Skip to content
Longterm Wiki has not been actively maintained since May 2026. Some data is still updated occasionally, but please don't rely on this site being current or accurate.
Longterm Wiki

OSWorld

Agentic
A benchmark for multimodal agents on real-world computer tasks across operating systems, testing GUI interaction and task completion.
Models Tested
3
Best Score
72.7%
Median Score
61.4%
Scoring: percentage
Introduced: 2024-04
Maintainer: CMU / HKU

Leaderboard (3 models)

#ModelDeveloperScore
🥇Claude Opus 4.6Anthropic
72.7%
🥈Claude Sonnet 4.5Anthropic
61.4%
🥉Claude Haiku 4.5Anthropic
50.7