096376,
      "avg_duration_per_task_ms": 83392.68266666666,
      "avg_output_tokens_per_task": 3561.7066666666665,
      "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7238232103746397,
      "accuracy_std_dev": 0.022993657910877263,
      "run_count": 7,
      "total_questions": 347,
      "avg_cost_per_task_usd": 0.38125144668587896,
      "avg_duration_per_task_ms": 402514.99711815565,
      "avg_output_tokens_per_task": 9102.43227665706,
      "last_run_timestamp": "2026-09-26T08:24:54.432Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Google",
      "accuracy": 0.7236368271604938,
      "accuracy_std_dev": 0.10808272562190256,
      "run_count": 26,
      "total_questions": 1296,
      "avg_cost_per_task_usd": 0.06827893055555555,
      "avg_duration_per_task_ms": 59208.529320987655,
      "avg_output_tokens_per_task": 5716.902777777777,
      "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Novita",
      "accuracy": 0.7236365454545454,
      "accuracy_std_dev": 0.030664462731500284,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.022053703636363636,
      "avg_duration_per_task_ms": 116837.80363636363,
      "avg_output_tokens_per_task": 6700.861818181818,
      "last_run_timestamp": "2026-09-26T16:00:25.697Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7233335,
      "accuracy_std_dev": 0.06333349999999999,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02959498,
      "avg_duration_per_task_ms": 615820.2,
      "avg_output_tokens_per_task": 25536.79,
      "last_run_timestamp": "2026-09-26T08:17:01.598Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "auto-routing",
      "accuracy": 0.7233330000000001,
      "accuracy_std_dev": 0.010000000000000009,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.09717974,
      "avg_duration_per_task_ms": 146982.06,
      "avg_output_tokens_per_task": 51201.48,
      "last_run_timestamp": "2026-09-01T17:23:14.324Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "StreamLake",
      "accuracy": 0.7231801206896552,
      "accuracy_std_dev": 0.03695962243048597,
      "run_count": 14,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.039466484195402295,
      "avg_duration_per_task_ms": 165156.0747126437,
      "avg_output_tokens_per_task": 8083.07040229885,
      "last_run_timestamp": "2026-09-26T08:24:54.334Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Fireworks",
      "accuracy": 0.7226788730593608,
      "accuracy_std_dev": 0.07496294197311405,
      "run_count": 22,
      "total_questions": 1095,
      "avg_cost_per_task_usd": 0.10164467945205478,
      "avg_duration_per_task_ms": 100638.51415525115,
      "avg_output_tokens_per_task": 7403.919634703197,
      "last_run_timestamp": "2026-09-26T08:26:47.303Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Together",
      "accuracy": 0.722499875,
      "accuracy_std_dev": 0.03558044972817764,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.29646921,
      "avg_duration_per_task_ms": 366904.835,
      "avg_output_tokens_per_task": 9370.25,
      "last_run_timestamp": "2026-09-26T08:26:14.256Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "Ionstream",
      "accuracy": 0.7222223333333333,
      "accuracy_std_dev": 0.013698571612479227,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.1329602666666667,
      "avg_duration_per_task_ms": 430783.74,
      "avg_output_tokens_per_task": 33596.9,
      "last_run_timestamp": "2026-09-26T08:22:24.836Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Makora",
      "accuracy": 0.722222,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 48,
      "avg_cost_per_task_usd": 0.019033145833333334,
      "avg_duration_per_task_ms": 65993.54166666667,
      "avg_output_tokens_per_task": 3355.9791666666665,
      "last_run_timestamp": "2026-09-12T08:12:22.607Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Parasail",
      "accuracy": 0.7215385384615385,
      "accuracy_std_dev": 0.032936427919473515,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.1025988369230769,
      "avg_duration_per_task_ms": 105651.07076923076,
      "avg_output_tokens_per_task": 8722.726153846153,
      "last_run_timestamp": "2026-09-26T08:27:14.783Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Novita",
      "accuracy": 0.7214112116788322,
      "accuracy_std_dev": 0.038322526853941594,
      "run_count": 22,
      "total_questions": 1096,
      "avg_cost_per_task_usd": 0.04304255018248176,
      "avg_duration_per_task_ms": 307146.9881386861,
      "avg_output_tokens_per_task": 9393.599452554745,
      "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-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Makora",
      "accuracy": 0.7211113333333333,
      "accuracy_std_dev": 0.019309173922142694,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.04666514,
      "avg_duration_per_task_ms": 136735.63333333333,
      "avg_output_tokens_per_task": 27417.226666666666,
      "last_run_timestamp": "2026-09-26T08:23:22.508Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.7211113333333333,
      "accuracy_std_dev": 0.011331263428624743,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012200653333333334,
      "avg_duration_per_task_ms": 38919.36666666667,
      "avg_output_tokens_per_task": 3634.8866666666668,
      "last_run_timestamp": "2026-09-26T08:23:23.813Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Makora",
      "accuracy": 0.7210882448979592,
      "accuracy_std_dev": 0.0076780598851532815,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.009489710204081632,
      "avg_duration_per_task_ms": 145499.9755102041,
      "avg_output_tokens_per_task": 16260.81224489796,
      "last_run_timestamp": "2026-09-26T08:06:43.933Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.0354337331437149,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.330923255,
      "avg_duration_per_task_ms": 203933.605,
      "avg_output_tokens_per_task": 10272.2525,
      "last_run_timestamp": "2026-09-26T08:10:43.603Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-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": "inclusionai/ling-3.0-flash-vl-20260910",
      "model_slug": "inclusionai/ling-3.0-flash-vl",
      "model_name": "inclusionAI: Ling 3.0 Flash VL",
      "provider_name": "Novita",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0037487199999999997,
      "avg_duration_per_task_ms": 165993.86,
      "avg_output_tokens_per_task": 15330.68,
      "last_run_timestamp": "2026-09-23T18:00:03.326Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-ultra-550b-a55b-20260604",
      "model_slug": "nvidia/nemotron-3-ultra-550b-a55b",
      "model_name": "NVIDIA: Nemotron 3 Ultra",
      "provider_name": "BaseTen",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.15154258,
      "avg_duration_per_task_ms": 105755.46,
      "avg_output_tokens_per_task": 13133.32,
      "last_run_timestamp": "2026-08-04T03:12:01.476Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Azure",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.01845912401316305,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.13026846666666667,
      "avg_duration_per_task_ms": 109723.80666666667,
      "avg_output_tokens_per_task": 4025.5433333333335,
      "last_run_timestamp": "2026-09-01T13:00:52.206Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.14489558,
      "avg_duration_per_task_ms": 248203.88,
      "avg_output_tokens_per_task": 4971.24,
      "last_run_timestamp": "2026-09-23T15:00:07.657Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.7-max-20260520",
      "model_slug": "qwen/qwen3.7-max",
      "model_name": "Qwen: Qwen3.7 Max",
      "provider_name": "AtlasCloud",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.402901,
      "avg_duration_per_task_ms": 196016.4,
      "avg_output_tokens_per_task": 16256.52,
      "last_run_timestamp": "2026-09-27T00:00:03.103Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "Parasail",
      "accuracy": 0.72,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.028463379999999996,
      "avg_duration_per_task_ms": 583244.64,
      "avg_output_tokens_per_task": 43866.54,
      "last_run_timestamp": "2026-08-08T08:27:52.152Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Crusoe",
      "accuracy": 0.72,
      "accuracy_std_dev": 0.0066669999999999785,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.037352619999999996,
      "avg_duration_per_task_ms": 206767.2,
      "avg_output_tokens_per_task": 11957.14,
      "last_run_timestamp": "2026-09-26T08:10:43.574Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "Novita",
      "accuracy": 0.7194656011450382,
      "accuracy_std_dev": 0.0453230167418833,
      "run_count": 21,
      "total_questions": 1048,
      "avg_cost_per_task_usd": 0.04039500858778625,
      "avg_duration_per_task_ms": 306786.39503816795,
      "avg_output_tokens_per_task": 8693.121183206107,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Alibaba",
      "accuracy": 0.7192630452261307,
      "accuracy_std_dev": 0.03505636837505419,
      "run_count": 20,
      "total_questions": 995,
      "avg_cost_per_task_usd": 0.11109498492462312,
      "avg_duration_per_task_ms": 145472.49246231155,
      "avg_output_tokens_per_task": 8249.809045226131,
      "last_run_timestamp": "2026-09-26T08:14:43.050Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "GMICloud",
      "accuracy": 0.7191666250000001,
      "accuracy_std_dev": 0.020666060140587385,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.0268572725,
      "avg_duration_per_task_ms": 236107.135,
      "avg_output_tokens_per_task": 10157.88,
      "last_run_timestamp": "2026-08-29T17:16:15.215Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "auto-routing",
      "accuracy": 0.7190475714285715,
      "accuracy_std_dev": 0.02815552369991236,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.3010150485714286,
      "avg_duration_per_task_ms": 421383.48,
      "avg_output_tokens_per_task": 9299.894285714287,
      "last_run_timestamp": "2026-09-26T08:07:18.968Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "DeepInfra",
      "accuracy": 0.7186897976878613,
      "accuracy_std_dev": 0.027015736117292515,
      "run_count": 7,
      "total_questions": 346,
      "avg_cost_per_task_usd": 0.02169187283236994,
      "avg_duration_per_task_ms": 314303.23988439306,
      "avg_output_tokens_per_task": 7575.85549132948,
      "last_run_timestamp": "2026-08-29T16:48:46.265Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "auto-routing",
      "accuracy": 0.7185660561797753,
      "accuracy_std_dev": 0.0276044455361813,
      "run_count": 25,
      "total_questions": 1246,
      "avg_cost_per_task_usd": 0.0226209036918138,
      "avg_duration_per_task_ms": 287670.0088282504,
      "avg_output_tokens_per_task": 8543.418138041734,
      "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": "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": 0.7184368957915831,
      "accuracy_std_dev": 0.08213389724443859,
      "run_count": 20,
      "total_questions": 998,
      "avg_cost_per_task_usd": 0.05849042284569138,
      "avg_duration_per_task_ms": 494076.376753507,
      "avg_output_tokens_per_task": 37893.49498997996,
      "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": "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Novita",
      "accuracy": 0.71777775,
      "accuracy_std_dev": 0.039283684924798404,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.038988519166666666,
      "avg_duration_per_task_ms": 167526.4525,
      "avg_output_tokens_per_task": 5438.175,
      "last_run_timestamp": "2026-08-29T17:21:26.058Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Phala",
      "accuracy": 0.7177420000000001,
      "accuracy_std_dev": 0.05717499919930039,
      "run_count": 5,
      "total_questions": 248,
      "avg_cost_per_task_usd": 0.2799884032258065,
      "avg_duration_per_task_ms": 377640.3185483871,
      "avg_output_tokens_per_task": 9988.213709677419,
      "last_run_timestamp": "2026-09-26T08:10:43.671Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7177330000000001,
      "accuracy_std_dev": 0.05885214850291977,
      "run_count": 12,
      "total_questions": 594,
      "avg_cost_per_task_usd": 0.12715290404040405,
      "avg_duration_per_task_ms": 1096410.3063973065,
      "avg_output_tokens_per_task": 23858.210437710437,
      "last_run_timestamp": "2026-09-26T08:12:28.093Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Friendli",
      "accuracy": 0.7176923076923077,
      "accuracy_std_dev": 0.02704492670577441,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.06838480846153847,
      "avg_duration_per_task_ms": 195077.87692307692,
      "avg_output_tokens_per_task": 9816.39,
      "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": "anthropic/claude-4.5-sonnet-20250929",
      "model_slug": "anthropic/claude-sonnet-4.5",
      "model_name": "Anthropic: Claude Sonnet 4.5",
      "provider_name": "Azure",
      "accuracy": 0.717499875,
      "accuracy_std_dev": 0.027474714386311185,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.35365771499999993,
      "avg_duration_per_task_ms": 238504.995,
      "avg_output_tokens_per_task": 10803.1675,
      "last_run_timestamp": "2026-09-26T08:10:56.834Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Alibaba",
      "accuracy": 0.7173334,
      "accuracy_std_dev": 0.025508314986294175,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.327251336,
      "avg_duration_per_task_ms": 427459.224,
      "avg_output_tokens_per_task": 9077.78,
      "last_run_timestamp": "2026-09-26T08:23:18.418Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7173334,
      "accuracy_std_dev": 0.050526018820405805,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.171608576,
      "avg_duration_per_task_ms": 290919.704,
      "avg_output_tokens_per_task": 9956.296,
      "last_run_timestamp": "2026-09-26T08:22:17.301Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Chutes",
      "accuracy": 0.7172150901287554,
      "accuracy_std_dev": 0.047887933631073305,
      "run_count": 14,
      "total_questions": 699,
      "avg_cost_per_task_usd": 0.027927947067238917,
      "avg_duration_per_task_ms": 173112.5693848355,
      "avg_output_tokens_per_task": 4396.263233190271,
      "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": "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": 0.7171427857142857,
      "accuracy_std_dev": 0.030728103205322304,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.06684413571428571,
      "avg_duration_per_task_ms": 229014.34285714285,
      "avg_output_tokens_per_task": 7689.901428571428,
      "last_run_timestamp": "2026-09-26T08:05:12.739Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "ModelRun",
      "accuracy": 0.7165880636792453,
      "accuracy_std_dev": 0.056348748072109975,
      "run_count": 17,
      "total_questions": 848,
      "avg_cost_per_task_usd": 0.01914819929245283,
      "avg_duration_per_task_ms": 55455.03773584906,
      "avg_output_tokens_per_task": 4000.7818396226417,
      "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": "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-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Novita",
      "accuracy": 0.7161458482142857,
      "accuracy_std_dev": 0.04047657579390825,
      "run_count": 18,
      "total_questions": 896,
      "avg_cost_per_task_usd": 0.03766043638392857,
      "avg_duration_per_task_ms": 222484.6875,
      "avg_output_tokens_per_task": 6775.514508928572,
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "GMICloud",
      "accuracy": 0.716111,
      "accuracy_std_dev": 0.031056479812324726,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.006504103333333333,
      "avg_duration_per_task_ms": 381723.9266666667,
      "avg_output_tokens_per_task": 12559.69,
      "last_run_timestamp": "2026-09-26T08:05:12.738Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "CoreWeave",
      "accuracy": 0.7155646528925619,
      "accuracy_std_dev": 0.07922982153103719,
      "run_count": 5,
      "total_questions": 242,
      "avg_cost_per_task_usd": 0.11734726446280992,
      "avg_duration_per_task_ms": 254637.0785123967,
      "avg_output_tokens_per_task": 42006.561983471074,
      "last_run_timestamp": "2026-08-29T17:18:52.408Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "OpenInference",
      "accuracy": 0.7155556666666666,
      "accuracy_std_dev": 0.040855070557058414,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.012743326666666667,
      "avg_duration_per_task_ms": 1224143.2066666668,
      "avg_output_tokens_per_task": 18669.653333333332,
      "last_run_timestamp": "2026-08-29T09:20:05.864Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "SiliconFlow",
      "accuracy": 0.7152931648106904,
      "accuracy_std_dev": 0.033648368891044014,
      "run_count": 18,
      "total_questions": 898,
      "avg_cost_per_task_usd": 0.047023087973273946,
      "avg_duration_per_task_ms": 373630.32516703784,
