{
  "api_version": "1",
  "updated": "2026-11-10",
  "latest_event": "2026-11-09",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "swe-bench-pro",
    "name": "SWE-bench Pro",
    "family": "swe-bench",
    "unit": "%",
    "higher_is_better": true,
    "status": "active",
    "active": true,
    "superseded_by": null,
    "retired": null,
    "benchmark_url": "https://scale.com/leaderboard/swe_bench_pro_public",
    "introduced": "2025-09-19",
    "creator": "Scale AI",
    "introduced_source": "https://scale.com/blog/swe-bench-pro",
    "settings": [
      {
        "setting": "public",
        "name": "Public set (731 tasks)",
        "url": "https://frontierlog.ai/benchmark/swe-bench-pro",
        "results": 29
      },
      {
        "setting": "corrected",
        "name": "Lab's corrected public set",
        "url": "https://frontierlog.ai/benchmark/swe-bench-pro/corrected",
        "results": 10
      },
      {
        "setting": "unstated",
        "name": "Set not stated",
        "url": "https://frontierlog.ai/benchmark/swe-bench-pro/unstated",
        "results": 46
      }
    ]
  },
  "total": 10,
  "results": [
    {
      "id": "qwen3-5-397b-a17b:swe-bench-pro:corrected:1",
      "model": "qwen3-5-397b-a17b",
      "model_name": "Qwen3.5-397B-A17B",
      "model_date": "2026-02-16",
      "event": "https://frontierlog.ai/event/2026-02-16-qwen3-5",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 50.9,
      "unit": "%",
      "printed": "50.9",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.5-397B-A17B",
      "source": "https://qwen.ai/blog?id=qwen3.6",
      "source_title": "Qwen3.6-Plus: Towards Real World Agents",
      "source_date": "2026-04-02",
      "where": "Qwen3.6-Plus post, 'Performance' section, 'Language' table, row 'SWE-bench Pro', column 'Qwen3.5-397B-A17B'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "internal harness with bash and file-edit tools; 200K context",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-19",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/6f4855b376c670b4.png"
    },
    {
      "id": "qwen3-6-plus:swe-bench-pro:corrected:1",
      "model": "qwen3-6-plus",
      "model_name": "Qwen3.6-Plus",
      "model_date": "2026-04-02",
      "event": "https://frontierlog.ai/event/2026-04-02-qwen3-6-plus",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 56.6,
      "unit": "%",
      "printed": "56.6",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.6-Plus",
      "source": "https://qwen.ai/blog?id=qwen3.6",
      "source_title": "Qwen3.6-Plus: Towards Real World Agents",
      "source_date": "2026-04-02",
      "where": "'Performance' section, 'Language' table, row 'SWE-bench Pro', column 'Qwen3.6-Plus'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "internal harness with bash and file-edit tools; 200K context",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-19",
      "check": "human",
      "corroborated": true,
      "corroborated_on": "2026-10-19",
      "corroborated_by": [
        {
          "url": "https://z.ai/blog/glm-5.1",
          "name": "GLM-5.1",
          "title": "GLM-5.1: Towards Long-Horizon Tasks",
          "lab": "zai",
          "date": "2026-04-07",
          "printed": "56.6",
          "where": "Benchmark table, row 'SWE-Bench Pro', column 'Qwen3.6-Plus'"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/6f4855b376c670b4.png"
    },
    {
      "id": "qwen3-6-35b-a3b:swe-bench-pro:corrected:1",
      "model": "qwen3-6-35b-a3b",
      "model_name": "Qwen3.6-35B-A3B",
      "model_date": "2026-04-15",
      "event": "https://frontierlog.ai/event/2026-04-15-qwen3-6-35b-a3b",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 49.5,
      "unit": "%",
      "printed": "49.5",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.6-35BA3B",
      "source": "https://qwen.ai/blog?id=qwen3.6-35b-a3b",
      "source_title": "Qwen3.6-35B-A3B: Agentic Coding Power, Now Open to All",
      "source_date": "2026-04-15",
      "where": "Language table (first table), Coding Agent block, row 'SWE-bench Pro', column 'Qwen3.6-35BA3B'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "internal harness with bash and file-edit tools; 200K context",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-19",
      "check": "human",
      "corroborated": true,
      "corroborated_on": "2026-10-19",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/Qwen/Qwen3.6-35B-A3B",
          "name": "Qwen3.6-35B-A3B card",
          "title": "Qwen/Qwen3.6-35B-A3B model card",
          "lab": "alibaba",
          "date": "2026-04-15",
          "printed": "49.5",
          "where": "HF model card, language benchmark HTML table, row SWE-bench Pro, column Qwen3.6-35BA3B; footnote: internal agent scaffold (bash + file-edit tools), 200K context"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/29389a297e0321ef.png"
    },
    {
      "id": "qwen3-7-max:swe-bench-pro:corrected: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": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 60.6,
      "unit": "%",
      "printed": "60.6",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.7-Max",
      "source": "https://qwen.ai/blog?id=qwen3.7",
      "source_title": "Qwen3.7: The Agent Frontier",
      "source_date": "2026-05-20",
      "where": "'Performance' table, row 'SWE-Pro', column 'Qwen3.7-Max'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "internal harness with bash and file-edit tools; 200K context",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-19",
      "check": "human",
      "corroborated": true,
      "corroborated_on": "2026-10-19",
      "corroborated_by": [
        {
          "url": "https://z.ai/blog/glm-5.2",
          "name": "GLM-5.2",
          "title": "GLM-5.2: Built for Long-Horizon Tasks",
          "lab": "zai",
          "date": "2026-06-16",
          "printed": "60.6",
          "where": "Full Benchmark Table, row 'SWE-bench Pro', column 'Qwen3.7-Max'"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/9d2174c2d8ef254a.png"
    },
    {
