{
  "api_version": "1",
  "updated": "2026-10-30",
  "latest_event": "2026-10-29",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "ifbench",
    "name": "IFBench",
    "family": "ifbench",
    "unit": "%",
    "higher_is_better": true,
    "status": "active",
    "active": true,
    "superseded_by": null,
    "retired": null,
    "benchmark_url": "https://github.com/allenai/IFBench",
    "introduced": "2025-07-03",
    "creator": "Ai2 and University of Washington",
    "introduced_source": "https://arxiv.org/abs/2507.02833",
    "settings": [
      {
        "setting": "default",
        "name": "",
        "url": "https://frontierlog.ai/benchmark/ifbench",
        "results": 18
      }
    ]
  },
  "total": 18,
  "results": [
    {
      "id": "minimax-m2:ifbench:default:1",
      "model": "minimax-m2",
      "model_name": "MiniMax M2",
      "model_date": "2025-10-27",
      "event": "https://frontierlog.ai/event/2025-10-27-minimax-m2",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 72.0,
      "unit": "%",
      "printed": "72.0",
      "reported_by": "minimax",
      "run_by": null,
      "label": "MiniMax-M2",
      "source": "https://www.minimax.io/news/minimax-m21",
      "source_title": "MiniMax M2.1: Significantly Enhanced Multi-Language Programming, Built for Real-World Complex Tasks",
      "source_date": "2025-12-23",
      "where": "later report: MiniMax-M2.1 post, 'Benchmarks' section, last (long-horizon tool use and intelligence) table image, row 'IFBench', column 'MiniMax-M2'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "internal run following Artificial Analysis Intelligence Index methods",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/MiniMaxAI/MiniMax-M2.1",
          "name": "MiniMax M2.1 card",
          "title": "MiniMaxAI/MiniMax-M2.1 model card",
          "lab": "minimax",
          "date": "2025-12-23",
          "printed": "72.0",
          "where": "Model card benchmark table, IFBench row, MiniMax-M2 column"
        },
        {
          "url": "https://github.com/MiniMax-AI/MiniMax-M2.1",
          "name": "MiniMax M2.1 card",
          "title": "MiniMax-M2.1 GitHub repository",
          "lab": "minimax",
          "date": "2025-12-23",
          "printed": "72.0",
          "where": "README benchmark table, IFBench row, MiniMax-M2 column"
        }
      ],
      "source_image": "https://frontierlog.ai/sources/78a04f05b9d1589b.png",
      "source_image_original": "https://file.cdn.minimax.io/public/6aaadb26-4395-46ba-9ebe-f858f3bc0c6b.png"
    },
    {
      "id": "mistral-large-3:ifbench:default:1",
      "model": "mistral-large-3",
      "model_name": "Mistral Large 3 675B",
      "model_date": "2025-12-02",
      "event": "https://frontierlog.ai/event/2025-12-02-mistral-3",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 37.8,
      "unit": "%",
      "printed": "37.8",
      "reported_by": "mistral",
      "run_by": null,
      "label": "Mistral Large 3",
      "source": "https://mistral.ai/news/mistral-small-4/",
      "source_title": "Introducing Mistral Small 4",
      "source_date": "2026-03-16",
      "where": "later report: Mistral Small 4 post, chart image 'Performance comparison across internal models' (first), group 'AllenAI IFBench', bar 'Mistral Large 3', value label",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "instruct model",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/4e38c748d409c83e.webp",
      "source_image_original": "https://mistral.ai/_astro/7d5b181d-776d-4406-aabc-b88414edf567_1u05nc.webp"
    },
    {
      "id": "minimax-m2-1:ifbench:default:1",
      "model": "minimax-m2-1",
      "model_name": "MiniMax M2.1",
      "model_date": "2025-12-23",
      "event": "https://frontierlog.ai/event/2025-12-23-minimax-m2-1",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 70.0,
      "unit": "%",
      "printed": "70.0",
      "reported_by": "minimax",
      "run_by": null,
      "label": "MiniMax-M2.1",
      "source": "https://www.minimax.io/news/minimax-m21",
      "source_title": "MiniMax M2.1: Significantly Enhanced Multi-Language Programming, Built for Real-World Complex Tasks",
      "source_date": "2025-12-23",
      "where": "'Benchmarks' section, last (long-horizon tool use and intelligence) table image, row 'IFBench', column 'MiniMax-M2.1'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "internal run following Artificial Analysis Intelligence Index methods",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/MiniMaxAI/MiniMax-M2.1",
          "name": "MiniMax M2.1 card",
          "title": "MiniMaxAI/MiniMax-M2.1 model card",
          "lab": "minimax",
          "date": "2025-12-23",
          "printed": "70.0",
          "where": "Model card benchmark table, IFBench row, MiniMax-M2.1 column"
        },
        {
          "url": "https://www.minimax.io/news/minimax-m25",
          "name": "MiniMax M2.5",
          "title": "MiniMax M2.5: Built for Real-World Productivity.",
