List your company

Computer Use on OSWorld (Verified)

75ScoreGPT-5.4
5.94423.87241.859.728Feb 2, 2026Feb 19, 2026Mar 9, 2026Mar 27, 2026Apr 13, 2026May 1, 2026May 19, 2026

Evaluation results

MethodLinks
GPT-5.42026.0575
Kimi-K2.62026.0573.1
Claude-Sonnet-4.62026.0572.5
Claude Opus 4.5Evaluation Source=External2026.0266.3
Kimi K2.5Evaluation Source=External2026.0263.3
Qwen-3.5-27B2026.0556.2
GUI-OWL-1.5-8B2026.0552.3
EvoCUA-8B2026.0546.1
Qwen-3.5-9B2026.0541.8
Qwen3-VL-235B-A22BEvaluation Source=External2026.0238.1
Gemini 3 ProEvaluation Source=Internal2026.0220.7
GPT-5.2 (xhigh)Evaluation Source=Internal2026.028.6