{
  "api_version": "1",
  "updated": "2026-10-30",
  "latest_event": "2026-10-29",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "ifeval",
    "name": "IFEval",
    "family": "ifeval",
    "unit": "%",
    "higher_is_better": true,
    "status": "retired",
    "active": false,
    "superseded_by": null,
    "retired": "2026-06-01",
    "benchmark_url": "https://github.com/google-research/google-research/tree/master/instruction_following_eval",
    "introduced": "2023-11-14",
    "creator": "Google and Yale University",
    "introduced_source": "https://arxiv.org/abs/2311.07911",
    "settings": [
      {
        "setting": "prompt-strict",
        "name": "Prompt-level strict",
        "url": "https://frontierlog.ai/benchmark/ifeval",
        "results": 3
      },
      {
        "setting": "unstated",
        "name": "Metric not stated",
        "url": "https://frontierlog.ai/benchmark/ifeval/unstated",
        "results": 10
      }
    ]
  },
  "total": 10,
  "results": [
    {
      "id": "qwen3-235b-a22b:ifeval:unstated:1",
      "model": "qwen3-235b-a22b",
      "model_name": "Qwen3-235B-A22B",
      "model_date": "2025-04-29",
      "event": "https://frontierlog.ai/event/2025-04-29-qwen3",
      "benchmark": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 83.4,
      "unit": "%",
      "printed": "83.4",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3-235B-A22B Thinking",
      "source": "https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507/blob/4cd68849b2a0fadb84866c703b133aa8c8636130/README.md",
      "source_title": "Qwen/Qwen3-235B-A22B-Thinking-2507 model card (revision of 2025-07-25)",
      "source_date": "2025-07-25",
      "where": "later report: Qwen3-235B-A22B-Thinking-2507 model card 'Performance' table, Alignment block, row 'IFEval', column 'Qwen3-235B-A22B Thinking'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "thinking mode",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/3d049bc43f1609a6.png"
    },
    {
      "id": "mistral-medium-3:ifeval:unstated:1",
      "model": "mistral-medium-3",
      "model_name": "Mistral Medium 3",
      "model_date": "2025-05-07",
      "event": "https://frontierlog.ai/event/2025-05-07-mistral-medium-3",
      "benchmark": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 89.4,
      "unit": "%",
      "printed": "89.4%",
      "reported_by": "mistral",
      "run_by": null,
      "label": "Mistral Medium 3",
      "source": "https://mistral.ai/news/mistral-medium-3/",
      "source_title": "Medium is the new large.",
      "source_date": "2025-05-07",
      "where": "benchmark table image ('Table Medium 6'), section INSTRUCTION FOLLOWING, row 'IfEval 0-shot', column 'Mistral Medium 3'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "0-shot",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/da41e4f28d4c6435.webp",
      "source_image_original": "https://mistral.ai/_astro/1a845a6e-34b8-4a0e-83d8-25a1436241a9_ZfoqEG.webp"
    },
    {
      "id": "qwen3-235b-a22b-thinking-2507:ifeval:unstated:1",
      "model": "qwen3-235b-a22b-thinking-2507",
      "model_name": "Qwen3-235B-A22B-Thinking-2507",
      "model_date": "2025-07-25",
      "event": "https://frontierlog.ai/event/2025-07-25-qwen3-235b-a22b-thinking-2507",
      "benchmark": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 87.8,
      "unit": "%",
      "printed": "87.8",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3-235B-A22B-Thinking-2507",
      "source": "https://huggingface.co/Qwen/Qwen3-235B-A22B-Thinking-2507/blob/4cd68849b2a0fadb84866c703b133aa8c8636130/README.md",
      "source_title": "Qwen/Qwen3-235B-A22B-Thinking-2507 model card (revision of 2025-07-25)",
      "source_date": "2025-07-25",
      "where": "model card 'Performance' table, Alignment block, row 'IFEval', column 'Qwen3-235B-A22B-Thinking-2507'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/3d049bc43f1609a6.png"
    },
    {
      "id": "qwen3-next:ifeval:unstated:1",
      "model": "qwen3-next",
      "model_name": "Qwen3-Next",
      "model_date": "2025-09-11",
      "event": "https://frontierlog.ai/event/2025-09-11-qwen3-next",
      "benchmark": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 88.9,
      "unit": "%",
      "printed": "88.9",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3-Next-80B-A3B-Thinking",
      "source": "https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Thinking/blob/3e64fd191eaaf4b3e2ccb5519500a0ae515c3e25/README.md",
      "source_title": "Qwen/Qwen3-Next-80B-A3B-Thinking model card (launch-day revision)",
      "source_date": "2025-09-11",
      "where": "model card 'Performance' table, Alignment block, row 'IFEval', column 'Qwen3-Next-80B-A3B-Thinking'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "thinking model",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/5d16cbb759bbfc69.png"
    },
    {