      "avg_output_tokens_per_task": 7851.192650334076,
      "last_run_timestamp": "2026-08-01T17:38:16.661Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "CoreWeave",
      "accuracy": 0.715238,
      "accuracy_std_dev": 0.036811758695907394,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.0628034857142857,
      "avg_duration_per_task_ms": 75005.29714285715,
      "avg_output_tokens_per_task": 9080.965714285714,
      "last_run_timestamp": "2026-09-26T08:24:48.840Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Reka",
      "accuracy": 0.715,
      "accuracy_std_dev": 0.04499999999999999,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.16922785,
      "avg_duration_per_task_ms": 299708.78,
      "avg_output_tokens_per_task": 9901.85,
      "last_run_timestamp": "2026-08-29T08:38:56.982Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.7147222083333333,
      "accuracy_std_dev": 0.03967363744034902,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.03867999416666666,
      "avg_duration_per_task_ms": 344445.3625,
      "avg_output_tokens_per_task": 12226.0975,
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Decart",
      "accuracy": 0.7145061296296297,
      "accuracy_std_dev": 0.033281830759235444,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.10411070679012345,
      "avg_duration_per_task_ms": 92462.91358024691,
      "avg_output_tokens_per_task": 10516.800925925925,
      "last_run_timestamp": "2026-09-26T08:22:18.077Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Parasail",
      "accuracy": 0.7144443333333333,
      "accuracy_std_dev": 0.0015716626723173276,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.04586062666666667,
      "avg_duration_per_task_ms": 212784.84,
      "avg_output_tokens_per_task": 30659.346666666668,
      "last_run_timestamp": "2026-09-26T08:22:27.237Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-mini-2025-08-07",
      "model_slug": "openai/gpt-5-mini",
      "model_name": "OpenAI: GPT-5 Mini",
      "provider_name": "OpenAI",
      "accuracy": 0.7142854897959183,
      "accuracy_std_dev": 0.039861000000000035,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.11979946938775511,
      "avg_duration_per_task_ms": 677914.6734693878,
      "avg_output_tokens_per_task": 54287.663265306124,
      "last_run_timestamp": "2026-09-01T17:30:15.922Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Inceptron",
      "accuracy": 0.7140255173041894,
      "accuracy_std_dev": 0.0345616737838867,
      "run_count": 11,
      "total_questions": 549,
      "avg_cost_per_task_usd": 0.06628586703096538,
      "avg_duration_per_task_ms": 272000.2349726776,
      "avg_output_tokens_per_task": 8738.27504553734,
      "last_run_timestamp": "2026-09-26T08:25:04.482Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "auto-routing",
      "accuracy": 0.7135835603217157,
      "accuracy_std_dev": 0.05817844462211832,
      "run_count": 15,
      "total_questions": 746,
      "avg_cost_per_task_usd": 0.024379721179624666,
      "avg_duration_per_task_ms": 139322.754691689,
      "avg_output_tokens_per_task": 6947.269436997319,
      "last_run_timestamp": "2026-09-26T16:00:18.484Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "NextBit",
      "accuracy": 0.7133335,
      "accuracy_std_dev": 0.026666499999999982,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.028499629999999998,
      "avg_duration_per_task_ms": 176126.96,
      "avg_output_tokens_per_task": 22001.83,
      "last_run_timestamp": "2026-09-26T08:08:19.643Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-pro-20260709",
      "model_slug": "openai/gpt-5.6-luna-pro",
      "model_name": "OpenAI: GPT-5.6 Luna Pro",
      "provider_name": "auto-routing",
      "accuracy": 0.7133335,
      "accuracy_std_dev": 0.013333499999999998,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.06446003,
      "avg_duration_per_task_ms": 348420.47,
      "avg_output_tokens_per_task": 34152.4,
      "last_run_timestamp": "2026-09-01T12:41:25.939Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.7133333913043478,
      "accuracy_std_dev": 0.03565804311334216,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.04301751304347827,
      "avg_duration_per_task_ms": 182842.47913043477,
      "avg_output_tokens_per_task": 6153.3460869565215,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Mancer 2",
      "accuracy": 0.7133333333333334,
      "accuracy_std_dev": 0.04268744285956807,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.04181986333333333,
      "avg_duration_per_task_ms": 299500.48,
      "avg_output_tokens_per_task": 13842.47,
      "last_run_timestamp": "2026-09-26T08:13:28.082Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Together",
      "accuracy": 0.7133333333333334,
      "accuracy_std_dev": 0.02211083193570266,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.03602225,
      "avg_duration_per_task_ms": 85193.79666666666,
      "avg_output_tokens_per_task": 3088.5833333333335,
      "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": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7133333333333334,
      "accuracy_std_dev": 0.019625871060640555,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.12069053999999999,
      "avg_duration_per_task_ms": 78204.86666666667,
      "avg_output_tokens_per_task": 4256.786666666667,
      "last_run_timestamp": "2026-09-01T12:59:40.225Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7133330000000001,
      "accuracy_std_dev": 0.02828427124746193,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.031477000000000005,
      "avg_duration_per_task_ms": 184080.18666666668,
      "avg_output_tokens_per_task": 16016.74,
      "last_run_timestamp": "2026-09-26T08:01:16.573Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-terra-20260709",
      "model_slug": "openai/gpt-5.6-terra",
      "model_name": "OpenAI: GPT-5.6 Terra",
      "provider_name": "OpenAI",
      "accuracy": 0.7133330000000001,
      "accuracy_std_dev": 0.040000000000000036,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.07565915000000001,
      "avg_duration_per_task_ms": 95308.62,
      "avg_output_tokens_per_task": 3758.94,
      "last_run_timestamp": "2026-09-01T13:00:41.334Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "CoreWeave",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.043553119999999994,
      "avg_duration_per_task_ms": 152550.44,
      "avg_output_tokens_per_task": 34185.24,
      "last_run_timestamp": "2026-09-26T08:22:24.717Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "SambaNova",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.14862698,
      "avg_duration_per_task_ms": 506333.72,
      "avg_output_tokens_per_task": 12832.14,
      "last_run_timestamp": "2026-09-26T16:00:25.688Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.6-20260420",
      "model_slug": "moonshotai/kimi-k2.6",
      "model_name": "MoonshotAI: Kimi K2.6",
      "provider_name": "GMICloud",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.09868048,
      "avg_duration_per_task_ms": 318461,
      "avg_output_tokens_per_task": 11794.24,
      "last_run_timestamp": "2026-08-29T17:18:09.782Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "OpenAI",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.57503108,
      "avg_duration_per_task_ms": 657829.88,
      "avg_output_tokens_per_task": 51958.96,
      "last_run_timestamp": "2026-09-01T13:00:31.045Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "DekaLLM",
      "accuracy": 0.713333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0723164,
      "avg_duration_per_task_ms": 346522.38,
      "avg_output_tokens_per_task": 59152.72,
      "last_run_timestamp": "2026-09-12T16:00:24.399Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "BaseTen",
      "accuracy": 0.7130758729674798,
      "accuracy_std_dev": 0.03492932464076567,
      "run_count": 20,
      "total_questions": 984,
      "avg_cost_per_task_usd": 0.09219651422764227,
      "avg_duration_per_task_ms": 75102.03760162601,
      "avg_output_tokens_per_task": 7773.928861788618,
      "last_run_timestamp": "2026-09-26T08:12:44.279Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "BaseTen",
      "accuracy": 0.713058298969072,
      "accuracy_std_dev": 0.025297499999999973,
      "run_count": 2,
      "total_questions": 97,
      "avg_cost_per_task_usd": 0.027766309278350512,
      "avg_duration_per_task_ms": 54271.969072164946,
      "avg_output_tokens_per_task": 11929.432989690722,
      "last_run_timestamp": "2026-09-26T08:12:00.291Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Phala",
      "accuracy": 0.7128571904761904,
      "accuracy_std_dev": 0.03649001975833554,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.043122668571428574,
      "avg_duration_per_task_ms": 201968.71333333335,
      "avg_output_tokens_per_task": 4435.060952380953,
      "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": "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": 0.7127330527950311,
      "accuracy_std_dev": 0.06987907508253477,
      "run_count": 13,
      "total_questions": 644,
      "avg_cost_per_task_usd": 0.04331319254658385,
      "avg_duration_per_task_ms": 245149.61645962732,
      "avg_output_tokens_per_task": 24723.41149068323,
      "last_run_timestamp": "2026-09-26T16:00:24.610Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "CoreWeave",
      "accuracy": 0.712,
      "accuracy_std_dev": 0.03249611258904671,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.08037907,
      "avg_duration_per_task_ms": 74455.182,
      "avg_output_tokens_per_task": 8504.276,
      "last_run_timestamp": "2026-09-26T08:11:40.822Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-2.4t-a95b-20260812",
      "model_slug": "qwen/qwen3.8-2.4t-a95b",
      "model_name": "Qwen: Qwen3.8 2.4T A95B",
      "provider_name": "Venice",
      "accuracy": 0.712,
      "accuracy_std_dev": 0.023626914974240687,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.34044856000000007,
      "avg_duration_per_task_ms": 188002.144,
      "avg_output_tokens_per_task": 11667.488,
      "last_run_timestamp": "2026-09-26T08:22:45.792Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Moonshot AI",
      "accuracy": 0.7119443333333334,
      "accuracy_std_dev": 0.045091638517825244,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.0304885225,
      "avg_duration_per_task_ms": 128203.68083333333,
      "avg_output_tokens_per_task": 5257.9591666666665,
      "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": "openai/gpt-5-2025-08-07",
      "model_slug": "openai/gpt-5",
      "model_name": "OpenAI: GPT-5",
      "provider_name": "Azure",
      "accuracy": 0.7117119324324325,
      "accuracy_std_dev": 0.01247848271581482,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.6902402702702702,
      "avg_duration_per_task_ms": 688360.5945945946,
      "avg_output_tokens_per_task": 56108.03378378379,
      "last_run_timestamp": "2026-09-01T13:02:58.619Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Morph",
      "accuracy": 0.7116669999999999,
      "accuracy_std_dev": 0.008660254037844347,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.013152875000000001,
      "avg_duration_per_task_ms": 1125576.36,
      "avg_output_tokens_per_task": 15885.82,
      "last_run_timestamp": "2026-09-26T08:11:57.022Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "AtlasCloud",
      "accuracy": 0.7115385708502023,
      "accuracy_std_dev": 0.03244045710975723,
      "run_count": 10,
      "total_questions": 494,
      "avg_cost_per_task_usd": 0.021128163967611337,
      "avg_duration_per_task_ms": 99825.94129554655,
      "avg_output_tokens_per_task": 6537.080971659919,
      "last_run_timestamp": "2026-09-26T16:00:20.169Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Together",
      "accuracy": 0.711111,
      "accuracy_std_dev": 0.016629528997138404,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.03228578,
      "avg_duration_per_task_ms": 111202.65333333334,
      "avg_output_tokens_per_task": 22255.406666666666,
      "last_run_timestamp": "2026-09-26T08:17:01.712Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "GMICloud",
      "accuracy": 0.7109091363636364,
      "accuracy_std_dev": 0.03827314426745951,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.04138232000000001,
      "avg_duration_per_task_ms": 208650.05909090908,
      "avg_output_tokens_per_task": 6790.424545454545,
      "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": "qwen/qwen3.5-35b-a3b-20260224",
      "model_slug": "qwen/qwen3.5-35b-a3b",
      "model_name": "Qwen: Qwen3.5-35B-A3B",
      "provider_name": "Parasail",
      "accuracy": 0.710762307174888,
      "accuracy_std_dev": 0.08669531485242507,
      "run_count": 18,
      "total_questions": 892,
      "avg_cost_per_task_usd": 0.05282534865470852,
      "avg_duration_per_task_ms": 474386.9360986547,
      "avg_output_tokens_per_task": 38674.68161434978,
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Venice",
      "accuracy": 0.7107022775919732,
      "accuracy_std_dev": 0.14302186991804494,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.10911909030100336,
      "avg_duration_per_task_ms": 144349.8110367893,
      "avg_output_tokens_per_task": 8000.060200668896,
      "last_run_timestamp": "2026-09-26T08:22:30.723Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "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": "openai/gpt-5.6-sol-20260709",
      "model_slug": "openai/gpt-5.6-sol",
      "model_name": "OpenAI: GPT-5.6 Sol",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.7094593108108108,
      "accuracy_std_dev": 0.025596061967072654,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.29917547972972974,
      "avg_duration_per_task_ms": 177661.23648648648,
      "avg_output_tokens_per_task": 4943.472972972973,
      "last_run_timestamp": "2026-09-01T13:00:41.783Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Venice",
      "accuracy": 0.709333,
      "accuracy_std_dev": 0.034409301068170535,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.23672154,
      "avg_duration_per_task_ms": 252134.184,
      "avg_output_tokens_per_task": 20169.304,
      "last_run_timestamp": "2026-09-26T08:14:47.038Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Fireworks",
      "accuracy": 0.7091838979591837,
      "accuracy_std_dev": 0.022986000000000006,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.021414438775510206,
      "avg_duration_per_task_ms": 174657.33673469388,
      "avg_output_tokens_per_task": 20207.65306122449,
      "last_run_timestamp": "2026-09-26T08:07:31.848Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Relace",
      "accuracy": 0.7091665,
      "accuracy_std_dev": 0.012774768970513724,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.01202336,
      "avg_duration_per_task_ms": 319678.925,
      "avg_output_tokens_per_task": 17962.82,
      "last_run_timestamp": "2026-09-26T08:14:04.244Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Friendli",
      "accuracy": 0.7090909545454546,
      "accuracy_std_dev": 0.07535893638416402,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.060541024545454535,
      "avg_duration_per_task_ms": 57240.09636363637,
      "avg_output_tokens_per_task": 5399.512727272728,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "StreamLake",
      "accuracy": 0.708889,
      "accuracy_std_dev": 0.019115994210782406,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.01979013333333333,
      "avg_duration_per_task_ms": 164929.67333333334,
      "avg_output_tokens_per_task": 14302.1,
      "last_run_timestamp": "2026-09-26T08:17:02.501Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.7088888888888889,
      "accuracy_std_dev": 0.03590109527559924,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.014436273333333333,
      "avg_duration_per_task_ms": 152553.27555555556,
      "avg_output_tokens_per_task": 4675.15,
      "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": "google/gemini-3.1-flash-lite-20260507",
      "model_slug": "google/gemini-3.1-flash-lite",
      "model_name": "Google: Gemini 3.1 Flash Lite",
      "provider_name": "Google",
      "accuracy": 0.7088886666666667,
      "accuracy_std_dev": 0.0339208519579853,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.08842117999999999,
      "avg_duration_per_task_ms": 151404.06666666668,
      "avg_output_tokens_per_task": 46789.46333333333,
      "last_run_timestamp": "2026-09-01T17:31:47.762Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "StreamLake",
      "accuracy": 0.7088886666666667,
      "accuracy_std_dev": 0.025724129999325985,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.062267260000000005,
      "avg_duration_per_task_ms": 262899.92,
      "avg_output_tokens_per_task": 8890.053333333333,
      "last_run_timestamp": "2026-09-26T08:07:15.090Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Together",
