OSWorld 2.0 Strict Pass Rate (Opus 5.5 System Card): leaderboard
OSWorld 2.0 (Yuan et al., 2026): 108 long-horizon computer-use tasks on a live Ubuntu VM, graded against weighted checkpoints. Strict pass rate is the percentage of tasks with every checkpoint satisfied, Pass@1 over five runs at max effort (higher is better). Anthropic's re-run on the September 10, 2026 task files with every screenshot retained, from Table 8.1.A of the Claude Opus 5.5 System Card (September 2026).
Metric: Strict Pass Rate (%). Source: www-cdn.anthropic.com. Saturation forecast: Around 2029. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) | 48.7 | #3 (Claude Opus 5.5) |
| 2 | Claude Fable 5.1 (Max) | 42.8 | #10 (Claude Fable 5.1) |
| 3 | Claude Opus 5 (Max) | 37.2 | #15 (Claude Opus 5) |
No result here: #5 GPT-6 Astra, #6 GPT-6.1 Sol, #9 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=osworld-2-0-strict-pass-rate-opus-5-5-system-card · How It Works · Data refreshed daily, snapshot 2026-11-07.