{
  "api_version": "1",
  "updated": "2026-11-10",
  "latest_event": "2026-11-09",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "osworld-2-0-2026-08-08",
    "name": "OSWorld 2.0 (v2026.08.08)",
    "family": "osworld",
    "unit": "%",
    "higher_is_better": true,
    "status": "superseded",
    "active": false,
    "superseded_by": "osworld-2-1",
    "retired": null,
    "benchmark_url": "https://github.com/xlang-ai/OSWorld-V2/tree/v2026.08.08",
    "introduced": "2026-08-08",
    "creator": "XLANG Lab",
    "introduced_source": "https://github.com/xlang-ai/OSWorld-V2",
    "settings": [
      {
        "setting": "offline-partial",
        "name": "Offline set, partial score",
        "url": "https://frontierlog.ai/benchmark/osworld-2-0-2026-08-08",
        "results": 14
      },
      {
        "setting": "anthropic-corrected-partial",
        "name": "Anthropic's corrected set, partial score",
        "url": "https://frontierlog.ai/benchmark/osworld-2-0-2026-08-08/anthropic-corrected-partial",
        "results": 2
      },
      {
        "setting": "scoring-unstated",
        "name": "Full set, scoring not stated",
        "url": "https://frontierlog.ai/benchmark/osworld-2-0-2026-08-08/scoring-unstated",
        "results": 1
      }
    ]
  },
  "total": 17,
  "results": [
    {
      "id": "gpt-5-6-sol:osworld-2-0-2026-08-08:offline-partial:1",
      "model": "gpt-5-6-sol",
      "model_name": "GPT-5.6 Sol",
      "model_date": "2026-07-09",
      "event": "https://frontierlog.ai/event/2026-07-09-gpt-5-6",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 65.7,
      "unit": "%",
      "printed": "65.7%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-5.6 Sol",
      "source": "https://openai.com/index/gpt-6-astra",
      "source_title": "GPT-6 Astra: A new generation of intelligence",
      "source_date": "2026-09-03",
      "where": "'Computer Use' table, row 'OSWorld 2.0 (v2026.08.08, offline set, partial score)', column 'GPT-5.6 Sol2'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "",
      "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/fbd1c3587f0ca67d.png"
    },
    {
      "id": "claude-opus-5:osworld-2-0-2026-08-08:anthropic-corrected-partial:1",
      "model": "claude-opus-5",
      "model_name": "Claude Opus 5",
      "model_date": "2026-07-24",
      "event": "https://frontierlog.ai/event/2026-07-24-claude-opus-5",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "anthropic-corrected-partial",
      "setting_name": "Anthropic's corrected set, partial score",
      "score": 75.4,
      "unit": "%",
      "printed": "75.4%",
      "reported_by": "anthropic",
      "run_by": null,
      "label": "Opus 5",
      "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
      "source_title": "Introducing Claude Fable 5.1 and Claude Mythos 5.1",
      "source_date": "2026-09-01",
      "where": "benchmark table (HTML), row 'Computer use / OSWorld 2.0 [2]', column 'Opus 5', the value tagged 'partial'",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-19",
      "check": "human",
      "corroborated": true,
      "corroborated_on": "2026-10-19",
      "corroborated_by": [
        {
          "url": "https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system-card",
          "name": "Claude Fable 5.1 and Mythos 5.1 card",
          "title": "Claude Fable 5.1 and Claude Mythos 5.1 System Card",
          "lab": "anthropic",
          "date": "2026-09-01",
          "printed": "75.4%",
          "where": "section 8.14.3 sentence: 'Claude Opus 5 scored 75.4% partial' on the August 2026 task release (Table 8.1.A OSWorld 2.0 (partial/strict) prints 75.4/39.6)"
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/f381a9ad69b68272.png"
    },
    {
      "id": "claude-fable-5-1:osworld-2-0-2026-08-08:anthropic-corrected-partial:1",
      "model": "claude-fable-5-1",
      "model_name": "Claude Fable 5.1",
      "model_date": "2026-09-01",
      "event": "https://frontierlog.ai/event/2026-09-01-claude-fable-5-1-mythos-5-1",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "anthropic-corrected-partial",
      "setting_name": "Anthropic's corrected set, partial score",
      "score": 77.9,
      "unit": "%",
      "printed": "77.9%",
      "reported_by": "anthropic",
      "run_by": null,
      "label": "Fable 5.1",
      "source": "https://www.anthropic.com/claude-fable-and-mythos-5-1",
      "source_title": "Introducing Claude Fable 5.1 and Claude Mythos 5.1",
