OSWorld 2.0: leaderboard
Long-horizon computer-use benchmark with 108 end-to-end desktop workflows. Tests agents on realistic GUI, file, browser, and application tasks with binary completion as the primary metric.
Metric: Binary Accuracy (%). Source: osworld-v2.xlang.ai. Saturation forecast: Around March 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 48.65 | #12 |
| 2 | GPT-5.6 Sol | 28.1 | #19 |
| 3 | Claude Opus 4.8 | 20.6 | #37 |
| 4 | Claude Opus 4.7 | 18.2 | #49 |
| 5 | GPT-5.5 | 13 | #31 |
| 6 | Claude Sonnet 4.6 | 9.3 | #90 |
| 7 | Kimi K2.6 | 4.6 | #104 |
| 8 | MiniMax-M3 | 4.6 | #136 |
| 9 | Qwen 3.7 Plus | 2.8 | #92 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=osworld-2-0 · How It Works · Data refreshed daily, snapshot 2026-10-20.