{
  "api_version": "1",
  "updated": "2026-10-20",
  "latest_event": "2026-10-18",
  "rights": "© Frontier Log, all rights reserved.",
  "benchmark": {
    "slug": "osworld-2-1",
    "name": "OSWorld 2.1",
    "family": "osworld",
    "unit": "%",
    "higher_is_better": true,
    "status": "active",
    "active": true,
    "superseded_by": null,
    "benchmark_url": "https://github.com/xlang-ai/OSWorld-V2/releases/tag/osworld-v2.1",
    "introduced": "2026-09-16",
    "creator": "XLANG Lab",
    "introduced_source": "https://github.com/xlang-ai/OSWorld-V2/releases/tag/osworld-v2.1",
    "settings": [
      {
        "setting": "partial",
        "name": "Full set, partial score",
        "url": "https://frontierlog.ai/benchmark/osworld-2-1",
        "results": 5
      },
      {
        "setting": "strict",
        "name": "Full set, strict pass rate",
        "url": "https://frontierlog.ai/benchmark/osworld-2-1/strict",
        "results": 3
      },
      {
        "setting": "offline-partial",
        "name": "Offline set, partial score",
        "url": "https://frontierlog.ai/benchmark/osworld-2-1/offline-partial",
        "results": 4
      },
      {
        "setting": "offline-strict",
        "name": "Offline set, strict pass rate",
        "url": "https://frontierlog.ai/benchmark/osworld-2-1/offline-strict",
        "results": 3
      }
    ]
  },
  "total": 3,
  "results": [
    {
      "id": "claude-sonnet-5:osworld-2-1:strict:1",
      "model": "claude-sonnet-5",
      "model_name": "Claude Sonnet 5",
      "model_date": "2026-06-30",
      "event": "https://frontierlog.ai/event/2026-06-30-claude-sonnet-5",
      "benchmark": "osworld-2-1",
      "benchmark_name": "OSWorld 2.1",
      "setting": "strict",
      "setting_name": "Full set, strict pass rate",
      "score": 25.6,
      "unit": "%",
      "printed": "25.6%",
      "reported_by": "anthropic",
      "label": "Claude Sonnet 5",
      "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
      "source_title": "Claude Sonnet 5.5 System Card",
      "source_date": "2026-09-28",
      "where": "section 8.13.3 OSWorld 2.1, text (PDF page 130), the Claude Sonnet 5 strict pass rate",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "max effort; mean of 5 runs; September 10, 2026 task files",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-20",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    },
    {
      "id": "claude-opus-5-5:osworld-2-1:strict:1",
      "model": "claude-opus-5-5",
      "model_name": "Claude Opus 5.5",
      "model_date": "2026-09-22",
      "event": "https://frontierlog.ai/event/2026-09-22-claude-opus-5-5",
      "benchmark": "osworld-2-1",
      "benchmark_name": "OSWorld 2.1",
      "setting": "strict",
      "setting_name": "Full set, strict pass rate",
      "score": 48.7,
      "unit": "%",
      "printed": "48.7%",
      "reported_by": "anthropic",
      "label": "Claude Opus 5.5",
      "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
      "source_title": "Claude Sonnet 5.5 System Card",
      "source_date": "2026-09-28",
      "where": "section 8.13.3 OSWorld 2.1, text (PDF page 130), the Claude Opus 5.5 strict pass rate",
      "own": false,
      "primary": true,
      "parallel": false,
      "notes": "max effort; mean of 5 runs; September 10, 2026 task files",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-20",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    },
    {
      "id": "claude-sonnet-5-5:osworld-2-1:strict:1",
      "model": "claude-sonnet-5-5",
      "model_name": "Claude Sonnet 5.5",
      "model_date": "2026-09-28",
      "event": "https://frontierlog.ai/event/2026-09-28-claude-sonnet-5-5",
      "benchmark": "osworld-2-1",
      "benchmark_name": "OSWorld 2.1",
      "setting": "strict",
      "setting_name": "Full set, strict pass rate",
      "score": 43.5,
      "unit": "%",
      "printed": "43.5%",
      "reported_by": "anthropic",
      "label": "Claude Sonnet 5.5",
      "source": "https://www.anthropic.com/claude-sonnet-5-5-system-card",
      "source_title": "Claude Sonnet 5.5 System Card",
      "source_date": "2026-09-28",
      "where": "section 8.13.3 OSWorld 2.1, text (PDF page 130), the Claude Sonnet 5.5 strict pass rate",
      "own": true,
      "primary": true,
      "parallel": false,
      "notes": "max effort; mean of 5 runs; September 10, 2026 task files",
      "caveat": "",
      "human_verified": true,
      "human_verified_on": "2026-10-20",
      "check": "human",
      "corroborated": false,
      "corroborated_on": null,
      "corroborated_by": []
    }
  ]
}