          "lab": "minimax",
          "date": "2026-02-12",
          "printed": "70.0",
          "where": "Appendix table image (AIME25 ~ AA-LCR, AA Intelligence Index methods), IFBench row, MiniMax-M2.1 column"
        },
        {
          "url": "https://github.com/MiniMax-AI/MiniMax-M2.1",
          "name": "MiniMax M2.1 card",
          "title": "MiniMax-M2.1 GitHub repository",
          "lab": "minimax",
          "date": "2025-12-23",
          "printed": "70.0",
          "where": "README benchmark table (vs M2 and others), IFBench row, MiniMax-M2.1 column"
        }
      ],
      "source_image": "https://frontierlog.ai/sources/78a04f05b9d1589b.png",
      "source_image_original": "https://file.cdn.minimax.io/public/6aaadb26-4395-46ba-9ebe-f858f3bc0c6b.png"
    },
    {
      "id": "qwen3-max-thinking:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 70.9,
      "unit": "%",
      "printed": "70.9",
      "reported_by": "alibaba",
      "run_by": null,
      "label": "Qwen3-Max-Thinking",
      "source": "https://qwen.ai/blog?id=qwen3-max-thinking",
      "source_title": "Pushing Qwen3-Max-Thinking Beyond its Limits",
      "source_date": "2026-01-26",
      "where": "evaluation table, Instruction Following & Alignment block, row 'IFBench', column 'Qwen3-Max-Thinking'",
      "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": "70.9",
          "where": "Benchmark Results > Language table, Instruction Following > IFBench row, Qwen3-Max-Thinking column"
        },
        {
          "url": "https://qwen.ai/blog?id=qwen3.5",
          "name": "Qwen3.5",
          "title": "Qwen3.5: Towards Native Multimodal Agents",
          "lab": "alibaba",
          "date": "2026-02-16",
          "printed": "70.9",
          "where": "Language table, Instruction Following > IFBench row, Qwen3-Max-Thinking column"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/aed8db26403e0575.png"
    },
    {
      "id": "minimax-m2-5:ifbench:default:1",
      "model": "minimax-m2-5",
      "model_name": "MiniMax M2.5",
      "model_date": "2026-02-12",
      "event": "https://frontierlog.ai/event/2026-02-12-minimax-m2-5",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 70.0,
      "unit": "%",
      "printed": "70.0",
      "reported_by": "minimax",
      "run_by": null,
      "label": "MiniMax-M2.5",
      "source": "https://www.minimax.io/news/minimax-m25",
      "source_title": "MiniMax M2.5: Built for Real-World Productivity.",
      "source_date": "2026-02-12",
      "where": "Appendix 'Further benchmark results' table image, row 'IFBench', column 'MiniMax-M2.5'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "internal run following Artificial Analysis Intelligence Index methods",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/5af8b315069f9686.png",
      "source_image_original": "https://file.cdn.minimax.io/public/8c019213-b0d5-4ee8-9273-6d9b799abeae.png"
    },
    {
      "id": "qwen3-5-397b-a17b:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 76.5,
      "unit": "%",
      "printed": "76.5",
      "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 'IFBench', 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": "76.5",
          "where": "Benchmark Results > Language table, Instruction Following > IFBench row, Qwen3.5-397B-A17B column"
        },
        {
          "url": "https://mistral.ai/news/vibe-remote-agents-mistral-medium-3-5/",
          "name": "Mistral Medium 3.5",
          "title": "Remote agents in Vibe. Powered by Mistral Medium 3.5.",
          "lab": "mistral",
          "date": "2026-05-22",
          "printed": "76.5",
          "where": "Image 'Math, instruction following vs competing models', Allenai Ifbench group, Qwen3.5 397B A17B bar (Mistral's table)"
        },
        {
          "url": "https://qwen.ai/blog?id=qwen3.6",
          "name": "Qwen3.6-Plus",
          "title": "Qwen3.6-Plus: Towards Real World Agents",
          "lab": "alibaba",
          "date": "2026-04-02",
          "printed": "76.5",
          "where": "Performance table, Instruction Following > IFBench row, Qwen3.5-397B-A17B column"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/1008bfdc99666f33.png"
    },
    {
      "id": "mistral-small-4:ifbench:default:1",
      "model": "mistral-small-4",
      "model_name": "Mistral Small 4 119B",
      "model_date": "2026-03-16",
      "event": "https://frontierlog.ai/event/2026-03-16-mistral-small-4",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 48.0,
      "unit": "%",
      "printed": "48",
      "reported_by": "mistral",
      "run_by": null,
      "label": "Mistral Small 4",
      "source": "https://mistral.ai/news/mistral-small-4/",
      "source_title": "Introducing Mistral Small 4",
      "source_date": "2026-03-16",
      "where": "chart image 'Performance comparison across internal models' (first), group 'AllenAI IFBench', bar 'Mistral Small 4', total (reasoning) value label above the bar",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "reasoning mode",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/4e38c748d409c83e.webp",