      "source_date": "2026-09-01",
      "where": "benchmark table (HTML), row 'Computer use / OSWorld 2.0 [2]', column 'Fable 5.1', the value tagged 'partial'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "tasks where safeguards intervened scored zero",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-18",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/f381a9ad69b68272.png"
    },
    {
      "id": "gpt-6-astra:osworld-2-0-2026-08-08:offline-partial:1",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 72.6,
      "unit": "%",
      "printed": "72.6%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT‑6 Astra",
      "source": "https://openai.com/index/gpt-6-astra",
      "source_title": "GPT-6 Astra: A new generation of intelligence",
      "source_date": "2026-09-03",
      "where": "Computer Use table, row 'OSWorld 2.0 (v2026.08.08, offline set, partial score)', column 'GPT‑6 Astra'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/fbd1c3587f0ca67d.png"
    },
    {
      "id": "gpt-6-astra:osworld-2-0-2026-08-08:offline-partial:2",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 62.2,
      "unit": "%",
      "printed": "62.2%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6 Astra (Low)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6 Astra', effort 'Low', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "low effort",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-11-07",
      "corroborated_by": [
        {
          "url": "https://openai.com/index/introducing-gpt-6-sol-and-luna",
          "name": "GPT-6 Sol and Luna",
          "title": "Introducing GPT-6 Sol and Luna",
          "lab": "openai",
          "date": "2026-09-22",
          "printed": "62.2%",
          "where": "\"OSWorld 2.0, offline set\" chart, point aria-label \"cost: $2.55; score: 62%; modelLabel: GPT-6 Astra; Model: GPT-6 Astra; Effort: low; Cost per task: $2.55; Score: 62.2%\""
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-astra:osworld-2-0-2026-08-08:offline-partial:3",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 69.3,
      "unit": "%",
      "printed": "69.3%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6 Astra (Medium)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6 Astra', effort 'Medium', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "medium effort",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-11-07",
      "corroborated_by": [
        {
          "url": "https://openai.com/index/introducing-gpt-6-sol-and-luna",
          "name": "GPT-6 Sol and Luna",
          "title": "Introducing GPT-6 Sol and Luna",
          "lab": "openai",
          "date": "2026-09-22",
          "printed": "69.3%",
          "where": "\"OSWorld 2.0, offline set\" chart, point aria-label \"cost: $5.10; score: 69%; modelLabel: GPT-6 Astra; Model: GPT-6 Astra; Effort: medium; Cost per task: $5.10; Score: 69.3%\""
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-astra:osworld-2-0-2026-08-08:offline-partial:4",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 70.0,
      "unit": "%",
      "printed": "70.0%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6 Astra (High)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6 Astra', effort 'High', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "high effort",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-11-07",
      "corroborated_by": [
        {
          "url": "https://openai.com/index/introducing-gpt-6-sol-and-luna",
          "name": "GPT-6 Sol and Luna",
          "title": "Introducing GPT-6 Sol and Luna",
          "lab": "openai",
          "date": "2026-09-22",
          "printed": "70.0%",
          "where": "\"OSWorld 2.0, offline set\" chart, point aria-label \"cost: $6.60; score: 70%; modelLabel: GPT-6 Astra; Model: GPT-6 Astra; Effort: high; Cost per task: $6.60; Score: 70.0%\""
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-astra:osworld-2-0-2026-08-08:offline-partial:5",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 71.3,
      "unit": "%",
      "printed": "71.3%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6 Astra (Xhigh)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6 Astra', effort 'Xhigh', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "xhigh effort",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-11-07",
      "corroborated_by": [
        {
          "url": "https://openai.com/index/introducing-gpt-6-sol-and-luna",
          "name": "GPT-6 Sol and Luna",
