Agentic computer use
OSWorld-Verified
Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Higher is better.
Rankings
Higher is better185%
Claude Fable 5
Anthropic · Jun 9 2026
283.4%
Claude Opus 4.8
Anthropic · May 28 2026
381.2%
Claude Sonnet 5
Anthropic · Jun 30 2026
480.8%
Muse Spark 1.1
Meta · Jul 9 2026
578.7%
GPT-5.5
OpenAI · Apr 23 2026
678.4%
Gemini 3.5 Flash
Google · May 19 2026
778%
Claude Opus 4.7
Anthropic · Apr 16 2026
876.2%
Gemini 3.1 Pro
Google · Feb 19 2026
975%
GPT-5.4
OpenAI · Mar 5 2026
1072.5%
Claude Sonnet 4.6
Anthropic · Feb 17 2026
1165.1%
Gemini 3.0 Flash
Google · Dec 17 2025
1253.3%
Muse Spark
Meta · Apr 8 2026