      "source_image_original": "https://mistral.ai/_astro/7d5b181d-776d-4406-aabc-b88414edf567_1u05nc.webp"
    },
    {
      "id": "mistral-small-4:ifbench:default:2",
      "model": "mistral-small-4",
      "model_name": "Mistral Small 4 119B",
      "model_date": "2026-03-16",
      "event": "https://frontierlog.ai/event/2026-03-16-mistral-small-4",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 35.7,
      "unit": "%",
      "printed": "35.7",
      "reported_by": "mistral",
      "run_by": null,
      "label": "Mistral Small 4",
      "source": "https://mistral.ai/news/mistral-small-4/",
      "source_title": "Introducing Mistral Small 4",
      "source_date": "2026-03-16",
      "where": "chart image 'Performance comparison across internal models' (first), group 'AllenAI IFBench', bar 'Mistral Small 4', instruct value label inside the solid part of the bar",
      "own": true,
      "primary": false,
      "parallel": false,
      "notes": "instruct mode, no reasoning",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/4e38c748d409c83e.webp",
      "source_image_original": "https://mistral.ai/_astro/7d5b181d-776d-4406-aabc-b88414edf567_1u05nc.webp"
    },
    {
      "id": "qwen3-5-omni:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 52.6,
      "unit": "%",
      "printed": "52.6",
      "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 'IFBench', column 'Qwen3.5-Omni-Plus'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Qwen3.5-Omni-Plus",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/f1a71324f95519d3.png"
    },
    {
      "id": "qwen3-5-omni-flash:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 38.4,
      "unit": "%",
      "printed": "38.4",
      "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 'IFBench', column 'Qwen3.5-Omni-Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "Qwen3.5-Omni-Flash",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/f1a71324f95519d3.png"
    },
    {
      "id": "qwen3-6-plus:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 74.2,
      "unit": "%",
      "printed": "74.2",
      "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, Instruction Following block, row 'IFBench', column 'Qwen3.6-Plus'",
      "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://qwen.ai/blog?id=qwen3.7",
          "name": "Qwen3.7-Max",
          "title": "Qwen3.7: The Agent Frontier",
          "lab": "alibaba",
          "date": "2026-05-20",
          "printed": "74.2",
          "where": "Performance table, General Capability > IFBench row, Qwen3.6-Plus column"
        },
        {
          "url": "https://qwen.ai/blog?id=qwen3.7-plus",
          "name": "Qwen3.7-Plus",
          "title": "Qwen3.7-Plus: Multimodal Agent Intelligence",
          "lab": "alibaba",
          "date": "2026-06-01",
          "printed": "74.2",
          "where": "Text Benchmarks table, General Capability > IFBench row, Qwen3.6-Plus column"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/4c1f8a6da26570d9.png"
    },
    {
      "id": "qwen3-7-max:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 79.1,
      "unit": "%",
      "printed": "79.1",
      "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, General Capability block, row 'IFBench', column 'Qwen3.7-Max'",
      "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://qwen.ai/blog?id=qwen3.8",
          "name": "Qwen3.8-Max",
          "title": "Qwen3.8-Max: A New Bar for Coding and Cowork",
          "lab": "alibaba",
          "date": "2026-08-03",
          "printed": "79.1",
          "where": "First performance table (Opus4.8, Fable5, GPT5.6 Sol (max), Qwen3.7-Max, Qwen3.8-Max), General Capabilities > IFBench row, Qwen3.7-Max column"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/2fc33c266fcb0a75.png"
    },
    {
      "id": "mistral-medium-3-5:ifbench:default:1",
      "model": "mistral-medium-3-5",
      "model_name": "Mistral Medium 3.5",
      "model_date": "2026-05-22",
      "event": "https://frontierlog.ai/event/2026-05-22-mistral-medium-3-5",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 69.0,
      "unit": "%",
      "printed": "69.0",
      "reported_by": "mistral",
      "run_by": null,
      "label": "Mistral Medium 3.5 128B",
      "source": "https://mistral.ai/news/vibe-remote-agents-mistral-medium-3-5/",
      "source_title": "Remote agents in Vibe. Powered by Mistral Medium 3.5.",
      "source_date": "2026-05-22",
      "where": "chart image 'Math, instruction following vs competing models', group 'Allenai Ifbench', bar 'Mistral Medium 3.5 128B', value label",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "maximum reasoning",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/f8299d5f5295f41f.webp",
      "source_image_original": "https://mistral.ai/_astro/088938d7-ef28-45e5-bff2-52f9d78aca06_Z1DYDxd.webp"