      "accuracy": 0.7087542545454545,
      "accuracy_std_dev": 0.05428965504228222,
      "run_count": 10,
      "total_questions": 495,
      "avg_cost_per_task_usd": 0.050922222222222216,
      "avg_duration_per_task_ms": 787128.494949495,
      "avg_output_tokens_per_task": 54707.670707070705,
      "last_run_timestamp": "2026-09-26T16:00:25.696Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7081481666666667,
      "accuracy_std_dev": 0.029106428330162543,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.030871916666666666,
      "avg_duration_per_task_ms": 204088.72444444444,
      "avg_output_tokens_per_task": 6242.532222222222,
      "last_run_timestamp": "2026-08-29T16:33:11.545Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.8-27b-20260814",
      "model_slug": "qwen/qwen3.8-27b",
      "model_name": "Qwen: Qwen3.8 27B",
      "provider_name": "Phala",
      "accuracy": 0.708,
      "accuracy_std_dev": 0.024367428169587355,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.117385436,
      "avg_duration_per_task_ms": 671239.06,
      "avg_output_tokens_per_task": 25704.02,
      "last_run_timestamp": "2026-09-26T08:22:25.076Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Together",
      "accuracy": 0.7073577859531772,
      "accuracy_std_dev": 0.05214988620366833,
      "run_count": 12,
      "total_questions": 598,
      "avg_cost_per_task_usd": 0.0982930769230769,
      "avg_duration_per_task_ms": 117571.96321070234,
      "avg_output_tokens_per_task": 8895.745819397993,
      "last_run_timestamp": "2026-09-26T08:22:36.629Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "auto-routing",
      "accuracy": 0.7073044861111111,
      "accuracy_std_dev": 0.04160167414852507,
      "run_count": 26,
      "total_questions": 1296,
      "avg_cost_per_task_usd": 0.018744344907407406,
      "avg_duration_per_task_ms": 142416.33564814815,
      "avg_output_tokens_per_task": 6045.990740740741,
      "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": "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": "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": "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": 0.70666675,
      "accuracy_std_dev": 0.03496013602930485,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.053884480000000005,
      "avg_duration_per_task_ms": 561783.395,
      "avg_output_tokens_per_task": 5383.42,
      "last_run_timestamp": "2026-09-26T08:28:05.282Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Venice",
      "accuracy": 0.7066666666666667,
      "accuracy_std_dev": 0.010886757164351354,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.042424746666666666,
      "avg_duration_per_task_ms": 131868.91333333333,
      "avg_output_tokens_per_task": 15613.653333333334,
      "last_run_timestamp": "2026-09-26T08:23:01.726Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "moonshotai/kimi-k3-20260715",
      "model_slug": "moonshotai/kimi-k3",
      "model_name": "MoonshotAI: Kimi K3",
      "provider_name": "Fireworks",
      "accuracy": 0.706237370221328,
      "accuracy_std_dev": 0.032130056570911926,
      "run_count": 10,
      "total_questions": 497,
      "avg_cost_per_task_usd": 0.37140224144869216,
      "avg_duration_per_task_ms": 217046.72837022133,
      "avg_output_tokens_per_task": 8219.961770623742,
      "last_run_timestamp": "2026-09-26T08:16:17.130Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "model_slug": "minimax/minimax-m2.7",
      "model_name": "MiniMax: MiniMax M2.7",
      "provider_name": "Minimax",
      "accuracy": 0.706111138888889,
      "accuracy_std_dev": 0.032974923600410835,
      "run_count": 36,
      "total_questions": 1800,
      "avg_cost_per_task_usd": 0.025842573888888876,
      "avg_duration_per_task_ms": 147179.83222222223,
      "avg_output_tokens_per_task": 6878.078888888889,
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "DeepInfra",
      "accuracy": 0.7057822423469389,
      "accuracy_std_dev": 0.08819684229599366,
      "run_count": 8,
      "total_questions": 392,
      "avg_cost_per_task_usd": 0.025861540816326536,
      "avg_duration_per_task_ms": 1299840.9795918367,
      "avg_output_tokens_per_task": 37856.7193877551,
      "last_run_timestamp": "2026-09-26T16:00:24.617Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "Venice",
      "accuracy": 0.7052709777282851,
      "accuracy_std_dev": 0.1455296969729975,
      "run_count": 18,
      "total_questions": 898,
      "avg_cost_per_task_usd": 0.08249193763919824,
      "avg_duration_per_task_ms": 252331.7171492205,
      "avg_output_tokens_per_task": 5844.012249443207,
      "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": "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Parasail",
      "accuracy": 0.704,
      "accuracy_std_dev": 0.047159304490206375,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.022172070000000002,
      "avg_duration_per_task_ms": 76616.802,
      "avg_output_tokens_per_task": 3006.926,
      "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": "xiaomi/mimo-v2.5-20260422",
      "model_slug": "xiaomi/mimo-v2.5",
      "model_name": "Xiaomi: MiMo-V2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.7036363636363636,
      "accuracy_std_dev": 0.0226280216677252,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.008178456363636363,
      "avg_duration_per_task_ms": 314792.5272727273,
      "avg_output_tokens_per_task": 14051.845454545455,
      "last_run_timestamp": "2026-09-26T08:22:46.390Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Wafer",
      "accuracy": 0.7033121966173361,
      "accuracy_std_dev": 0.09340545024178089,
      "run_count": 19,
      "total_questions": 946,
      "avg_cost_per_task_usd": 0.11567343551797042,
      "avg_duration_per_task_ms": 101375.4778012685,
      "avg_output_tokens_per_task": 8107.707188160676,
      "last_run_timestamp": "2026-09-26T08:26:14.351Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.70277775,
      "accuracy_std_dev": 0.05294707729284183,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.039451974166666674,
      "avg_duration_per_task_ms": 188705.17083333334,
      "avg_output_tokens_per_task": 5188.7,
      "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": "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": "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": "minimax/minimax-m3-20260531",
      "model_slug": "minimax/minimax-m3",
      "model_name": "MiniMax: MiniMax M3",
      "provider_name": "Minimax",
      "accuracy": 0.7004762142857143,
      "accuracy_std_dev": 0.07309331857307827,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.024353129999999997,
      "avg_duration_per_task_ms": 88796.36857142857,
      "avg_output_tokens_per_task": 6881.517142857143,
      "last_run_timestamp": "2026-09-26T16:00:24.613Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "Decart",
      "accuracy": 0.7,
      "accuracy_std_dev": 0,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.00948338,
      "avg_duration_per_task_ms": 87248.24,
      "avg_output_tokens_per_task": 5625.9,
      "last_run_timestamp": "2026-09-26T08:10:43.550Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "Near AI",
      "accuracy": 0.7,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.0112321,
      "avg_duration_per_task_ms": 596563.34,
      "avg_output_tokens_per_task": 4097.46,
      "last_run_timestamp": "2026-09-12T09:21:58.569Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.5-0127",
      "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": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "Novita",
      "accuracy": 0.6994872307692308,
      "accuracy_std_dev": 0.031968156822673593,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.03178479384615384,
      "avg_duration_per_task_ms": 200835.56307692308,
      "avg_output_tokens_per_task": 5041.824615384616,
      "last_run_timestamp": "2026-08-01T17:38:23.802Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Inceptron",
      "accuracy": 0.6989348609022557,
      "accuracy_std_dev": 0.11823452094841348,
      "run_count": 11,
      "total_questions": 532,
      "avg_cost_per_task_usd": 0.05432932518796992,
      "avg_duration_per_task_ms": 113791.10338345864,
      "avg_output_tokens_per_task": 7205.951127819549,
      "last_run_timestamp": "2026-09-26T08:08:33.476Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Google",
      "accuracy": 0.6955437880386983,
      "accuracy_std_dev": 0.0690966463341572,
      "run_count": 23,
      "total_questions": 1137,
      "avg_cost_per_task_usd": 0.09496287071240103,
      "avg_duration_per_task_ms": 258193.17766051012,
      "avg_output_tokens_per_task": 6953.824978012313,
      "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": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "Parasail",
      "accuracy": 0.695,
      "accuracy_std_dev": 0.038405728739343024,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.020702055,
      "avg_duration_per_task_ms": 135918.28,
      "avg_output_tokens_per_task": 3857.625,
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "auto-routing",
      "accuracy": 0.6949153322033899,
      "accuracy_std_dev": 0.055158692701843356,
      "run_count": 12,
      "total_questions": 590,
      "avg_cost_per_task_usd": 0.02756941186440678,
      "avg_duration_per_task_ms": 850370.9288135594,
      "avg_output_tokens_per_task": 36904.2406779661,
      "last_run_timestamp": "2026-09-26T16:00:18.485Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.6948148333333334,
      "accuracy_std_dev": 0.02859300565843564,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.024409364444444435,
      "avg_duration_per_task_ms": 211749.8111111111,
      "avg_output_tokens_per_task": 5141.246666666667,
      "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": "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": "qwen/qwen3.5-397b-a17b-20260216",
      "model_slug": "qwen/qwen3.5-397b-a17b",
      "model_name": "Qwen: Qwen3.5 397B A17B",
      "provider_name": "Novita",
      "accuracy": 0.6946935773542601,
      "accuracy_std_dev": 0.21264485878543257,
      "run_count": 18,
      "total_questions": 892,
      "avg_cost_per_task_usd": 0.10174442825112107,
      "avg_duration_per_task_ms": 197126.54596412557,
      "avg_output_tokens_per_task": 10724.269058295964,
      "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": "DigitalOcean",
      "accuracy": 0.694166625,
      "accuracy_std_dev": 0.06141286114475023,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.3036356975,
      "avg_duration_per_task_ms": 394520.48,
      "avg_output_tokens_per_task": 9370.85,
      "last_run_timestamp": "2026-09-26T08:08:19.644Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Mistral",
      "accuracy": 0.6933331428571429,
      "accuracy_std_dev": 0.02519760633715926,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.11111235000000001,
      "avg_duration_per_task_ms": 77728.44714285714,
      "avg_output_tokens_per_task": 9319.027142857143,
      "last_run_timestamp": "2026-09-26T08:24:41.403Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemma-4-26b-a4b-it-20260403",
      "model_slug": "google/gemma-4-26b-a4b-it",
      "model_name": "Google: Gemma 4 26B A4B ",
      "provider_name": "Ionstream",
      "accuracy": 0.693333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.015592,
      "avg_duration_per_task_ms": 1084747.56,
      "avg_output_tokens_per_task": 17745.48,
      "last_run_timestamp": "2026-07-31T03:44:20.825Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "stepfun/step-3.7-flash-20260528",
      "model_slug": "stepfun/step-3.7-flash",
      "model_name": "StepFun: Step 3.7 Flash",
      "provider_name": "Ambient",
      "accuracy": 0.693333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.02187568,
      "avg_duration_per_task_ms": 236451.76,
      "avg_output_tokens_per_task": 11760.74,
      "last_run_timestamp": "2026-07-18T08:04:03.248Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-27b-20260422",
      "model_slug": "qwen/qwen3.6-27b",
      "model_name": "Qwen: Qwen3.6 27B",
      "provider_name": "Phala",
      "accuracy": 0.6913580481481482,
      "accuracy_std_dev": 0.1030309433019321,
      "run_count": 11,
      "total_questions": 540,
      "avg_cost_per_task_usd": 0.1584400537037037,
      "avg_duration_per_task_ms": 922948.7851851851,
      "avg_output_tokens_per_task": 40801.18148148148,
      "last_run_timestamp": "2026-09-26T08:11:26.492Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Venice",
      "accuracy": 0.6905443180515759,
      "accuracy_std_dev": 0.13570623589352024,
      "run_count": 14,
      "total_questions": 698,
      "avg_cost_per_task_usd": 0.055779584527220635,
      "avg_duration_per_task_ms": 503090.9255014327,
      "avg_output_tokens_per_task": 6839.69770773639,
      "last_run_timestamp": "2026-09-26T08:22:17.866Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-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": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "Azure",
      "accuracy": 0.6900003333333334,
      "accuracy_std_dev": 0.012472191289246483,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.04195019333333334,
      "avg_duration_per_task_ms": 246231.66666666666,
      "avg_output_tokens_per_task": 31170.286666666667,
      "last_run_timestamp": "2026-09-24T09:00:05.215Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v3.2-20251201",
      "model_slug": "deepseek/deepseek-v3.2",
      "model_name": "DeepSeek: DeepSeek V3.2",
      "provider_name": "Phala",
      "accuracy": 0.69,
      "accuracy_std_dev": 0.032088096328701056,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.18925239666666666,
      "avg_duration_per_task_ms": 1912773.68,
      "avg_output_tokens_per_task": 13075.366666666667,
      "last_run_timestamp": "2026-08-29T16:50:22.450Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.038729833462074155,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.04847448249999999,
      "avg_duration_per_task_ms": 102036.4875,
      "avg_output_tokens_per_task": 3335.44,
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "SiliconFlow",
      "accuracy": 0.69,
      "accuracy_std_dev": 0.04999999999999999,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.04272526,
      "avg_duration_per_task_ms": 284841.43,
      "avg_output_tokens_per_task": 3214.01,
      "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": "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "auto-routing",
      "accuracy": 0.6893332666666667,
      "accuracy_std_dev": 0.03976225718352628,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.0426068,
      "avg_duration_per_task_ms": 444126.1826666667,
      "avg_output_tokens_per_task": 13135.968,
      "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-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Friendli",
      "accuracy": 0.689300549382716,
      "accuracy_std_dev": 0.06636481757428885,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.10339247685185184,
      "avg_duration_per_task_ms": 75157.70987654322,
      "avg_output_tokens_per_task": 8117.143518518518,
      "last_run_timestamp": "2026-09-26T08:22:29.056Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.688717923076923,
      "accuracy_std_dev": 0.1628459988939345,
      "run_count": 13,
      "total_questions": 650,
      "avg_cost_per_task_usd": 0.052740692307692315,
      "avg_duration_per_task_ms": 145669.94461538462,
      "avg_output_tokens_per_task": 17932.95076923077,
      "last_run_timestamp": "2026-09-26T16:00:25.686Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "moonshotai/kimi-k2-thinking-20251106",
      "model_slug": "moonshotai/kimi-k2-thinking",
      "model_name": "MoonshotAI: Kimi K2 Thinking",
      "provider_name": "auto-routing",
      "accuracy": 0.6880856967871486,
      "accuracy_std_dev": 0.11576392654211201,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.047004783132530134,
      "avg_duration_per_task_ms": 101938.27108433735,
      "avg_output_tokens_per_task": 4039.1506024096384,
      "last_run_timestamp": "2026-08-01T17:38:27.943Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.5-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-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "OpenAI",
      "accuracy": 0.686667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.1282767,
      "avg_duration_per_task_ms": 113186.12,
      "avg_output_tokens_per_task": 4988.34,
      "last_run_timestamp": "2026-09-23T15:00:14.019Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Azure",
      "accuracy": 0.6866666666666666,
      "accuracy_std_dev": 0.005443174461857887,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.009258993333333333,
      "avg_duration_per_task_ms": 105180.14666666667,
      "avg_output_tokens_per_task": 8673.2,
      "last_run_timestamp": "2026-09-23T06:00:08.557Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.6-luna-20260709",