      "id": "qwen3-7-plus:swe-bench-pro:corrected: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": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 57.6,
      "unit": "%",
      "printed": "57.6",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.7-Plus",
      "source": "https://qwen.ai/blog?id=qwen3.7-plus",
      "source_title": "Qwen3.7-Plus: Multimodal Agent Intelligence",
      "source_date": "2026-06-01",
      "where": "language benchmark table (first table), Coding Agent block, row 'SWE-Pro', column 'Qwen3.7-Plus'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "internal harness with bash and file-edit tools; 200K context",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-19",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/0b4fd6fbae7120cb.png"
    },
    {
      "id": "qwen3-7-plus:swe-bench-pro:corrected:2",
      "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": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 55.8,
      "unit": "%",
      "printed": "55.8",
      "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), row 'Agentic coding / SWE-bench Pro', column 'Qwen3.7-Plus'",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "Claude Code harness; 256K context",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-19",
      "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": "55.8",
          "where": "Qwen3.8-Flash-Next HF model card, HTML table, row Agentic coding / SWE-bench Pro, column Qwen3.7-Plus; footnote 2: Claude Code harness, 256K context"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/ce5ffcfe0edfb313.png"
    },
    {
      "id": "longcat-2-0:swe-bench-pro:corrected:1",
      "model": "longcat-2-0",
      "model_name": "LongCat-2.0",
      "model_date": "2026-06-30",
      "event": "https://frontierlog.ai/event/2026-06-30-longcat-2-0",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 59.5,
      "unit": "%",
      "printed": "59.5",
      "reported_by": "meituan",
      "run_by": null,
      "label": "LongCat-2.0",
      "source": "https://longcat.chat/blog/longcat-2.0/",
      "source_title": "Introducing LongCat-2.0",
      "source_date": "2026-06-30",
      "where": "'Evaluations' section table (rendered by the page's script; load it in a browser), row 'SWE-bench Pro' under 'CODE AGENT', column 'LongCat-2.0'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code harness",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-23",
      "check": "human",
      "corroborated": true,
      "corroborated_on": "2026-10-23",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/meituan-longcat/LongCat-2.0",
          "name": "LongCat-2.0 card",
          "title": "LongCat-2.0 model card (Hugging Face)",
          "lab": "meituan",
          "date": "2026-06-30",
          "printed": "59.5",
          "where": "Evaluation Results table, Code Agent section, row 'SWE-bench Pro', column 'LongCat-2.0' (cards say only 'measured in-house under a unified harness'; per-row harness/timeout not stated; card intro names Claude Code among supported harnesses)"
        },
        {
          "url": "https://github.com/meituan-longcat/LongCat-2.0",
          "name": "LongCat-2.0 card",
          "title": "LongCat-2.0 README (GitHub)",
          "lab": "meituan",
          "date": "2026-06-30",
          "printed": "59.5",
          "where": "README Evaluation Results table, Code Agent section, row 'SWE-bench Pro', column 'LongCat-2.0' (cards say only 'measured in-house under a unified harness'; per-row harness/timeout not stated; card intro names Claude Code among supported harnesses)"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/958070cf1766c220.png"
    },
    {
      "id": "qwen3-8-max:swe-bench-pro:corrected: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": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 67.7,
      "unit": "%",
      "printed": "67.7",
      "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), row 'SWE-bench Pro' under 'Coding Agent', column 'Qwen3.8-Max' (read via the qwen.ai article API, qwen.ai/api/v2/article/?language=en-US&path=qwen3.8&type=qwen_ai, article 65dfab20-97fb-4932-97cd-4f7da4ab51f5)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code harness",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-17",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/10d02db2e5d51286.png"
    },
    {
      "id": "qwen3-8-flash-next:swe-bench-pro:corrected: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": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 62.5,
      "unit": "%",
      "printed": "62.5",
      "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), row 'Agentic coding / SWE-bench Pro', column 'Qwen3.8-Flash-Next'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code harness; 256K context",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-19",
      "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": "62.5",
          "where": "README Language table, row 'Agentic coding / SWE-bench Pro', column 'Qwen3.8-Flash-Next'"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/ce5ffcfe0edfb313.png"
    },
    {
      "id": "qwen3-8-omni-flash:swe-bench-pro:corrected:1",
      "model": "qwen3-8-omni-flash",
      "model_name": "Qwen3.8-Omni-Flash",
      "model_date": "2026-09-18",
      "event": "https://frontierlog.ai/event/2026-09-18-qwen3-8-omni-flash",
      "benchmark": "swe-bench-pro",
      "benchmark_name": "SWE-bench Pro",
      "setting": "corrected",
      "setting_name": "Lab's corrected public set",
      "score": 63.3,
      "unit": "%",
      "printed": "63.3",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.8-Omni-Flash",
      "source": "https://qwen.ai/blog?id=qwen3.8-omni-flash",
      "source_title": "Qwen3.8-Omni-Flash: Omni Senses. Agentic Delivery.",
      "source_date": "2026-09-18",
      "where": "text benchmark table (Coding and Agent block), row 'SWE-bench Pro', column 'Qwen3.8-Omni-Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Claude Code harness; 256K context",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-20",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/404338e9a04ccbae.png"
    }
  ]
}