OSWorld (OSWorld)

A computer-use benchmark for GUI task completion across the broader OSWorld task suite.

Top Models on OSWorld — March 2026

As of March 2026, Claude Opus 4.6 leads the OSWorld leaderboard with 72.2% , followed by Claude Opus 4.5 (66.3%) and Kimi K2.5 (63.3%).

4 modelsAgenticCurrentDisplay onlyUpdated April 1, 2026

According to BenchLM.ai, Claude Opus 4.6 leads the OSWorld benchmark with a score of 72.2%, followed by Claude Opus 4.5 (66.3%) and Kimi K2.5 (63.3%). The scores show moderate spread, with meaningful differences between the top tier and mid-tier models.

4 models have been evaluated on OSWorld. The benchmark falls in the Agentic category. This category carries a 22% weight in BenchLM.ai's overall scoring system. OSWorld is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About OSWorld

Year

2026

Tasks

Computer-use tasks

Format

Interactive GUI evaluation

Difficulty

Broad computer-use suite

BenchLM tracks plain OSWorld as a display-only provider-table reference and preserves OSWorld-Verified as the weighted core benchmark key.

GLM-5V-Turbo

BenchLM freshness & provenance

Version

OSWorld 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Leaderboard (4 models)

#1Claude Opus 4.6
72.2%
#2Claude Opus 4.5
66.3%
#3Kimi K2.5
63.3%
#4GLM-5V-Turbo
62.3%

FAQ

What does OSWorld measure?

A computer-use benchmark for GUI task completion across the broader OSWorld task suite.

Which model scores highest on OSWorld?

Claude Opus 4.6 by Anthropic currently leads with a score of 72.2% on OSWorld.

How many models are evaluated on OSWorld?

4 AI models have been evaluated on OSWorld on BenchLM.

Last updated: April 1, 2026 · BenchLM version OSWorld 2026

Weekly LLM Benchmark Digest

Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.

Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.