      "model_slug": "openai/gpt-5.6-luna",
      "model_name": "OpenAI: GPT-5.6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6866663333333333,
      "accuracy_std_dev": 0.024944382578492918,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.013524920000000001,
      "avg_duration_per_task_ms": 73418.06,
      "avg_output_tokens_per_task": 5390.6866666666665,
      "last_run_timestamp": "2026-09-01T17:31:44.289Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Cloudflare",
      "accuracy": 0.6820986111111111,
      "accuracy_std_dev": 0.16187486847591467,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.06144617438271606,
      "avg_duration_per_task_ms": 160426.91666666666,
      "avg_output_tokens_per_task": 7038.361111111111,
      "last_run_timestamp": "2026-09-26T08:23:24.525Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "Xiaomi",
      "accuracy": 0.6816666666666666,
      "accuracy_std_dev": 0.037822685726367385,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.005967318333333332,
      "avg_duration_per_task_ms": 82883.09833333333,
      "avg_output_tokens_per_task": 4062.4183333333335,
      "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": "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": "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": "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": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Cloudflare",
      "accuracy": 0.6802083125,
      "accuracy_std_dev": 0.12506379958681027,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.11843063625000003,
      "avg_duration_per_task_ms": 132394.155,
      "avg_output_tokens_per_task": 7828.2725,
      "last_run_timestamp": "2026-09-26T08:16:17.531Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Inceptron",
      "accuracy": 0.68,
      "accuracy_std_dev": 0.03999999999999998,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.021491569999999998,
      "avg_duration_per_task_ms": 71328.98,
      "avg_output_tokens_per_task": 3122.245,
      "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": "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": "openai/gpt-6-astra-20260903",
      "model_slug": "openai/gpt-6-astra",
      "model_name": "OpenAI: GPT-6 Astra",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.68,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.65240112,
      "avg_duration_per_task_ms": 90576.24,
      "avg_output_tokens_per_task": 3843.06,
      "last_run_timestamp": "2026-09-26T09:24:36.382Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "Together",
      "accuracy": 0.68,
      "accuracy_std_dev": 0,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.088178175,
      "avg_duration_per_task_ms": 74028.285,
      "avg_output_tokens_per_task": 3643.675,
      "last_run_timestamp": "2026-09-19T08:08:39.705Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "thinkingmachines/inkling-20260715",
      "model_slug": "thinkingmachines/inkling",
      "model_name": "Thinking Machines: Inkling",
      "provider_name": "DeepInfra",
      "accuracy": 0.68,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.06724904000000001,
      "avg_duration_per_task_ms": 102008.58,
      "avg_output_tokens_per_task": 2510.66,
      "last_run_timestamp": "2026-08-22T08:23:03.574Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.68,
      "accuracy_std_dev": 0.04472135954999579,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.006251205000000001,
      "avg_duration_per_task_ms": 157772.555,
      "avg_output_tokens_per_task": 5287.735,
      "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": "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "auto-routing",
      "accuracy": 0.6783333333333333,
      "accuracy_std_dev": 0.06706133179576904,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.0056993750000000004,
      "avg_duration_per_task_ms": 43275.725,
      "avg_output_tokens_per_task": 3567.115,
      "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": "deepseek/deepseek-v3.2-exp",
      "model_slug": "deepseek/deepseek-v3.2-exp",
      "model_name": "DeepSeek: DeepSeek V3.2 Exp",
      "provider_name": "Novita",
      "accuracy": 0.6781481111111112,
      "accuracy_std_dev": 0.03020960748005119,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.050625214444444455,
      "avg_duration_per_task_ms": 461844.62888888887,
      "avg_output_tokens_per_task": 6960.154444444444,
      "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": "deepseek/deepseek-v4-flash-vision-exp-20260821",
      "model_slug": "deepseek/deepseek-v4-flash-vision-exp",
      "model_name": "DeepSeek: DeepSeek V4 Flash Vision Exp",
      "provider_name": "Fireworks",
      "accuracy": 0.6777776666666667,
      "accuracy_std_dev": 0.013698571612479227,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.0340762,
      "avg_duration_per_task_ms": 244093.44,
      "avg_output_tokens_per_task": 27086.806666666667,
      "last_run_timestamp": "2026-09-19T08:01:53.393Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "DeepInfra",
      "accuracy": 0.6774691543209876,
      "accuracy_std_dev": 0.13202811911229878,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.04778613271604938,
      "avg_duration_per_task_ms": 151992.8487654321,
      "avg_output_tokens_per_task": 7276.706790123457,
      "last_run_timestamp": "2026-09-26T08:30:17.801Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.7-20251222",
      "model_slug": "z-ai/glm-4.7",
      "model_name": "Z.ai: GLM 4.7",
      "provider_name": "DeepInfra",
      "accuracy": 0.6770175263157895,
      "accuracy_std_dev": 0.07095745902363905,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.02569281263157895,
      "avg_duration_per_task_ms": 247756.84421052632,
      "avg_output_tokens_per_task": 6891.394736842105,
      "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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6768421052631579,
      "accuracy_std_dev": 0.09630283912571388,
      "run_count": 19,
      "total_questions": 950,
      "avg_cost_per_task_usd": 0.031000393684210526,
      "avg_duration_per_task_ms": 226486.0452631579,
      "avg_output_tokens_per_task": 8460.411578947369,
      "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": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "Nebius",
      "accuracy": 0.675675972972973,
      "accuracy_std_dev": 0.02284471405428887,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.04042408108108108,
      "avg_duration_per_task_ms": 518660.3783783784,
      "avg_output_tokens_per_task": 20990.527027027027,
      "last_run_timestamp": "2026-09-26T08:11:26.490Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-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": "moonshotai/kimi-k2.5-0127",
      "model_slug": "moonshotai/kimi-k2.5",
      "model_name": "MoonshotAI: Kimi K2.5",
      "provider_name": "Venice",
      "accuracy": 0.6739257299651568,
      "accuracy_std_dev": 0.124700233270467,
      "run_count": 23,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.04174416724738675,
      "avg_duration_per_task_ms": 113875.53397212544,
      "avg_output_tokens_per_task": 5766.850174216028,
      "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": "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": "openai/gpt-6-sol-20260922",
      "model_slug": "openai/gpt-6-sol",
      "model_name": "OpenAI: GPT-6 Sol",
      "provider_name": "Azure",
      "accuracy": 0.6733334999999999,
      "accuracy_std_dev": 0.00666650000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.14733292,
      "avg_duration_per_task_ms": 130030.77,
      "avg_output_tokens_per_task": 5038.48,
      "last_run_timestamp": "2026-09-23T15:13:09.897Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.2-chat-20251211",
      "model_slug": "openai/gpt-5.2-chat",
      "model_name": "OpenAI: GPT-5.2 Chat",
      "provider_name": "auto-routing",
      "accuracy": 0.673333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.1116277,
      "avg_duration_per_task_ms": 79582.44,
      "avg_output_tokens_per_task": 4280.42,
      "last_run_timestamp": "2026-08-01T08:05:57.504Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-pro-20260922",
      "model_slug": "openai/gpt-6-luna-pro",
      "model_name": "OpenAI: GPT-6 Luna Pro",
      "provider_name": "OpenAI",
      "accuracy": 0.673333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.039595319999999996,
      "avg_duration_per_task_ms": 304692.26,
      "avg_output_tokens_per_task": 31675.6,
      "last_run_timestamp": "2026-09-24T09:00:04.937Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6683333333333333,
      "accuracy_std_dev": 0.06147128206370479,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.07486647166666667,
      "avg_duration_per_task_ms": 623524.1416666667,
      "avg_output_tokens_per_task": 16533.455,
      "last_run_timestamp": "2026-09-26T08:15:33.177Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.6681159130434783,
      "accuracy_std_dev": 0.032598105721055856,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.12360180173913042,
      "avg_duration_per_task_ms": 117551.10695652173,
      "avg_output_tokens_per_task": 12432.393913043477,
      "last_run_timestamp": "2026-09-26T08:11:26.494Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Baidu",
      "accuracy": 0.668,
      "accuracy_std_dev": 0.056,
      "run_count": 10,
      "total_questions": 500,
      "avg_cost_per_task_usd": 0.009852277999999999,
      "avg_duration_per_task_ms": 49533.244,
      "avg_output_tokens_per_task": 3541.168,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Minimax",
      "accuracy": 0.6677193421052631,
      "accuracy_std_dev": 0.061111115868114546,
      "run_count": 38,
      "total_questions": 1900,
      "avg_cost_per_task_usd": 0.024342005789473677,
      "avg_duration_per_task_ms": 113629.73684210527,
      "avg_output_tokens_per_task": 7236.974736842105,
      "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": "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": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "auto-routing",
      "accuracy": 0.6674998750000001,
      "accuracy_std_dev": 0.034710984293295034,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.11729118250000001,
      "avg_duration_per_task_ms": 121864.08,
      "avg_output_tokens_per_task": 12614.4,
      "last_run_timestamp": "2026-09-26T08:10:43.582Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.2-20260616",
      "model_slug": "z-ai/glm-5.2",
      "model_name": "Z.ai: GLM 5.2",
      "provider_name": "Phala",
      "accuracy": 0.667439403400309,
      "accuracy_std_dev": 0.1898774300860501,
      "run_count": 13,
      "total_questions": 647,
      "avg_cost_per_task_usd": 0.0928107310664606,
      "avg_duration_per_task_ms": 257455.93199381762,
      "avg_output_tokens_per_task": 7895.010819165379,
      "last_run_timestamp": "2026-09-26T08:24:45.132Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Minimax",
      "accuracy": 0.6666666774193548,
      "accuracy_std_dev": 0.059532208685657846,
      "run_count": 31,
      "total_questions": 1550,
      "avg_cost_per_task_usd": 0.019698243870967743,
      "avg_duration_per_task_ms": 105867.46322580645,
      "avg_output_tokens_per_task": 6375.503870967742,
      "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/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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "auto-routing",
      "accuracy": 0.6648484545454546,
      "accuracy_std_dev": 0.03613567256846351,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.015155524545454546,
      "avg_duration_per_task_ms": 85746.93454545454,
      "avg_output_tokens_per_task": 5712.461818181818,
      "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "CoreWeave",
      "accuracy": 0.6633333333333333,
      "accuracy_std_dev": 0.016555205069369833,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.01703684,
      "avg_duration_per_task_ms": 144424.58666666667,
      "avg_output_tokens_per_task": 21218.593333333334,
      "last_run_timestamp": "2026-09-26T08:04:53.526Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "OpenAI",
      "accuracy": 0.66,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00819156,
      "avg_duration_per_task_ms": 91056.4,
      "avg_output_tokens_per_task": 8317.36,
      "last_run_timestamp": "2026-09-23T06:00:08.542Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Anthropic",
      "accuracy": 0.658750125,
      "accuracy_std_dev": 0.021339771978851493,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.113134535,
      "avg_duration_per_task_ms": 123266.8825,
      "avg_output_tokens_per_task": 12095.085,
      "last_run_timestamp": "2026-09-26T08:08:19.644Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.6575,
      "accuracy_std_dev": 0.028173569173961615,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.00993166625,
      "avg_duration_per_task_ms": 80526.21125,
      "avg_output_tokens_per_task": 4264.58,
      "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": "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": "Phala",
      "accuracy": 0.6565315,
      "accuracy_std_dev": 0.011837568509716087,
      "run_count": 3,
      "total_questions": 148,
      "avg_cost_per_task_usd": 0.014823912162162164,
      "avg_duration_per_task_ms": 271525.86486486485,
      "avg_output_tokens_per_task": 19985.412162162163,
      "last_run_timestamp": "2026-09-19T08:01:36.456Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "DeepInfra",
      "accuracy": 0.6555556666666666,
      "accuracy_std_dev": 0.03189478845969805,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.023210153333333334,
      "avg_duration_per_task_ms": 207735.58,
      "avg_output_tokens_per_task": 33693.153333333335,
      "last_run_timestamp": "2026-09-26T08:25:43.443Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "AkashML",
      "accuracy": 0.655102163265306,
      "accuracy_std_dev": 0.015608943341559031,
      "run_count": 5,
      "total_questions": 245,
      "avg_cost_per_task_usd": 0.012813330612244898,
      "avg_duration_per_task_ms": 278688.9183673469,
      "avg_output_tokens_per_task": 16201.359183673469,
      "last_run_timestamp": "2026-08-29T16:33:39.228Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "anthropic/claude-4.5-haiku-20251001",
      "model_slug": "anthropic/claude-haiku-4.5",
      "model_name": "Anthropic: Claude Haiku 4.5",
      "provider_name": "Google",
      "accuracy": 0.6549999166666667,
      "accuracy_std_dev": 0.02926570645242861,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.12179273749999998,
      "avg_duration_per_task_ms": 120187.11666666667,
      "avg_output_tokens_per_task": 12516.864166666666,
      "last_run_timestamp": "2026-09-26T08:13:57.277Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Novita",
      "accuracy": 0.6536110833333333,
      "accuracy_std_dev": 0.045804700989560625,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.016743790000000005,
      "avg_duration_per_task_ms": 95651.83166666667,
      "avg_output_tokens_per_task": 6607.111666666667,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Inceptron",
      "accuracy": 0.6534847727272728,
      "accuracy_std_dev": 0.034822512390343396,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.014245854545454545,
      "avg_duration_per_task_ms": 157291.03181818183,
      "avg_output_tokens_per_task": 6496.27,
      "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "Nebius",
      "accuracy": 0.6533335,
      "accuracy_std_dev": 0.053333500000000034,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.22863909999999998,
      "avg_duration_per_task_ms": 717866.38,
      "avg_output_tokens_per_task": 11422.01,
      "last_run_timestamp": "2026-09-26T08:13:25.231Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-6-luna-20260922",
      "model_slug": "openai/gpt-6-luna",
      "model_name": "OpenAI: GPT-6 Luna",
      "provider_name": "Amazon Bedrock",
      "accuracy": 0.653333,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.00923216,
      "avg_duration_per_task_ms": 145753.66,
      "avg_output_tokens_per_task": 8192.38,
      "last_run_timestamp": "2026-09-23T06:00:08.549Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.65226668,
      "accuracy_std_dev": 0.04159829731584696,
      "run_count": 25,
      "total_questions": 1250,
      "avg_cost_per_task_usd": 0.0068952256000000016,
      "avg_duration_per_task_ms": 200888.324,
      "avg_output_tokens_per_task": 7369.9624,
      "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": "Friendli",
      "accuracy": 0.65222225,
      "accuracy_std_dev": 0.05094171990802333,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.019855066666666674,
      "avg_duration_per_task_ms": 138088.43583333332,
      "avg_output_tokens_per_task": 6770.496666666667,