          "title": "Introducing GPT-6 Sol and Luna",
          "lab": "openai",
          "date": "2026-09-22",
          "printed": "71.3%",
          "where": "\"OSWorld 2.0, offline set\" chart, point aria-label \"cost: $7.17; score: 71%; modelLabel: GPT-6 Astra; Model: GPT-6 Astra; Effort: xhigh; Cost per task: $7.17; Score: 71.3%\""
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-astra:osworld-2-0-2026-08-08:offline-partial:6",
      "model": "gpt-6-astra",
      "model_name": "GPT-6 Astra",
      "model_date": "2026-09-03",
      "event": "https://frontierlog.ai/event/2026-09-03-gpt-6-astra",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 73.5,
      "unit": "%",
      "printed": "73.5%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6 Astra (Max)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6 Astra', effort 'Max', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": false,
      "primary": false,
      "parallel": false,
      "notes": "max effort",
      "caveat": "",
      "human_verified": false,
      "human_verified_on": null,
      "check": "corroborated",
      "corroborated": true,
      "corroborated_on": "2026-11-07",
      "corroborated_by": [
        {
          "url": "https://openai.com/index/introducing-gpt-6-sol-and-luna",
          "name": "GPT-6 Sol and Luna",
          "title": "Introducing GPT-6 Sol and Luna",
          "lab": "openai",
          "date": "2026-09-22",
          "printed": "73.5%",
          "where": "\"OSWorld 2.0, offline set\" chart, point aria-label \"cost: $9.07; score: 73%; modelLabel: GPT-6 Astra; Model: GPT-6 Astra; Effort: max; Cost per task: $9.07; Score: 73.5%\""
        }
      ],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-sol:osworld-2-0-2026-08-08:offline-partial:1",
      "model": "gpt-6-sol",
      "model_name": "GPT-6 Sol",
      "model_date": "2026-09-22",
      "event": "https://frontierlog.ai/event/2026-09-22-gpt-6-sol-and-luna",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 60.5,
      "unit": "%",
      "printed": "60.5%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6 Sol (xhigh)",
      "source": "https://openai.com/index/introducing-gpt-6-sol-and-luna",
      "source_title": "Introducing GPT-6 Sol and Luna",
      "source_date": "2026-09-22",
      "where": "'Computer use' section, paragraph 'On OSWorld 2.0 offline, GPT-6 Sol at xhigh effort achieves a similar score to Claude Opus 5 ...' (first figure of the pair)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "xhigh effort",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-20",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    },
    {
      "id": "gpt-6-1-sol:osworld-2-0-2026-08-08:offline-partial:1",
      "model": "gpt-6-1-sol",
      "model_name": "GPT-6.1 Sol",
      "model_date": "2026-09-29",
      "event": "https://frontierlog.ai/event/2026-09-29-gpt-6-1-sol",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 59.0,
      "unit": "%",
      "printed": "59.0%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6.1 Sol (Low)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6.1 Sol', effort 'Low', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": true,
      "primary": false,
      "parallel": false,
      "notes": "low effort",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-11-08",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-1-sol:osworld-2-0-2026-08-08:offline-partial:2",
      "model": "gpt-6-1-sol",
      "model_name": "GPT-6.1 Sol",
      "model_date": "2026-09-29",
      "event": "https://frontierlog.ai/event/2026-09-29-gpt-6-1-sol",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 66.8,
      "unit": "%",
      "printed": "66.8%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6.1 Sol (Medium)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6.1 Sol', effort 'Medium', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": true,
      "primary": false,
      "parallel": false,
      "notes": "medium effort",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-11-08",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-1-sol:osworld-2-0-2026-08-08:offline-partial:3",
      "model": "gpt-6-1-sol",
      "model_name": "GPT-6.1 Sol",
      "model_date": "2026-09-29",
      "event": "https://frontierlog.ai/event/2026-09-29-gpt-6-1-sol",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 69.6,
      "unit": "%",
