Which AI model scores highest on OSWorld 2.0?

Claude Opus 5 by Anthropic holds the best OSWorld 2.0 result among tracked models, at 70.6% (released Jul 24 2026). Higher scores are better on this benchmark.

What are the top 5 models on OSWorld 2.0?

1. Claude Opus 5 (Anthropic) — 70.6%; 2. Muse Spark 1.3 (Meta) — 66.9%; 3. Gemini 3.8 Flash (Google) — 59%; 4. Gemini 3.7 Flash (Google) — 38.1%; 5. Qwen3.8-Flash-Next (Qwen) — 19.4%.

How many models have a published OSWorld 2.0 score?

5 tracked models have a published OSWorld 2.0 score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.

What is the best open model on OSWorld 2.0?

Qwen3.8-Flash-Next by Qwen is the highest-ranked model with downloadable weights on OSWorld 2.0, scoring 19.4% at rank 5 overall.

Agentic computer use

OSWorld 2.0

Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Version 2.0 is a harder, refreshed task set. Higher is better.

Rankings

Higher is better

OSWorld 2.0 — frequently asked questions

: Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Version 2.0 is a harder, refreshed task set. Higher is better.
: Claude Opus 5 by Anthropic holds the best OSWorld 2.0 result among tracked models, at 70.6% (released Jul 24 2026). Higher scores are better on this benchmark.
: 1. Claude Opus 5 (Anthropic) — 70.6%; 2. Muse Spark 1.3 (Meta) — 66.9%; 3. Gemini 3.8 Flash (Google) — 59%; 4. Gemini 3.7 Flash (Google) — 38.1%; 5. Qwen3.8-Flash-Next (Qwen) — 19.4%.
: 5 tracked models have a published OSWorld 2.0 score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.
: Qwen3.8-Flash-Next by Qwen is the highest-ranked model with downloadable weights on OSWorld 2.0, scoring 19.4% at rank 5 overall.