      "last_run_timestamp": "2026-08-29T16:14:58.314Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "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": "z-ai/glm-5.3-flash-20260826",
      "model_slug": "z-ai/glm-5.3-flash",
      "model_name": "Z.ai: GLM 5.3 Flash",
      "provider_name": "DigitalOcean",
      "accuracy": 0.651111,
      "accuracy_std_dev": 0.09803010371649448,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.010906386666666667,
      "avg_duration_per_task_ms": 303175.14,
      "avg_output_tokens_per_task": 4777.48,
      "last_run_timestamp": "2026-09-26T08:13:57.291Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.7-20260318",
      "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": "nvidia/nemotron-3.5-lightning-20260807",
      "model_slug": "nvidia/nemotron-3.5-lightning",
      "model_name": "NVIDIA: Nemotron 3.5 Lightning",
      "provider_name": "auto-routing",
      "accuracy": 0.6476509932885907,
      "accuracy_std_dev": 0.013206828848070314,
      "run_count": 3,
      "total_questions": 149,
      "avg_cost_per_task_usd": 0.018146744966442953,
      "avg_duration_per_task_ms": 194939.0268456376,
      "avg_output_tokens_per_task": 22002.395973154362,
      "last_run_timestamp": "2026-09-26T08:13:57.279Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "auto-routing",
      "accuracy": 0.6466667083333333,
      "accuracy_std_dev": 0.04311917808284303,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.016928467500000002,
      "avg_duration_per_task_ms": 184800.4625,
      "avg_output_tokens_per_task": 7229.21,
      "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": "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "SambaNova",
      "accuracy": 0.6466666666666666,
      "accuracy_std_dev": 0.053748384988656986,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.027454210000000003,
      "avg_duration_per_task_ms": 53778.18666666667,
      "avg_output_tokens_per_task": 3634.7233333333334,
      "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": "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": 0.645833375,
      "accuracy_std_dev": 0.02526261098905604,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.11683335,
      "avg_duration_per_task_ms": 147559.335,
      "avg_output_tokens_per_task": 12607.1725,
      "last_run_timestamp": "2026-09-26T08:22:25.227Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "Venice",
      "accuracy": 0.6456595058430717,
      "accuracy_std_dev": 0.16063234813378102,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.04617719449081804,
      "avg_duration_per_task_ms": 194681.6978297162,
      "avg_output_tokens_per_task": 5492.883138564273,
      "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "Venice",
      "accuracy": 0.6440944614173228,
      "accuracy_std_dev": 0.07748571441835636,
      "run_count": 13,
      "total_questions": 635,
      "avg_cost_per_task_usd": 0.020678966929133857,
      "avg_duration_per_task_ms": 227010.8503937008,
      "avg_output_tokens_per_task": 22675.00472440945,
      "last_run_timestamp": "2026-09-26T16:00:24.609Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-5.4-nano-20260317",
      "model_slug": "openai/gpt-5.4-nano",
      "model_name": "OpenAI: GPT-5.4 Nano",
      "provider_name": "OpenAI",
      "accuracy": 0.6433335,
      "accuracy_std_dev": 0.0033334999999999892,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.034682500000000005,
      "avg_duration_per_task_ms": 159195.12,
      "avg_output_tokens_per_task": 18151.16,
      "last_run_timestamp": "2026-09-01T17:31:43.288Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260423",
      "model_slug": "deepseek/deepseek-v4-flash",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0423",
      "provider_name": "OpenInference",
      "accuracy": 0.6433333333333333,
      "accuracy_std_dev": 0.14455548467015086,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.10531282,
      "avg_duration_per_task_ms": 1453976.54,
      "avg_output_tokens_per_task": 18878.353333333333,
      "last_run_timestamp": "2026-08-08T08:21:31.992Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "WandB",
      "accuracy": 0.6433333125,
      "accuracy_std_dev": 0.0463680879845163,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.03536953125,
      "avg_duration_per_task_ms": 94585.0525,
      "avg_output_tokens_per_task": 4734.50875,
      "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "DeepInfra",
      "accuracy": 0.6431451290322581,
      "accuracy_std_dev": 0.07480941255490514,
      "run_count": 10,
      "total_questions": 496,
      "avg_cost_per_task_usd": 0.028669461693548386,
      "avg_duration_per_task_ms": 341904.4637096774,
      "avg_output_tokens_per_task": 10131.977822580646,
      "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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "Novita",
      "accuracy": 0.6427272272727272,
      "accuracy_std_dev": 0.0360504277719316,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.014966787272727274,
      "avg_duration_per_task_ms": 83301.79454545454,
      "avg_output_tokens_per_task": 5789.916363636364,
      "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-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": "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": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Venice",
      "accuracy": 0.64,
      "accuracy_std_dev": 0.025647213408087847,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.218415728,
      "avg_duration_per_task_ms": 1219489.072,
      "avg_output_tokens_per_task": 31443.888,
      "last_run_timestamp": "2026-08-22T16:17:16.779Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "AkashML",
      "accuracy": 0.6395555333333334,
      "accuracy_std_dev": 0.08663140524841761,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.018671574666666666,
      "avg_duration_per_task_ms": 106669.63333333333,
      "avg_output_tokens_per_task": 3706.6453333333334,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "DeepInfra",
      "accuracy": 0.6392073065902578,
      "accuracy_std_dev": 0.050170485537427534,
      "run_count": 14,
      "total_questions": 698,
      "avg_cost_per_task_usd": 0.011382749283667623,
      "avg_duration_per_task_ms": 156016.08022922635,
      "avg_output_tokens_per_task": 5066.042979942694,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Phala",
      "accuracy": 0.63875,
      "accuracy_std_dev": 0.031021421810178213,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.016371615000000003,
      "avg_duration_per_task_ms": 231785.61,
      "avg_output_tokens_per_task": 12424.35125,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "WandB",
      "accuracy": 0.6383647665094339,
      "accuracy_std_dev": 0.026715950245675562,
      "run_count": 17,
      "total_questions": 848,
      "avg_cost_per_task_usd": 0.011237323113207546,
      "avg_duration_per_task_ms": 166991.90094339623,
      "avg_output_tokens_per_task": 10001.11320754717,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "auto-routing",
      "accuracy": 0.6382114250871079,
      "accuracy_std_dev": 0.03539796279192529,
      "run_count": 23,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.013505973867595823,
      "avg_duration_per_task_ms": 198911.70383275262,
      "avg_output_tokens_per_task": 13526.754355400697,
      "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": "z-ai/glm-5-20260211",
      "model_slug": "z-ai/glm-5",
      "model_name": "Z.ai: GLM 5",
      "provider_name": "DeepInfra",
      "accuracy": 0.6372548823529411,
      "accuracy_std_dev": 0.058613911349040856,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.028036784705882355,
      "avg_duration_per_task_ms": 144149.16352941177,
      "avg_output_tokens_per_task": 4911.451764705882,
      "last_run_timestamp": "2026-07-18T16:00:10.793Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-5.1-20260406",
      "model_slug": "z-ai/glm-5.1",
      "model_name": "Z.ai: GLM 5.1",
      "provider_name": "DeepInfra",
      "accuracy": 0.6365,
      "accuracy_std_dev": 0.07836859560946591,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.06096936600000001,
      "avg_duration_per_task_ms": 147673.036,
      "avg_output_tokens_per_task": 6219.183,
      "last_run_timestamp": "2026-08-22T16:12:20.461Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-v4-flash-20260731",
      "model_slug": "deepseek/deepseek-v4-flash-0731",
      "model_name": "DeepSeek: DeepSeek V4 Flash 0731",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6360152471264368,
      "accuracy_std_dev": 0.13554242758288354,
      "run_count": 7,
      "total_questions": 348,
      "avg_cost_per_task_usd": 0.04763444827586207,
      "avg_duration_per_task_ms": 2222096.040229885,
      "avg_output_tokens_per_task": 71368.38505747127,
      "last_run_timestamp": "2026-09-26T08:08:19.641Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6352777083333333,
      "accuracy_std_dev": 0.04398987634338833,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.019619210000000005,
      "avg_duration_per_task_ms": 153720.92,
      "avg_output_tokens_per_task": 6503.970833333334,
      "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": "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": "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "auto-routing",
      "accuracy": 0.6303740787269682,
      "accuracy_std_dev": 0.08664725240494583,
      "run_count": 24,
      "total_questions": 1194,
      "avg_cost_per_task_usd": 0.05638804941373535,
      "avg_duration_per_task_ms": 323905.2269681742,
      "avg_output_tokens_per_task": 9046.836683417085,
      "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": "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": "qwen/qwen3.5-9b-20260310",
      "model_slug": "qwen/qwen3.5-9b",
      "model_name": "Qwen: Qwen3.5-9B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.6280864185185184,
      "accuracy_std_dev": 0.08641467998795974,
      "run_count": 11,
      "total_questions": 540,
      "avg_cost_per_task_usd": 0.01978988888888889,
      "avg_duration_per_task_ms": 730210.6111111111,
      "avg_output_tokens_per_task": 25514.716666666667,
      "last_run_timestamp": "2026-09-26T16:00:20.168Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "DeepInfra",
      "accuracy": 0.6280215229007634,
      "accuracy_std_dev": 0.052631933879496495,
      "run_count": 21,
      "total_questions": 1048,
      "avg_cost_per_task_usd": 0.03512998282442747,
      "avg_duration_per_task_ms": 195080.4961832061,
      "avg_output_tokens_per_task": 6126.911259541985,
      "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": "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": "xiaomi/mimo-v2.5-pro-20260422",
      "model_slug": "xiaomi/mimo-v2.5-pro",
      "model_name": "Xiaomi: MiMo-V2.5-Pro",
      "provider_name": "AtlasCloud",
      "accuracy": 0.6258440128913444,
      "accuracy_std_dev": 0.22029611561511347,
      "run_count": 11,
      "total_questions": 543,
      "avg_cost_per_task_usd": 0.020638675874769797,
      "avg_duration_per_task_ms": 355913.0128913444,
      "avg_output_tokens_per_task": 11160.08655616943,
      "last_run_timestamp": "2026-09-26T08:14:40.334Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "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": 0.6233333125,
      "accuracy_std_dev": 0.051153368250192906,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.05906399125,
      "avg_duration_per_task_ms": 293035.57125,
      "avg_output_tokens_per_task": 14125.07375,
      "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": "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": 0.6222934707297514,
      "accuracy_std_dev": 0.051089021576005936,
      "run_count": 25,
      "total_questions": 1247,
      "avg_cost_per_task_usd": 0.007040149959903768,
      "avg_duration_per_task_ms": 147687.18203688852,
      "avg_output_tokens_per_task": 7570.056134723336,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "DeepInfra",
      "accuracy": 0.6220456222493888,
      "accuracy_std_dev": 0.040780983283176246,
      "run_count": 33,
      "total_questions": 1636,
      "avg_cost_per_task_usd": 0.014479357579462098,
      "avg_duration_per_task_ms": 232578.68643031784,
      "avg_output_tokens_per_task": 14319.570293398532,
      "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": 0.6214190008032129,
      "accuracy_std_dev": 0.13040949799473348,
      "run_count": 25,
      "total_questions": 1245,
      "avg_cost_per_task_usd": 0.012814858634538153,
      "avg_duration_per_task_ms": 167474.9903614458,
      "avg_output_tokens_per_task": 5278.168674698795,
      "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": 0.6207692692307692,
      "accuracy_std_dev": 0.15044459474544053,
      "run_count": 26,
      "total_questions": 1300,
      "avg_cost_per_task_usd": 0.029133950769230768,
      "avg_duration_per_task_ms": 194019.7423076923,
      "avg_output_tokens_per_task": 5212.69076923077,
      "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": "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": 0.6204417565485362,
      "accuracy_std_dev": 0.06516705835591563,
      "run_count": 26,
      "total_questions": 1298,
      "avg_cost_per_task_usd": 0.00849505469953775,
      "avg_duration_per_task_ms": 165342.74576271186,
      "avg_output_tokens_per_task": 8391.972265023112,
      "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": "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": "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": "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": "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": "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": 0.6148485,
      "accuracy_std_dev": 0.04287317952429849,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.05723192454545456,
      "avg_duration_per_task_ms": 395544.69636363635,
      "avg_output_tokens_per_task": 20268.843636363636,
      "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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "SambaNova",
      "accuracy": 0.613394855704698,
      "accuracy_std_dev": 0.05588312171671833,
      "run_count": 24,
      "total_questions": 1192,
      "avg_cost_per_task_usd": 0.02671478104026846,
      "avg_duration_per_task_ms": 57602.19043624161,
      "avg_output_tokens_per_task": 15352.657718120805,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Parasail",
      "accuracy": 0.6120478419388831,
      "accuracy_std_dev": 0.0713164995507085,
      "run_count": 19,
      "total_questions": 949,
      "avg_cost_per_task_usd": 0.019381050579557434,
      "avg_duration_per_task_ms": 108067.79030558483,
      "avg_output_tokens_per_task": 9183.234984193889,
      "last_run_timestamp": "2026-08-01T17:40:52.342Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_slug": "qwen/qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen: Qwen3 235B A22B Thinking 2507",
      "provider_name": "Novita",
      "accuracy": 0.612021883424408,
      "accuracy_std_dev": 0.055254347051473786,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.09755443624772316,
      "avg_duration_per_task_ms": 391161.3069216758,
      "avg_output_tokens_per_task": 18802.257741347905,
      "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": "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": "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": "moonshotai/kimi-k2.7-code-20260612",
      "model_slug": "moonshotai/kimi-k2.7-code",
      "model_name": "MoonshotAI: Kimi K2.7 Code",
      "provider_name": "GMICloud",
      "accuracy": 0.604,
      "accuracy_std_dev": 0.19783738552457672,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.131018624,
      "avg_duration_per_task_ms": 356662.58,
      "avg_output_tokens_per_task": 7644.116,
      "last_run_timestamp": "2026-09-26T08:24:48.090Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Cerebras",
      "accuracy": 0.603913,
      "accuracy_std_dev": 0.03467796953586577,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.039784393043478264,
      "avg_duration_per_task_ms": 20972.738260869566,
      "avg_output_tokens_per_task": 12506.93304347826,
      "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": "deepseek/deepseek-v4-pro-20260813",
      "model_slug": "deepseek/deepseek-v4-pro-0813",
      "model_name": "DeepSeek: DeepSeek V4 Pro 0813",
      "provider_name": "DigitalOcean",
      "accuracy": 0.6004564109589041,
      "accuracy_std_dev": 0.09338877055873593,
      "run_count": 6,
      "total_questions": 292,
      "avg_cost_per_task_usd": 0.485756647260274,
      "avg_duration_per_task_ms": 2654805.914383562,
      "avg_output_tokens_per_task": 83086.6404109589,
      "last_run_timestamp": "2026-09-26T08:17:02.590Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "AtlasCloud",
      "accuracy": 0.5991111333333333,
      "accuracy_std_dev": 0.04144548035651441,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.02262494266666667,
