{
  "api_version": "1",
  "updated": "2026-10-29",
  "latest_event": "2026-10-28",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "agents-last-exam",
    "name": "Agents' Last Exam v1",
    "family": "agents-last-exam",
    "unit": "%",
    "higher_is_better": true,
    "status": "superseded",
    "active": false,
    "superseded_by": "agents-last-exam-1-1",
    "retired": null,
    "benchmark_url": "https://agents-last-exam.org",
    "introduced": "2026-06-03",
    "creator": "UC Berkeley RDI",
    "introduced_source": "https://arxiv.org/abs/2606.05405v1",
    "settings": [
      {
        "setting": "pass",
        "name": "Pass rate",
        "url": "https://frontierlog.ai/benchmark/agents-last-exam",
        "results": 8
      },
      {
        "setting": "score",
        "name": "Score",
        "url": "https://frontierlog.ai/benchmark/agents-last-exam/score",
        "results": 4
      },
      {
        "setting": "cli-pass",
        "name": "CLI subset, pass rate",
        "url": "https://frontierlog.ai/benchmark/agents-last-exam/cli-pass",
        "results": 8
      },
      {
        "setting": "unstated",
        "name": "Measure not stated",
        "url": "https://frontierlog.ai/benchmark/agents-last-exam/unstated",
        "results": 8
      }
    ]
  },
  "total": 4,
  "results": [
    {
      "id": "qwen3-7-max:agents-last-exam:score:1",
      "model": "qwen3-7-max",
      "model_name": "Qwen3.7-Max",
      "model_date": "2026-05-20",
      "event": "https://frontierlog.ai/event/2026-05-20-qwen3-7-max",
      "benchmark": "agents-last-exam",
      "benchmark_name": "Agents' Last Exam v1",
      "setting": "score",
      "setting_name": "Score",
      "score": 31.1,
      "unit": "%",
      "printed": "31.1",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.7-Max",
      "source": "https://qwen.ai/blog?id=qwen3.8",
      "source_title": "Qwen3.8-Max: A New Bar for Coding and Cowork",
      "source_date": "2026-08-03",
      "where": "later report: Qwen3.8-Max post, 'Full Benchmark Table' (first table), General Agent block, row 'Agents' Last Exam (Pass / Score)', column 'Qwen3.7-Max', second of the two values",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "harness and subset not stated",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    },
    {
      "id": "qwen3-7-plus:agents-last-exam:score:1",
      "model": "qwen3-7-plus",
      "model_name": "Qwen3.7-Plus",
      "model_date": "2026-06-01",
      "event": "https://frontierlog.ai/event/2026-06-01-qwen3-7-plus",
      "benchmark": "agents-last-exam",
      "benchmark_name": "Agents' Last Exam v1",
      "setting": "score",
      "setting_name": "Score",
      "score": 33.6,
      "unit": "%",
      "printed": "33.6",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.7-Plus",
      "source": "https://qwen.ai/blog?id=qwen3.8-flash-next",
      "source_title": "Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-Efficiency",
      "source_date": "2026-08-26",
      "where": "later report: Qwen3.8-Flash-Next post, Language table (first table), Agent block, row 'Agents' Last Exam', column 'Qwen3.7-Plus', the value labelled 'Score'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "harness and subset not stated",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-29",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/Qwen/Qwen3.8-Flash-Next",
          "name": "Qwen3.8-Flash-Next card",
          "title": "Qwen/Qwen3.8-Flash-Next model card",
          "lab": "alibaba",
          "date": "2026-08-26",
          "printed": "33.6",
          "where": "model card benchmark table, row \"Frontier agentic tasks / Agents' Last Exam\", column \"Qwen3.7-Plus\", sub-value labelled \"Score\""
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/1ce175beb0b20cf1.png"
    },
    {
      "id": "qwen3-8-max:agents-last-exam:score:1",
      "model": "qwen3-8-max",
      "model_name": "Qwen3.8-Max",
      "model_date": "2026-08-03",
      "event": "https://frontierlog.ai/event/2026-08-03-qwen3-8-max",
      "benchmark": "agents-last-exam",
      "benchmark_name": "Agents' Last Exam v1",
      "setting": "score",
      "setting_name": "Score",
      "score": 52.4,
      "unit": "%",
      "printed": "52.4",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.8-Max",
      "source": "https://qwen.ai/blog?id=qwen3.8",
      "source_title": "Qwen3.8-Max: A New Bar for Coding and Cowork",
      "source_date": "2026-08-03",
      "where": "'Full Benchmark Table' (first table), General Agent block, row 'Agents' Last Exam (Pass / Score)', column 'Qwen3.8-Max' (last column), second of the two values",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "harness and subset not stated",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    },
    {
      "id": "qwen3-8-flash-next:agents-last-exam:score:1",
      "model": "qwen3-8-flash-next",
      "model_name": "Qwen3.8-Flash-Next",
      "model_date": "2026-08-26",
      "event": "https://frontierlog.ai/event/2026-08-26-qwen3-8-flash-next",
      "benchmark": "agents-last-exam",
      "benchmark_name": "Agents' Last Exam v1",
      "setting": "score",
      "setting_name": "Score",
      "score": 51.2,
      "unit": "%",
      "printed": "51.2",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.8-Flash-Next",
      "source": "https://qwen.ai/blog?id=qwen3.8-flash-next",
      "source_title": "Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-Efficiency",
      "source_date": "2026-08-26",
      "where": "Language table (first table), Agent block, row 'Frontier agentic tasks / Agents' Last Exam', column 'Qwen3.8-Flash-Next', the value labelled 'Score'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "harness and subset not stated",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-29",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/Qwen/Qwen3.8-Flash-Next",
          "name": "Qwen3.8-Flash-Next card",
          "title": "Qwen/Qwen3.8-Flash-Next model card",
          "lab": "alibaba",
          "date": "2026-08-26",
          "printed": "51.2",
          "where": "model card benchmark table, row \"Frontier agentic tasks / Agents' Last Exam\", column \"Qwen3.8-Flash-Next\", sub-value labelled \"Score\""
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/1ce175beb0b20cf1.png"
    }
  ]
}