Skip to content
Bencher

TestsBenchmarks · Agentic

OSWorld 2.0 (strict pass rate)

OSWorld 2.0 tasks counted only if fully passed.

% solvedHigher is betterThe test's websiteOfficial page ↗

The best 3 · 3 models tested

Top 3 · best setting per model · 3 models, 3 results (0 independent)

Measured by independent testersIndependent Reported by the makerVendor-reported
OSWorld 2.0 (strict pass rate) leaderboard
Thinking levelSetting
1Claude Fable 5.1Anthropic41.7%Maxadaptive thinking, effort=maxMaker's own figureVendor-reportedAnthropic ↗—1 Sep 2026
2Claude Opus 5Anthropic39.6%Maxadaptive thinking, effort=maxMaker's own figureVendor-reportedAnthropic ↗—1 Sep 2026
3Claude Fable 5Anthropic36.1%Maxadaptive thinking, effort=maxMaker's own figureVendor-reportedAnthropic ↗—1 Sep 2026