      "avg_duration_per_task_ms": 44422.645333333334,
      "avg_output_tokens_per_task": 2008.0613333333333,
      "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": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "Venice",
      "accuracy": 0.5955284268292683,
      "accuracy_std_dev": 0.10689581341902113,
      "run_count": 23,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.07586419512195122,
      "avg_duration_per_task_ms": 273117.974738676,
      "avg_output_tokens_per_task": 6308.23606271777,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "Krea",
      "accuracy": 0.595,
      "accuracy_std_dev": 0.08500000000000002,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.15232006,
      "avg_duration_per_task_ms": 976036.98,
      "avg_output_tokens_per_task": 119494.16,
      "last_run_timestamp": "2026-09-26T08:22:20.613Z"
    },
    {
      "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": "xiaomi/mimo-v2-flash-20251210",
      "model_slug": "xiaomi/mimo-v2-flash",
      "model_name": "Xiaomi: MiMo-V2-Flash",
      "provider_name": "Novita",
      "accuracy": 0.5933333333333334,
      "accuracy_std_dev": 0.02748737083745107,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.00537365,
      "avg_duration_per_task_ms": 42524.18666666667,
      "avg_output_tokens_per_task": 1858.8366666666666,
      "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": "BaseTen",
      "accuracy": 0.593333,
      "accuracy_std_dev": 0.020000000000000018,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.04552234,
      "avg_duration_per_task_ms": 32685.39,
      "avg_output_tokens_per_task": 3125.98,
      "last_run_timestamp": "2026-09-26T08:22:32.737Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Phala",
      "accuracy": 0.5911110833333333,
      "accuracy_std_dev": 0.06849353668711902,
      "run_count": 12,
      "total_questions": 600,
      "avg_cost_per_task_usd": 0.03914455666666666,
      "avg_duration_per_task_ms": 242677.80666666667,
      "avg_output_tokens_per_task": 10777.653333333334,
      "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-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": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Io Net",
      "accuracy": 0.588,
      "accuracy_std_dev": 0.027129319932501058,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.013733939999999998,
      "avg_duration_per_task_ms": 254046.672,
      "avg_output_tokens_per_task": 9671.628,
      "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": "deepseek/deepseek-v4.1-flash-20260910",
      "model_slug": "deepseek/deepseek-v4.1-flash",
      "model_name": "DeepSeek: DeepSeek V4.1 Flash",
      "provider_name": "OpenInference",
      "accuracy": 0.586667,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.15373889999999998,
      "avg_duration_per_task_ms": 5757447.74,
      "avg_output_tokens_per_task": 125764.32,
      "last_run_timestamp": "2026-09-23T03:50:04.452Z"
    },
    {
      "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": 0.5866667142857143,
      "accuracy_std_dev": 0.06443891163999389,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.058708722857142855,
      "avg_duration_per_task_ms": 207994.24285714285,
      "avg_output_tokens_per_task": 9282.182857142858,
      "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": "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": 0.5816638697829716,
      "accuracy_std_dev": 0.09157507085172394,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.013990112687813019,
      "avg_duration_per_task_ms": 68737.19449081803,
      "avg_output_tokens_per_task": 2743.449916527546,
      "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": "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": "auto-routing",
      "accuracy": 0.5783332916666667,
      "accuracy_std_dev": 0.06758403856340586,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.02616774333333334,
      "avg_duration_per_task_ms": 66057.43916666666,
      "avg_output_tokens_per_task": 2955.8108333333334,
      "last_run_timestamp": "2026-08-29T16:43:55.088Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "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": "minimax/minimax-m2.1",
      "model_slug": "minimax/minimax-m2.1",
      "model_name": "MiniMax: MiniMax M2.1",
      "provider_name": "AtlasCloud",
      "accuracy": 0.575686294117647,
      "accuracy_std_dev": 0.14503435534089426,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.009561005882352942,
      "avg_duration_per_task_ms": 71150.81176470588,
      "avg_output_tokens_per_task": 3692.6835294117645,
      "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": "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": "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": "qwen/qwen3-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Together",
      "accuracy": 0.57,
      "accuracy_std_dev": 0.03316624790355398,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.04864015,
      "avg_duration_per_task_ms": 28286.98,
      "avg_output_tokens_per_task": 1797.17,
      "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": "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": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "Novita",
      "accuracy": 0.5685557531992688,
      "accuracy_std_dev": 0.05835119458074531,
      "run_count": 22,
      "total_questions": 1094,
      "avg_cost_per_task_usd": 0.092415978976234,
      "avg_duration_per_task_ms": 695361.1864716636,
      "avg_output_tokens_per_task": 15899.566727605119,
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "Novita",
      "accuracy": 0.5653333,
      "accuracy_std_dev": 0.04655220391034133,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.12865686099999998,
      "avg_duration_per_task_ms": 196202.004,
      "avg_output_tokens_per_task": 3859.015,
      "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/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": 0.5633528237259816,
      "accuracy_std_dev": 0.139614019301722,
      "run_count": 24,
      "total_questions": 1197,
      "avg_cost_per_task_usd": 0.024817531328320798,
      "avg_duration_per_task_ms": 144439.48621553884,
      "avg_output_tokens_per_task": 7778.0726817042605,
      "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": "minimax/minimax-m2.5-20260211",
      "model_slug": "minimax/minimax-m2.5",
      "model_name": "MiniMax: MiniMax M2.5",
      "provider_name": "Chutes",
      "accuracy": 0.5590908181818182,
      "accuracy_std_dev": 0.11381447181493397,
      "run_count": 11,
      "total_questions": 550,
      "avg_cost_per_task_usd": 0.030268118181818182,
      "avg_duration_per_task_ms": 253612.38363636364,
      "avg_output_tokens_per_task": 13343.18909090909,
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "SiliconFlow",
      "accuracy": 0.556,
      "accuracy_std_dev": 0.04963869458396342,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.032140744,
      "avg_duration_per_task_ms": 123464.168,
      "avg_output_tokens_per_task": 1432.936,
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "auto-routing",
      "accuracy": 0.5544422404809619,
      "accuracy_std_dev": 0.0556916211755413,
      "run_count": 20,
      "total_questions": 998,
      "avg_cost_per_task_usd": 0.1281926412825651,
      "avg_duration_per_task_ms": 173153.67835671344,
      "avg_output_tokens_per_task": 3644.451903807615,
      "last_run_timestamp": "2026-08-01T17:38:23.537Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "google/gemini-2.5-flash",
      "model_slug": "google/gemini-2.5-flash",
      "model_name": "Google: Gemini 2.5 Flash",
      "provider_name": "Google 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": "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": 0.5405240859232175,
      "accuracy_std_dev": 0.0807969531074554,
      "run_count": 22,
      "total_questions": 1094,
      "avg_cost_per_task_usd": 0.08495231718464354,
      "avg_duration_per_task_ms": 201381.6352833638,
      "avg_output_tokens_per_task": 33884.773308957956,
      "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": "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": "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": "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": 0.5200001111111111,
      "accuracy_std_dev": 0.02454520646772421,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.02573156,
      "avg_duration_per_task_ms": 449048.8288888889,
      "avg_output_tokens_per_task": 52732.037777777776,
      "last_run_timestamp": "2026-09-26T16:00:18.487Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": "deepseek/deepseek-v3.1-terminus",
      "model_slug": "deepseek/deepseek-v3.1-terminus",
      "model_name": "DeepSeek: DeepSeek V3.1 Terminus",
      "provider_name": "AtlasCloud",
      "accuracy": 0.5093939090909091,
      "accuracy_std_dev": 0.26566983901244945,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.022748017272727274,
      "avg_duration_per_task_ms": 109776.03,
      "avg_output_tokens_per_task": 2644.2045454545455,
      "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-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": "WandB",
      "accuracy": 0.5083333125,
      "accuracy_std_dev": 0.0195078848024291,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.031913726249999996,
      "avg_duration_per_task_ms": 788235.6275,
      "avg_output_tokens_per_task": 119361.61125,
      "last_run_timestamp": "2026-07-18T16:00:10.167Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "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": "openai/gpt-oss-20b",
      "model_slug": "openai/gpt-oss-20b",
      "model_name": "OpenAI: gpt-oss-20b",
      "provider_name": "DeepInfra",
      "accuracy": 0.5077973684210526,
      "accuracy_std_dev": 0.01607029428505164,
      "run_count": 7,
      "total_questions": 342,
      "avg_cost_per_task_usd": 0.021449652046783625,
      "avg_duration_per_task_ms": 1262349.283625731,
      "avg_output_tokens_per_task": 119951.4649122807,
      "last_run_timestamp": "2026-08-29T17:16:15.676Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.5033669747474748,
      "accuracy_std_dev": 0.027292446564651823,
      "run_count": 8,
      "total_questions": 396,
      "avg_cost_per_task_usd": 0.025276146464646466,
      "avg_duration_per_task_ms": 577739.3712121212,
      "avg_output_tokens_per_task": 53598.469696969696,
      "last_run_timestamp": "2026-09-26T16:00:25.689Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.5022222,
      "accuracy_std_dev": 0.048422017234862634,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.016079876,
      "avg_duration_per_task_ms": 34106.436,
      "avg_output_tokens_per_task": 1702.6026666666667,
      "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": "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-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Ionstream",
      "accuracy": 0.49866136646586345,
      "accuracy_std_dev": 0.20252963395096038,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.010686857429718876,
      "avg_duration_per_task_ms": 38588.640562249,
      "avg_output_tokens_per_task": 1877.5230923694778,
      "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": "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": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "WandB",
      "accuracy": 0.4979166875,
      "accuracy_std_dev": 0.05113807646548552,
      "run_count": 16,
      "total_questions": 800,
      "avg_cost_per_task_usd": 0.112073925,
      "avg_duration_per_task_ms": 113062.30875,
      "avg_output_tokens_per_task": 1623.53625,
      "last_run_timestamp": "2026-07-18T08:04:08.872Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Novita",
      "accuracy": 0.49777777777777776,
      "accuracy_std_dev": 0.023934065809486697,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.015927537777777778,
      "avg_duration_per_task_ms": 341008.0911111111,
      "avg_output_tokens_per_task": 17430.457777777778,
      "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": "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": "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": 0.49476185714285714,
      "accuracy_std_dev": 0.035539193018527655,
      "run_count": 7,
      "total_questions": 350,
      "avg_cost_per_task_usd": 0.02792878571428572,
      "avg_duration_per_task_ms": 578258.1285714286,
      "avg_output_tokens_per_task": 53943.23142857143,
      "last_run_timestamp": "2026-09-26T16:00:20.149Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.49275365217391304,
      "accuracy_std_dev": 0.05838113659290146,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.042801485217391314,
      "avg_duration_per_task_ms": 191615.06608695653,
      "avg_output_tokens_per_task": 2487.1382608695653,
      "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": "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": 0.485,
      "accuracy_std_dev": 0.044440972086577935,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.05964879999999999,
      "avg_duration_per_task_ms": 16505.4875,
      "avg_output_tokens_per_task": 1559.855,
      "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": "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": "nvidia/nemotron-3-nano-30b-a3b",
      "model_slug": "nvidia/nemotron-3-nano-30b-a3b",
      "model_name": "NVIDIA: Nemotron 3 Nano 30B A3B",
      "provider_name": "Crusoe",
      "accuracy": 0.48243642152466365,
      "accuracy_std_dev": 0.023113682722310937,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.028583977578475337,
      "avg_duration_per_task_ms": 385123.331838565,
      "avg_output_tokens_per_task": 66962.30269058296,
      "last_run_timestamp": "2026-09-26T16:00:20.148Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-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": 0.48060604545454544,
      "accuracy_std_dev": 0.05092710756247356,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.022301401818181816,
      "avg_duration_per_task_ms": 29490.40818181818,
      "avg_output_tokens_per_task": 2280.981818181818,
      "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": "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Google",
      "accuracy": 0.47856448265895946,
      "accuracy_std_dev": 0.04258833481538036,
      "run_count": 14,
      "total_questions": 692,
      "avg_cost_per_task_usd": 0.09286534826589596,
      "avg_duration_per_task_ms": 54402.82225433526,
      "avg_output_tokens_per_task": 3776.9132947976877,
      "last_run_timestamp": "2026-08-15T14:15:58.674Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-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": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "SiliconFlow",
      "accuracy": 0.47626414551083596,
      "accuracy_std_dev": 0.030462766110460025,
      "run_count": 13,
      "total_questions": 646,
      "avg_cost_per_task_usd": 0.030607054179566565,
      "avg_duration_per_task_ms": 1315956.8513931888,
      "avg_output_tokens_per_task": 23348.664086687306,
      "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": "openai/gpt-5-nano-2025-08-07",
      "model_slug": "openai/gpt-5-nano",
      "model_name": "OpenAI: GPT-5 Nano",
      "provider_name": "auto-routing",
      "accuracy": 0.47619069387755103,
      "accuracy_std_dev": 0.04472200000000001,
      "run_count": 2,
      "total_questions": 98,
      "avg_cost_per_task_usd": 0.042756918367346936,
      "avg_duration_per_task_ms": 697970.0612244898,
      "avg_output_tokens_per_task": 99743.38775510204,
      "last_run_timestamp": "2026-09-01T17:27:02.043Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-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": 0.4757525267558528,
      "accuracy_std_dev": 0.0504992076723975,
      "run_count": 24,
      "total_questions": 1196,
      "avg_cost_per_task_usd": 0.015268672240802679,
      "avg_duration_per_task_ms": 37530.696488294314,
      "avg_output_tokens_per_task": 2374.680602006689,
      "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": "deepseek/deepseek-r1-0528",
      "model_slug": "deepseek/deepseek-r1-0528",
      "model_name": "DeepSeek: R1 0528",
      "provider_name": "SiliconFlow",
      "accuracy": 0.47535832526881727,
      "accuracy_std_dev": 0.009377392067212614,
      "run_count": 15,
      "total_questions": 744,
      "avg_cost_per_task_usd": 0.11317686962365592,
      "avg_duration_per_task_ms": 1748531.8629032257,
      "avg_output_tokens_per_task": 29728.6061827957,
      "last_run_timestamp": "2026-08-29T16:20:58.736Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-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": "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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "DeepInfra",
      "accuracy": 0.4725397142857143,
      "accuracy_std_dev": 0.037145532599612084,
