{
  "api_version": "1",
  "updated": "2026-11-10",
  "latest_event": "2026-11-09",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "tau2-bench",
    "name": "τ²-bench",
    "family": "tau-bench",
    "unit": "%",
    "higher_is_better": true,
    "status": "retired",
    "active": false,
    "superseded_by": null,
    "retired": "2026-04-23",
    "benchmark_url": "https://github.com/sierra-research/tau2-bench",
    "introduced": "2025-06-09",
    "creator": "Sierra",
    "introduced_source": "https://arxiv.org/abs/2506.07982",
    "settings": [
      {
        "setting": "telecom",
        "name": "Telecom",
        "url": "https://frontierlog.ai/benchmark/tau2-bench",
        "results": 42
      },
      {
        "setting": "retail",
        "name": "Retail, set not stated",
        "url": "https://frontierlog.ai/benchmark/tau2-bench/retail",
        "results": 31
      },
      {
        "setting": "airline",
        "name": "Airline, set not stated",
        "url": "https://frontierlog.ai/benchmark/tau2-bench/airline",
        "results": 26
      },
      {
        "setting": "airline-anthropic-corrected",
        "name": "Airline, Anthropic's corrected set",
        "url": "https://frontierlog.ai/benchmark/tau2-bench/airline-anthropic-corrected",
        "results": 2
      }
    ]
  },
  "total": 2,
  "results": [
    {
      "id": "claude-opus-4-5:tau2-bench:airline-anthropic-corrected:1",
      "model": "claude-opus-4-5",
      "model_name": "Claude Opus 4.5",
      "model_date": "2025-11-24",
      "event": "https://frontierlog.ai/event/2025-11-24-claude-opus-4-5",
      "benchmark": "tau2-bench",
      "benchmark_name": "τ²-bench",
      "setting": "airline-anthropic-corrected",
      "setting_name": "Airline, Anthropic's corrected set",
      "score": 87.8,
      "unit": "%",
      "printed": "87.8%",
      "reported_by": "anthropic",
      "run_by": null,
      "label": "Claude Opus 4.5",
      "source": "https://www.anthropic.com/claude-opus-4-5-system-card",
      "source_title": "Claude Opus 4.5 System Card",
      "source_date": "2025-11-24",
      "where": "Table 2.8.A (p.26), row 'Airline (corrected)', column 'Claude Opus 4.5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "prompt addendum to the policy; Opus 4.1 simulates the user",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    },
    {
      "id": "longcat-flash-thinking-2601:tau2-bench:airline-anthropic-corrected:1",
      "model": "longcat-flash-thinking-2601",
      "model_name": "LongCat-Flash-Thinking-2601",
      "model_date": "2026-01-20",
      "event": "https://frontierlog.ai/event/2026-01-20-longcat-flash-thinking-2601",
      "benchmark": "tau2-bench",
      "benchmark_name": "τ²-bench",
      "setting": "airline-anthropic-corrected",
      "setting_name": "Airline, Anthropic's corrected set",
      "score": 76.5,
      "unit": "%",
      "printed": "76.5",
      "reported_by": "meituan",
      "run_by": null,
      "label": "LongCat-Flash-Thinking-2601",
      "source": "https://huggingface.co/meituan-longcat/LongCat-Flash-Thinking-2601",
      "source_title": "LongCat-Flash-Thinking-2601 model card (Hugging Face)",
      "source_date": "2026-01-20",
      "where": "model card, Evaluation Results table, row 'τ²-Airline (Avg@4)', column 'LongCat-Flash-Thinking-2601'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "average of 4 runs; airline environment fixes from the Claude Opus 4.5 report",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-28",
      "corroborated_by": [
        {
          "url": "https://github.com/meituan-longcat/LongCat-Flash-Thinking-2601",
          "name": "LongCat-Flash-Thinking-2601 card",
          "title": "LongCat-Flash-Thinking-2601 GitHub repository",
          "lab": "meituan",
          "date": "2026-01-20",
          "printed": "76.5",
          "where": "GitHub README, τ²-Airline (Avg@4); note: airline environment fixes from the Claude Opus 4.5 report row, LongCat-Flash-Thinking-2601 column"
        },
        {
          "url": "https://github.com/meituan-longcat/LongCat-Flash-Thinking-2601/blob/main/LongCat_Flash_Thinking_2601_Technical_Report.pdf",
          "name": "LongCat-Flash-Thinking-2601 card",
          "title": "LongCat-Flash-Thinking-2601 Technical Report",
          "lab": "meituan",
          "date": "2026-01-20",
          "printed": "76.5",
          "where": "Technical Report (PDF in GitHub repo), main evaluation table, τ²-Airline (Avg@4); note: airline environment fixes from the Claude Opus 4.5 report row, LongCat-Flash-Thinking-2601 column"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/5d8fd20342a87cc4.png"
    }
  ]
}