    },
    {
      "id": "qwen3-7-plus:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 79.1,
      "unit": "%",
      "printed": "79.1",
      "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": "'Performance' section, language table (HTML), General Capability block, row 'IFBench', column 'Qwen3.7-Plus'",
      "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/Qwen3.8-Flash-Next",
          "name": "Qwen3.8-Flash-Next card",
          "title": "Qwen/Qwen3.8-Flash-Next model card",
          "lab": "alibaba",
          "date": "2026-08-26",
          "printed": "79.1",
          "where": "Benchmark Results table, General > Instruction following IFBench row, Qwen3.7-Plus column"
        },
        {
          "url": "https://qwen.ai/blog?id=qwen3.8-flash-next",
          "name": "Qwen3.8-Flash-Next",
          "title": "Qwen3.8-Flash-Next: A New Architecture, Towards Ultimate Cost-Efficiency",
          "lab": "alibaba",
          "date": "2026-08-26",
          "printed": "79.1",
          "where": "Performance table, General > Instruction following IFBench row, Qwen3.7-Plus column"
        },
        {
          "url": "https://qwen.ai/blog?id=qwen3.8-omni-flash",
          "name": "Qwen3.8-Omni-Flash",
          "title": "Qwen3.8-Omni-Flash: Omni Senses. Agentic Delivery.",
          "lab": "alibaba",
          "date": "2026-09-18",
          "printed": "79.1",
          "where": "General Text Capabilities table, IFBench row, Qwen3.7-Plus column"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/de169a5a059c61f9.png"
    },
    {
      "id": "inkling:ifbench:default:1",
      "model": "inkling",
      "model_name": "Inkling",
      "model_date": "2026-07-15",
      "event": "https://frontierlog.ai/event/2026-07-15-inkling",
      "benchmark": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 79.8,
      "unit": "%",
      "printed": "79.8%",
      "reported_by": "thinking-machines",
      "run_by": null,
      "label": "Inkling effort=0.99",
      "source": "https://thinkingmachines.ai/news/introducing-inkling/",
      "source_title": "Inkling: Our Open-Weights Model",
      "source_date": "2026-07-15",
      "where": "section 'Benchmarking Inkling', main benchmark table, row 'IFBench' (Reasoning), column 'Inkling effort=0.99'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "effort 0.99",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-10-30",
      "corroborated_by": [
        {
          "url": "https://huggingface.co/thinkingmachines/Inkling",
          "name": "Inkling card",
          "title": "thinkingmachines/Inkling model card",
          "lab": "thinking-machines",
          "date": "2026-07-15",
          "printed": "79.8%",
          "where": "Benchmark table (Inkling results at effort=0.99), Chat > IFBench row, Inkling column"
        },
        {
          "url": "https://thinkingmachines.ai/news/inkling-small/",
          "name": "Thinking Machines post",
          "title": "Introducing Inkling-Small",
          "lab": "thinking-machines",
          "date": "2026-07-30",
          "printed": "79.8%",
          "where": "Benchmark table (all evals at effort 0.99), Chat > IFBench row, Inkling column (7th)"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/a919a67037d3c8a6.png"
    },
    {
      "id": "qwen3-8-max:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 82.8,
      "unit": "%",
      "printed": "82.8",
      "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, language/agent HTML table), General Capabilities block, row 'IFBench', column 'Qwen3.8-Max'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/bfa2cf10e4f8b439.png"
    },
    {
      "id": "qwen3-8-flash-next:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 81.3,
      "unit": "%",
      "printed": "81.3",
      "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 'Instruction following / IFBench', column 'Qwen3.8-Flash-Next'",
      "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/Qwen3.8-Flash-Next",
          "name": "Qwen3.8-Flash-Next card",
          "title": "Qwen/Qwen3.8-Flash-Next model card",
          "lab": "alibaba",
          "date": "2026-08-26",
          "printed": "81.3",
          "where": "Benchmark Results table, General > Instruction following IFBench row, Qwen3.8-Flash-Next column"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/1ce175beb0b20cf1.png"
    },
    {
      "id": "qwen3-8-omni-flash:ifbench:default: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": "ifbench",
      "benchmark_name": "IFBench",
      "setting": "default",
      "setting_name": "",
      "score": 81.5,
      "unit": "%",
      "printed": "81.5",
      "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, row 'IFBench (Instruction following)', column 'Qwen3.8-Omni-Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-30",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/0f628ae53818eafc.png"
    }
  ]
}