      "run_count": 21,
      "total_questions": 1050,
      "avg_cost_per_task_usd": 0.01800119238095238,
      "avg_duration_per_task_ms": 57921.885714285716,
      "avg_output_tokens_per_task": 1903.9952380952382,
      "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": "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-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": 0.4704347391304348,
      "accuracy_std_dev": 0.03718740744464026,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.015934632173913044,
      "avg_duration_per_task_ms": 67733.77913043478,
      "avg_output_tokens_per_task": 2419.831304347826,
      "last_run_timestamp": "2026-08-29T16:22:50.200Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-oss-120b",
      "model_slug": "openai/gpt-oss-120b",
      "model_name": "OpenAI: gpt-oss-120b",
      "provider_name": "Together",
      "accuracy": 0.4674228813263525,
      "accuracy_std_dev": 0.04047160730787914,
      "run_count": 23,
      "total_questions": 1146,
      "avg_cost_per_task_usd": 0.01647780541012216,
      "avg_duration_per_task_ms": 158446.44415357767,
      "avg_output_tokens_per_task": 11545.342931937173,
      "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": "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": 0.4666666666666667,
      "accuracy_std_dev": 0.009428090415820616,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.012639673333333332,
      "avg_duration_per_task_ms": 75585.13333333333,
      "avg_output_tokens_per_task": 1995.3766666666668,
      "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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Nebius",
      "accuracy": 0.46561775,
      "accuracy_std_dev": 0.03549794632569826,
      "run_count": 23,
      "total_questions": 1144,
      "avg_cost_per_task_usd": 0.01601312237762238,
      "avg_duration_per_task_ms": 458927.9431818182,
      "avg_output_tokens_per_task": 13622.401223776224,
      "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-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": 0.46516165886287625,
      "accuracy_std_dev": 0.03724010007510945,
      "run_count": 24,
      "total_questions": 1196,
      "avg_cost_per_task_usd": 0.015563879598662204,
      "avg_duration_per_task_ms": 101650.67391304347,
      "avg_output_tokens_per_task": 2399.9096989966556,
      "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-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": "Together",
      "accuracy": 0.4625,
      "accuracy_std_dev": 0.01198957880828179,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.17673909249999997,
      "avg_duration_per_task_ms": 37836.6175,
      "avg_output_tokens_per_task": 1280.4725,
      "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": "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": "SiliconFlow",
      "accuracy": 0.46,
      "accuracy_std_dev": 0.04242640687119284,
      "run_count": 4,
      "total_questions": 200,
      "avg_cost_per_task_usd": 0.01170933,
      "avg_duration_per_task_ms": 112399.705,
      "avg_output_tokens_per_task": 1534.125,
      "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/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Alibaba",
      "accuracy": 0.4599999444444445,
      "accuracy_std_dev": 0.0405517410689003,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.017139246666666674,
      "avg_duration_per_task_ms": 199527.3188888889,
      "avg_output_tokens_per_task": 16295.887777777778,
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Google",
      "accuracy": 0.4593055833333333,
      "accuracy_std_dev": 0.04214152751237655,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.03695082416666666,
      "avg_duration_per_task_ms": 45466.5775,
      "avg_output_tokens_per_task": 2368.125,
      "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-coder-next-2025-02-03",
      "model_slug": "qwen/qwen3-coder-next",
      "model_name": "Qwen: Qwen3 Coder Next",
      "provider_name": "Novita",
      "accuracy": 0.45861108333333334,
      "accuracy_std_dev": 0.16222959206063398,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.02081920916666667,
      "avg_duration_per_task_ms": 35347.055,
      "avg_output_tokens_per_task": 1812.1541666666667,
      "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-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": 0.45848490909090905,
      "accuracy_std_dev": 0.04206453621078092,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.012749035454545457,
      "avg_duration_per_task_ms": 51588.314545454545,
      "avg_output_tokens_per_task": 2288.438181818182,
      "last_run_timestamp": "2026-08-29T16:43:11.444Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "auto-routing",
      "accuracy": 0.4582608260869565,
      "accuracy_std_dev": 0.050742022611100224,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.05333932869565218,
      "avg_duration_per_task_ms": 66926.48956521739,
      "avg_output_tokens_per_task": 1968.3739130434783,
      "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-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-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Google",
      "accuracy": 0.45686270588235295,
      "accuracy_std_dev": 0.04373821409889245,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.029164419999999996,
      "avg_duration_per_task_ms": 27415.028235294118,
      "avg_output_tokens_per_task": 1606.5,
      "last_run_timestamp": "2026-08-29T16:58:32.929Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "AtlasCloud",
      "accuracy": 0.45644446666666666,
      "accuracy_std_dev": 0.06152943015594697,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.08120367199999999,
      "avg_duration_per_task_ms": 51087.073333333334,
      "avg_output_tokens_per_task": 1477.2133333333334,
      "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.4558333333333333,
      "accuracy_std_dev": 0.0388641151252955,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.019138924166666665,
      "avg_duration_per_task_ms": 114207.36,
      "avg_output_tokens_per_task": 2395.9258333333332,
      "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": "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": 0.4537778,
      "accuracy_std_dev": 0.03489385074422139,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.017512640000000003,
      "avg_duration_per_task_ms": 279411.34266666666,
      "avg_output_tokens_per_task": 7481.072,
      "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "SambaNova",
      "accuracy": 0.452549,
      "accuracy_std_dev": 0.12023692617494842,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.08110716235294117,
      "avg_duration_per_task_ms": 124311.29176470588,
      "avg_output_tokens_per_task": 7596.611764705883,
      "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-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Novita",
      "accuracy": 0.45217395652173914,
      "accuracy_std_dev": 0.04587961773700564,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.06969070695652176,
      "avg_duration_per_task_ms": 68529.97913043479,
      "avg_output_tokens_per_task": 1958.8730434782608,
      "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": 0.45190471428571427,
      "accuracy_std_dev": 0.05152768095670046,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.02711271714285714,
      "avg_duration_per_task_ms": 89812.11571428571,
      "avg_output_tokens_per_task": 1856.527142857143,
      "last_run_timestamp": "2026-07-04T16:00:25.259Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Alibaba",
      "accuracy": 0.4518840869565217,
      "accuracy_std_dev": 0.0396095334977008,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.1784364313043478,
      "avg_duration_per_task_ms": 61838.67391304348,
      "avg_output_tokens_per_task": 1962.6434782608696,
      "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-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": 0.451310823434992,
      "accuracy_std_dev": 0.07639218105535146,
      "run_count": 25,
      "total_questions": 1246,
      "avg_cost_per_task_usd": 0.020971863563402894,
      "avg_duration_per_task_ms": 51140.44863563403,
      "avg_output_tokens_per_task": 2040.5256821829855,
      "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": "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/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "WandB",
      "accuracy": 0.44571428571428573,
      "accuracy_std_dev": 0.046860627048162073,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.011708047142857144,
      "avg_duration_per_task_ms": 28176.38,
      "avg_output_tokens_per_task": 1525.6285714285714,
      "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": "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": 0.4442857142857143,
      "accuracy_std_dev": 0.04609218801400479,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.02215216285714286,
      "avg_duration_per_task_ms": 36972.67428571429,
      "avg_output_tokens_per_task": 1609.7742857142857,
      "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": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "auto-routing",
      "accuracy": 0.4411765294117647,
      "accuracy_std_dev": 0.05550646750988218,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.016216156470588234,
      "avg_duration_per_task_ms": 197255.68705882353,
      "avg_output_tokens_per_task": 10840.736470588236,
      "last_run_timestamp": "2026-08-01T17:38:32.120Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "auto-routing",
      "accuracy": 0.44073454757929886,
      "accuracy_std_dev": 0.07859861818914539,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.015726146076794657,
      "avg_duration_per_task_ms": 440405.62771285477,
      "avg_output_tokens_per_task": 13779.416527545909,
      "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": "openai/gpt-4.1-mini-2025-04-14",
      "model_slug": "openai/gpt-4.1-mini",
      "model_name": "OpenAI: GPT-4.1 Mini",
      "provider_name": "auto-routing",
      "accuracy": 0.44,
      "accuracy_std_dev": 0.013332999999999984,
      "run_count": 2,
      "total_questions": 100,
      "avg_cost_per_task_usd": 0.02857548,
      "avg_duration_per_task_ms": 74491.22,
      "avg_output_tokens_per_task": 2449.6,
      "last_run_timestamp": "2026-09-01T17:30:53.043Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-2024-05-13",
      "model_slug": "openai/gpt-4o-2024-05-13",
      "model_name": "OpenAI: GPT-4o (2024-05-13)",
      "provider_name": "OpenAI",
      "accuracy": 0.44,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.7291186999999999,
      "avg_duration_per_task_ms": 35507.54,
      "avg_output_tokens_per_task": 3272.36,
      "last_run_timestamp": "2026-08-01T08:05:57.702Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "AtlasCloud",
      "accuracy": 0.43972229166666665,
      "accuracy_std_dev": 0.05926957176640695,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.03497607833333332,
      "avg_duration_per_task_ms": 88448.07583333334,
      "avg_output_tokens_per_task": 2351.838333333333,
      "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": "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/qwen3.6-35b-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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "SiliconFlow",
      "accuracy": 0.4370816445783133,
      "accuracy_std_dev": 0.023605829668696253,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.07084799598393576,
      "avg_duration_per_task_ms": 649532.4809236948,
      "avg_output_tokens_per_task": 12783.55421686747,
      "last_run_timestamp": "2026-08-29T16:10:09.206Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o",
      "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": "Parasail",
      "accuracy": 0.4359688229398664,
      "accuracy_std_dev": 0.11718231670692668,
      "run_count": 18,
      "total_questions": 898,
      "avg_cost_per_task_usd": 0.009257310690423163,
      "avg_duration_per_task_ms": 55520.83184855234,
      "avg_output_tokens_per_task": 1748.1336302895322,
      "last_run_timestamp": "2026-08-29T17:44:16.832Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-coder-480b-a35b-07-25",
      "model_slug": "qwen/qwen3-coder",
      "model_name": "Qwen: Qwen3 Coder 480B A35B",
      "provider_name": "Venice",
      "accuracy": 0.43594414571949,
      "accuracy_std_dev": 0.06435200699203476,
      "run_count": 22,
      "total_questions": 1098,
      "avg_cost_per_task_usd": 0.011169334244080141,
      "avg_duration_per_task_ms": 99438.00273224043,
      "avg_output_tokens_per_task": 1867.516393442623,
      "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": "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": 0.435,
      "accuracy_std_dev": 0.029580398915498084,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.012494922499999998,
      "avg_duration_per_task_ms": 17298.03,
      "avg_output_tokens_per_task": 1157.0075,
      "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": "DeepInfra",
      "accuracy": 0.43422520234113715,
      "accuracy_std_dev": 0.04823165186928528,
      "run_count": 24,
      "total_questions": 1196,
      "avg_cost_per_task_usd": 0.015668311036789298,
      "avg_duration_per_task_ms": 173534.67224080267,
      "avg_output_tokens_per_task": 2336.4021739130435,
      "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-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": "Novita",
      "accuracy": 0.43236108333333334,
      "accuracy_std_dev": 0.05038259370549571,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.0171767125,
      "avg_duration_per_task_ms": 123156.45833333333,
      "avg_output_tokens_per_task": 2417.2791666666667,
      "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": 0.43172692570281124,
      "accuracy_std_dev": 0.039473708180097544,
      "run_count": 20,
      "total_questions": 996,
      "avg_cost_per_task_usd": 0.04319263554216868,
      "avg_duration_per_task_ms": 361663.5903614458,
      "avg_output_tokens_per_task": 8970.2640562249,
      "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-coder-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-coder-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 Coder 30B A3B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.4311113333333333,
      "accuracy_std_dev": 0.006285236475706817,
      "run_count": 3,
      "total_questions": 150,
      "avg_cost_per_task_usd": 0.023601506666666664,
      "avg_duration_per_task_ms": 216917.58666666667,
      "avg_output_tokens_per_task": 3146.213333333333,
      "last_run_timestamp": "2026-08-15T08:02:48.904Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "Alibaba",
      "accuracy": 0.4310830042194093,
      "accuracy_std_dev": 0.06328266797880157,
      "run_count": 19,
      "total_questions": 948,
      "avg_cost_per_task_usd": 0.02028051582278481,
      "avg_duration_per_task_ms": 151585.09915611814,
      "avg_output_tokens_per_task": 10897.122362869199,
      "last_run_timestamp": "2026-08-01T17:38:29.532Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-235b-a22b-07-25",
      "model_slug": "qwen/qwen3-235b-a22b-2507",
      "model_name": "Qwen: Qwen3 235B A22B Instruct 2507",
      "provider_name": "Alibaba",
      "accuracy": 0.4302777916666667,
      "accuracy_std_dev": 0.04311821560545223,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.02642653999999999,
      "avg_duration_per_task_ms": 53021.551666666666,
      "avg_output_tokens_per_task": 2356.3941666666665,
      "last_run_timestamp": "2026-08-29T16:10:58.039Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "Novita",
      "accuracy": 0.42857142857142855,
      "accuracy_std_dev": 0.03681171064778607,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.02296006857142857,
      "avg_duration_per_task_ms": 94686.24857142857,
      "avg_output_tokens_per_task": 2287.175714285714,
      "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": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "DeepInfra",
      "accuracy": 0.42799069860627176,
      "accuracy_std_dev": 0.06327294802110665,
      "run_count": 23,
      "total_questions": 1148,
      "avg_cost_per_task_usd": 0.013504289198606273,
      "avg_duration_per_task_ms": 250638.181184669,
      "avg_output_tokens_per_task": 12498.641986062717,
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "DeepInfra",
      "accuracy": 0.42499063860830527,
      "accuracy_std_dev": 0.04030000519747692,
      "run_count": 18,
      "total_questions": 891,
      "avg_cost_per_task_usd": 0.020175196408529744,
      "avg_duration_per_task_ms": 230086.4859708193,
      "avg_output_tokens_per_task": 2539.8159371492707,
      "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/qwen3-30b-a3b-04-28",
      "model_slug": "qwen/qwen3-30b-a3b",
      "model_name": "Qwen: Qwen3 30B A3B",
      "provider_name": "DeepInfra",