      "id": "qwen3-max-thinking:ifeval:unstated:1",
      "model": "qwen3-max-thinking",
      "model_name": "Qwen3-Max-Thinking",
      "model_date": "2026-01-26",
      "event": "https://frontierlog.ai/event/2026-01-26-qwen3-max-thinking",
      "benchmark": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 93.4,
      "unit": "%",
      "printed": "93.4",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3-Max-Thinking",
      "source": "https://qwen.ai/blog?id=qwen3.5",
      "source_title": "Qwen3.5: Towards Native Multimodal Agents",
      "source_date": "2026-02-16",
      "where": "later report: Qwen3.5 post, 'Language' table, Instruction Following block, row 'IFEval', column 'Qwen3-Max-Thinking'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/Qwen/Qwen3.5-397B-A17B/blob/7cad2bae11cb49ca79f7d6a0954de2e2756f4e27/README.md",
          "name": "Qwen3.5 card",
          "title": "Qwen/Qwen3.5-397B-A17B model card (launch-day revision)",
          "lab": "alibaba",
          "date": "2026-02-16",
          "printed": "93.4",
          "where": "model card benchmark table (HTML), Instruction Following, row IFEval, column Qwen3-Max-Thinking"
        }
      ]
    },
    {
      "id": "qwen3-5-397b-a17b:ifeval:unstated: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": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 92.6,
      "unit": "%",
      "printed": "92.6",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.5-397B-A17B",
      "source": "https://qwen.ai/blog?id=qwen3.5",
      "source_title": "Qwen3.5: Towards Native Multimodal Agents",
      "source_date": "2026-02-16",
      "where": "'Performance' section, 'Language' table, Instruction Following block, row 'IFEval', column 'Qwen3.5-397B-A17B'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/Qwen/Qwen3.5-397B-A17B/blob/7cad2bae11cb49ca79f7d6a0954de2e2756f4e27/README.md",
          "name": "Qwen3.5 card",
          "title": "Qwen/Qwen3.5-397B-A17B model card (launch-day revision)",
          "lab": "alibaba",
          "date": "2026-02-16",
          "printed": "92.6",
          "where": "model card benchmark table (HTML), Instruction Following, row IFEval, column Qwen3.5-397B-A17B"
        }
      ]
    },
    {
      "id": "qwen3-5-omni:ifeval:unstated:1",
      "model": "qwen3-5-omni",
      "model_name": "Qwen3.5-Omni",
      "model_date": "2026-03-30",
      "event": "https://frontierlog.ai/event/2026-03-30-qwen3-5-omni",
      "benchmark": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 89.7,
      "unit": "%",
      "printed": "89.7",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.5-Omni-Plus",
      "source": "https://qwen.ai/blog?id=qwen3.5-omni",
      "source_title": "Qwen3.5-Omni: Scaling Up, Toward Native Omni-Modal AGI",
      "source_date": "2026-03-30",
      "where": "'Text' HTML table, Instruction Following block, row 'IFEval', column 'Qwen3.5-Omni-Plus'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Qwen3.5-Omni-Plus",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/f1a71324f95519d3.png"
    },
    {
      "id": "qwen3-5-omni-flash:ifeval:unstated:1",
      "model": "qwen3-5-omni-flash",
      "model_name": "Qwen3.5-Omni-Flash",
      "model_date": "2026-03-30",
      "event": "https://frontierlog.ai/event/2026-03-30-qwen3-5-omni",
      "benchmark": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 85.2,
      "unit": "%",
      "printed": "85.2",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3.5-Omni-Flash",
      "source": "https://qwen.ai/blog?id=qwen3.5-omni",
      "source_title": "Qwen3.5-Omni: Scaling Up, Toward Native Omni-Modal AGI",
      "source_date": "2026-03-30",
      "where": "'Text' HTML table, Instruction Following block, row 'IFEval', column 'Qwen3.5-Omni-Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Qwen3.5-Omni-Flash",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/f1a71324f95519d3.png"
    },
    {
      "id": "qwen3-7-max:ifeval:unstated: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": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 94.3,
      "unit": "%",
      "printed": "94.3",
      "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 HTML table, General Capability block, row 'IFEval', column 'Qwen3.7-Max'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    },
    {
      "id": "qwen3-7-plus:ifeval:unstated: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": "ifeval",
      "benchmark_name": "IFEval",
      "setting": "unstated",
      "setting_name": "Metric not stated",
      "score": 94.6,
      "unit": "%",
      "printed": "94.6",
      "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": "'Text Benchmarks' HTML table, General Capability block, row 'IFEval', column 'Qwen3.7-Plus'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "double_entry",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/de169a5a059c61f9.png"
    }
  ]
}