      "printed": "69.6%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6.1 Sol (High)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6.1 Sol', effort 'High', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": true,
      "primary": false,
      "parallel": false,
      "notes": "high effort",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-11-08",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-1-sol:osworld-2-0-2026-08-08:offline-partial:4",
      "model": "gpt-6-1-sol",
      "model_name": "GPT-6.1 Sol",
      "model_date": "2026-09-29",
      "event": "https://frontierlog.ai/event/2026-09-29-gpt-6-1-sol",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 69.4,
      "unit": "%",
      "printed": "69.4%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6.1 Sol (Xhigh)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6.1 Sol', effort 'Xhigh', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": true,
      "primary": false,
      "parallel": false,
      "notes": "xhigh effort",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-11-08",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gpt-6-1-sol:osworld-2-0-2026-08-08:offline-partial:5",
      "model": "gpt-6-1-sol",
      "model_name": "GPT-6.1 Sol",
      "model_date": "2026-09-29",
      "event": "https://frontierlog.ai/event/2026-09-29-gpt-6-1-sol",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 71.4,
      "unit": "%",
      "printed": "71.4%",
      "reported_by": "openai",
      "run_by": null,
      "label": "GPT-6.1 Sol (Max)",
      "source": "https://openai.com/index/introducing-gpt-6-1-sol",
      "source_title": "Introducing GPT-6.1 Sol",
      "source_date": "2026-09-29",
      "where": "\"OSWorld 2.0, offline set\" chart under \"Computer use\" (score vs cost per task; no value labels): point 'GPT-6.1 Sol', effort 'Max', read from the point's aria-label (load the page in a browser and hover or inspect the SVG point)",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "max effort",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-11-08",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_capture": "https://frontierlog.ai/captures/624489870d3f5366.png"
    },
    {
      "id": "gemini-4-argon:osworld-2-0-2026-08-08:offline-partial:1",
      "model": "gemini-4-argon",
      "model_name": "Gemini 4 Argon",
      "model_date": "2026-09-30",
      "event": "https://frontierlog.ai/event/2026-09-30-gemini-4-argon",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "offline-partial",
      "setting_name": "Offline set, partial score",
      "score": 69.2,
      "unit": "%",
      "printed": "69.2%",
      "reported_by": "google",
      "run_by": null,
      "label": "Gemini 4 Argon",
      "source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/",
      "source_title": "Gemini 4 Argon: our next era of frontier intelligence",
      "source_date": "2026-09-30",
      "where": "benchmark table (image), row 'OSWorld-2.0', 'Offline subset, Partial score', column 'Gemini 4 Argon'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-16",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/481de187c66d4b4d.gif",
      "source_image_original": "https://storage.googleapis.com/gweb-uniblog-publish-prod/original_images/gemini-4-argon_table_blog.gif"
    },
    {
      "id": "minimax-m3-1-flash:osworld-2-0-2026-08-08:scoring-unstated:1",
      "model": "minimax-m3-1-flash",
      "model_name": "MiniMax M3.1 Flash",
      "model_date": "2026-10-19",
      "event": "https://frontierlog.ai/event/2026-10-19-minimax-m3-1-flash",
      "benchmark": "osworld-2-0-2026-08-08",
      "benchmark_name": "OSWorld 2.0 (v2026.08.08)",
      "setting": "scoring-unstated",
      "setting_name": "Full set, scoring not stated",
      "score": 61.7,
      "unit": "%",
      "printed": "61.7",
      "reported_by": "minimax",
      "run_by": null,
      "label": "MiniMax M3.1 Flash",
      "source": "https://www.minimax.io/blog/minimax-m31f",
      "source_title": "MiniMax M3.1 Flash: Redefining Agent Performance and Cost",
      "source_date": "2026-10-19",
      "where": "benchmark table image (first image), row 'OSWorld 2.0', column 'MiniMax M3.1 Flash'",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-26",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": [],
      "source_image": "https://frontierlog.ai/sources/5525d0bcceba0858.m31-flash-en-2-1792403273121-webp",
      "source_image_original": "https://filecdn.minimax.chat/public/m31-flash-en-2-1792403273121-webp-1.m31-flash-en-2-1792403273121-webp"
    }
  ]
}