      "accuracy": 0.4203703888888889,
      "accuracy_std_dev": 0.04136388473339579,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.01674081333333333,
      "avg_duration_per_task_ms": 216767.63555555555,
      "avg_output_tokens_per_task": 11157.038888888888,
      "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": "auto-routing",
      "accuracy": 0.413888875,
      "accuracy_std_dev": 0.04059353491866295,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.03607687833333334,
      "avg_duration_per_task_ms": 151776.1875,
      "avg_output_tokens_per_task": 3792.6,
      "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": "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": 0.4136363181818182,
      "accuracy_std_dev": 0.04711807131739889,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.02628960454545455,
      "avg_duration_per_task_ms": 119464.79909090909,
      "avg_output_tokens_per_task": 2606.97,
      "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": "moonshotai/kimi-k2-0905",
      "model_slug": "moonshotai/kimi-k2-0905",
      "model_name": "MoonshotAI: Kimi K2 0905",
      "provider_name": "AtlasCloud",
      "accuracy": 0.40956652017937223,
      "accuracy_std_dev": 0.04505436755688195,
      "run_count": 9,
      "total_questions": 446,
      "avg_cost_per_task_usd": 0.7669130919282512,
      "avg_duration_per_task_ms": 375326.5717488789,
      "avg_output_tokens_per_task": 14899.289237668161,
      "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": "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": 0.4070327283236994,
      "accuracy_std_dev": 0.022772654274634263,
      "run_count": 21,
      "total_questions": 1038,
      "avg_cost_per_task_usd": 0.006255626204238921,
      "avg_duration_per_task_ms": 19009.97302504817,
      "avg_output_tokens_per_task": 331.504816955684,
      "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": 0.40347830434782606,
      "accuracy_std_dev": 0.03645492416006468,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.017620779130434784,
      "avg_duration_per_task_ms": 171847.04,
      "avg_output_tokens_per_task": 2749.1495652173912,
      "last_run_timestamp": "2026-08-29T16:17:08.788Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "WandB",
      "accuracy": 0.4032566666666667,
      "accuracy_std_dev": 0.0558289469572196,
      "run_count": 14,
      "total_questions": 696,
      "avg_cost_per_task_usd": 0.22458803448275863,
      "avg_duration_per_task_ms": 344124.2744252874,
      "avg_output_tokens_per_task": 17850.82471264368,
      "last_run_timestamp": "2026-07-18T16:00:08.906Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.40094352201257866,
      "accuracy_std_dev": 0.042095981346942486,
      "run_count": 13,
      "total_questions": 636,
      "avg_cost_per_task_usd": 0.05157444339622641,
      "avg_duration_per_task_ms": 190922.58805031446,
      "avg_output_tokens_per_task": 3934.8223270440253,
      "last_run_timestamp": "2026-09-26T08:09:25.362Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.40055558333333335,
      "accuracy_std_dev": 0.02035940571929975,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.009198192499999995,
      "avg_duration_per_task_ms": 42365.11,
      "avg_output_tokens_per_task": 393.8333333333333,
      "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": "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": "AkashML",
      "accuracy": 0.39416666666666667,
      "accuracy_std_dev": 0.026056356301720738,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.008334941666666665,
      "avg_duration_per_task_ms": 21964.856666666667,
      "avg_output_tokens_per_task": 381.19166666666666,
      "last_run_timestamp": "2026-08-29T16:10:09.208Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "Inceptron",
      "accuracy": 0.3933333333333333,
      "accuracy_std_dev": 0.03197221015541812,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.004903370000000001,
      "avg_duration_per_task_ms": 11677.77,
      "avg_output_tokens_per_task": 233.46333333333334,
      "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": "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": "DeepInfra",
      "accuracy": 0.3860600450751252,
      "accuracy_std_dev": 0.0672991061241096,
      "run_count": 24,
      "total_questions": 1198,
      "avg_cost_per_task_usd": 0.03661595993322203,
      "avg_duration_per_task_ms": 170385.86227045074,
      "avg_output_tokens_per_task": 2372.5033388981637,
      "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": "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": 0.3856944583333333,
      "accuracy_std_dev": 0.07826521681329195,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.038991605833333345,
      "avg_duration_per_task_ms": 59134.56,
      "avg_output_tokens_per_task": 1840.6966666666667,
      "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": "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": 0.38527770833333336,
      "accuracy_std_dev": 0.033332128566743696,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.04909040916666666,
      "avg_duration_per_task_ms": 79084.84416666666,
      "avg_output_tokens_per_task": 2047.975,
      "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/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "auto-routing",
      "accuracy": 0.3837681304347826,
      "accuracy_std_dev": 0.032112499031163325,
      "run_count": 23,
      "total_questions": 1150,
      "avg_cost_per_task_usd": 0.016648859999999998,
      "avg_duration_per_task_ms": 124100.78260869565,
      "avg_output_tokens_per_task": 2798.5713043478263,
      "last_run_timestamp": "2026-08-29T16:10:20.675Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "meta-llama/llama-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.38358462814070354,
      "accuracy_std_dev": 0.035037700884760284,
      "run_count": 24,
      "total_questions": 1194,
      "avg_cost_per_task_usd": 0.013517370184254609,
      "avg_duration_per_task_ms": 48228.662479061975,
      "avg_output_tokens_per_task": 593.5661641541038,
      "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": "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": "WandB",
      "accuracy": 0.38313723529411764,
      "accuracy_std_dev": 0.04464047933477813,
      "run_count": 17,
      "total_questions": 850,
      "avg_cost_per_task_usd": 0.01614717176470588,
      "avg_duration_per_task_ms": 39209.35647058824,
      "avg_output_tokens_per_task": 2326.121176470588,
      "last_run_timestamp": "2026-07-18T08:03:14.950Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "AtlasCloud",
      "accuracy": 0.38311113333333335,
      "accuracy_std_dev": 0.05225957113597045,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.012468884,
      "avg_duration_per_task_ms": 66002.08133333334,
      "avg_output_tokens_per_task": 1901.1146666666666,
      "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": "deepseek/deepseek-chat-v3-0324",
      "model_slug": "deepseek/deepseek-chat-v3-0324",
      "model_name": "DeepSeek: DeepSeek V3 0324",
      "provider_name": "AtlasCloud",
      "accuracy": 0.38285714285714284,
      "accuracy_std_dev": 0.03194382824999699,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.01867379142857143,
      "avg_duration_per_task_ms": 97094.68714285715,
      "avg_output_tokens_per_task": 2232.7085714285713,
      "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/qwen3-30b-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": 0.3796296111111111,
      "accuracy_std_dev": 0.024492127175298337,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.03998439333333333,
      "avg_duration_per_task_ms": 13676.166666666666,
      "avg_output_tokens_per_task": 510.91333333333336,
      "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": "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": 0.3709090909090909,
      "accuracy_std_dev": 0.09668333303668551,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.04139647545454545,
      "avg_duration_per_task_ms": 141167.0809090909,
      "avg_output_tokens_per_task": 5048.265454545454,
      "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": 0.37,
      "accuracy_std_dev": 0.04898979485566354,
      "run_count": 8,
      "total_questions": 400,
      "avg_cost_per_task_usd": 0.04354687000000001,
      "avg_duration_per_task_ms": 64045.5575,
      "avg_output_tokens_per_task": 2587.9975,
      "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": 0.3628333,
      "accuracy_std_dev": 0.054805493135360074,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.019168134000000003,
      "avg_duration_per_task_ms": 43635.281,
      "avg_output_tokens_per_task": 2269.592,
      "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": 0.3574878985507246,
      "accuracy_std_dev": 0.02831418715169525,
      "run_count": 14,
      "total_questions": 690,
      "avg_cost_per_task_usd": 0.06476017536231883,
      "avg_duration_per_task_ms": 326809.4072463768,
      "avg_output_tokens_per_task": 6876.569565217392,
      "last_run_timestamp": "2026-09-26T08:14:46.867Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3.6-35b-a3b-20260415",
      "model_slug": "qwen/qwen3.6-35b-a3b",
      "model_name": "Qwen: Qwen3.6 35B A3B",
      "provider_name": "Venice",
      "accuracy": 0.3555554444444445,
      "accuracy_std_dev": 0.3294587300632867,
      "run_count": 9,
      "total_questions": 450,
      "avg_cost_per_task_usd": 0.021335997777777782,
      "avg_duration_per_task_ms": 65350.72222222222,
      "avg_output_tokens_per_task": 11586.551111111112,
      "last_run_timestamp": "2026-09-26T16:00:24.611Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_slug": "qwen/qwen3-30b-a3b-instruct-2507",
      "model_name": "Qwen: Qwen3 30B A3B Instruct 2507",
      "provider_name": "Nebius",
      "accuracy": 0.35151509090909094,
      "accuracy_std_dev": 0.04450312047374685,
      "run_count": 22,
      "total_questions": 1100,
      "avg_cost_per_task_usd": 0.023887789090909096,
      "avg_duration_per_task_ms": 76519.86636363636,
      "avg_output_tokens_per_task": 3035.681818181818,
      "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": 0.3461934413580247,
      "accuracy_std_dev": 0.030214013647418478,
      "run_count": 13,
      "total_questions": 648,
      "avg_cost_per_task_usd": 0.07313393055555555,
      "avg_duration_per_task_ms": 211770.98302469135,
      "avg_output_tokens_per_task": 6862.71450617284,
      "last_run_timestamp": "2026-09-26T08:01:16.588Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "auto-routing",
      "accuracy": 0.3268333,
      "accuracy_std_dev": 0.04588001392556459,
      "run_count": 20,
      "total_questions": 1000,
      "avg_cost_per_task_usd": 0.020283186000000005,
      "avg_duration_per_task_ms": 68109.949,
      "avg_output_tokens_per_task": 2751.024,
      "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-8b-instruct",
      "model_slug": "qwen/qwen3-vl-8b-instruct",
      "model_name": "Qwen: Qwen3 VL 8B Instruct",
      "provider_name": "Novita",
      "accuracy": 0.3257778,
      "accuracy_std_dev": 0.06723601430602502,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.010765525333333335,
      "avg_duration_per_task_ms": 44415.11066666667,
      "avg_output_tokens_per_task": 1680.8693333333333,
      "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-30b-a3b-instruct",
      "model_slug": "qwen/qwen3-vl-30b-a3b-instruct",
      "model_name": "Qwen: Qwen3 VL 30B A3B Instruct",
      "provider_name": "Alibaba",
      "accuracy": 0.3242857857142857,
      "accuracy_std_dev": 0.042134165636483804,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.07503298571428571,
      "avg_duration_per_task_ms": 131659.5957142857,
      "avg_output_tokens_per_task": 7119.688571428572,
      "last_run_timestamp": "2026-09-26T08:09:25.338Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "z-ai/glm-4.6",
      "model_slug": "z-ai/glm-4.6",
      "model_name": "Z.ai: GLM 4.6",
      "provider_name": "AtlasCloud",
      "accuracy": 0.31383302376599637,
      "accuracy_std_dev": 0.12215659515952029,
      "run_count": 22,
      "total_questions": 1094,
      "avg_cost_per_task_usd": 0.10114592687385739,
      "avg_duration_per_task_ms": 309255.3491773309,
      "avg_output_tokens_per_task": 6164.430530164534,
      "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": "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": 0.31037033333333336,
      "accuracy_std_dev": 0.06921873819598595,
      "run_count": 18,
      "total_questions": 900,
      "avg_cost_per_task_usd": 0.006170470000000001,
      "avg_duration_per_task_ms": 68229.11888888889,
      "avg_output_tokens_per_task": 1564.5266666666666,
      "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": 0.3,
      "accuracy_std_dev": 0.12837445228704972,
      "run_count": 5,
      "total_questions": 250,
      "avg_cost_per_task_usd": 0.168574564,
      "avg_duration_per_task_ms": 489750.116,
      "avg_output_tokens_per_task": 17487.74,
      "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": "Together",
      "accuracy": 0.293888875,
      "accuracy_std_dev": 0.03696429813269432,
      "run_count": 24,
      "total_questions": 1200,
      "avg_cost_per_task_usd": 0.10173790416666667,
      "avg_duration_per_task_ms": 65960.12166666667,
      "avg_output_tokens_per_task": 1260.8666666666666,
      "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-3.3-70b-instruct",
      "model_slug": "meta-llama/llama-3.3-70b-instruct",
      "model_name": "Meta: Llama 3.3 70B Instruct",
      "provider_name": "DeepInfra",
      "accuracy": 0.2926605816513761,
      "accuracy_std_dev": 0.03933661777666503,
      "run_count": 22,
      "total_questions": 1090,
      "avg_cost_per_task_usd": 0.012492933944954132,
      "avg_duration_per_task_ms": 163693.22110091744,
      "avg_output_tokens_per_task": 2119.8,
      "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": "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": 0.2904954257028112,
      "accuracy_std_dev": 0.19463944059054425,
      "run_count": 10,
      "total_questions": 498,
      "avg_cost_per_task_usd": 0.1360842630522088,
      "avg_duration_per_task_ms": 731552.421686747,
      "avg_output_tokens_per_task": 5033.198795180723,
      "last_run_timestamp": "2026-09-26T08:05:39.969Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "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": 0.28177779999999997,
      "accuracy_std_dev": 0.018129828637542793,
      "run_count": 15,
      "total_questions": 750,
      "avg_cost_per_task_usd": 0.0062486493333333325,
      "avg_duration_per_task_ms": 80386.89066666667,
      "avg_output_tokens_per_task": 1861.8466666666666,
      "last_run_timestamp": "2026-07-11T08:03:53.497Z"
    },
    {
      "benchmark": "tau_bench_verified_airline",
      "benchmark_page_url": "https://openrouter.ai/benchmarks/tau2-bench-airline",
      "model_permaslug": "openai/gpt-4o-mini",
      "model_slug": "openai/gpt-4o-mini",
      "model_name": "OpenAI: GPT-4o-mini",
      "provider_name": "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": "deepseek/deepseek-chat-v3.1",
      "model_slug": "deepseek/deepseek-chat-v3.1",
      "model_name": "DeepSeek: DeepSeek V3.1",
      "provider_name": "Chutes",
      "accuracy": 0.26,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.09888683999999999,
      "avg_duration_per_task_ms": 848196.66,
      "avg_output_tokens_per_task": 16087.38,
      "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": "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": 0.24666666666666667,
      "accuracy_std_dev": 0.029814239699997202,
      "run_count": 6,
      "total_questions": 300,
      "avg_cost_per_task_usd": 0.07892822666666666,
      "avg_duration_per_task_ms": 52263.34333333333,
      "avg_output_tokens_per_task": 10821.06,
      "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": "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": 0.2057142857142857,
      "accuracy_std_dev": 0.05038626311013829,
      "run_count": 14,
      "total_questions": 700,
      "avg_cost_per_task_usd": 0.08984258285714286,
      "avg_duration_per_task_ms": 35641.26,
      "avg_output_tokens_per_task": 1649.8585714285714,
      "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": "qwen/qwen3-32b-04-28",
      "model_slug": "qwen/qwen3-32b",
      "model_name": "Qwen: Qwen3 32B",
      "provider_name": "Chutes",
      "accuracy": 0.18,
      "accuracy_std_dev": null,
      "run_count": 1,
      "total_questions": 50,
      "avg_cost_per_task_usd": 0.015152460000000001,
      "avg_duration_per_task_ms": 1356163.7,
      "avg_output_tokens_per_task": 13773.38,
      "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": "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": "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": "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"
    }
  ]
}
