SkillsBench Leaderboard

Benchmark Model Rank Results
agents-on-skillsbench-v1-1-with-skillsGLM 5.1 (OpenHands)#1Pass Rate: 58.4
agents-on-skillsbench-v1-1-with-skillsKimi K2.6 (OpenHands)#2Pass Rate: 54
agents-on-skillsbench-v1-1-with-skillsMiniMax M3 (OpenHands)#3Pass Rate: 53
agents-on-skillsbench-v1-1-with-skillsDeepSeek V4 Pro (OpenHands)#4Pass Rate: 50.1
agents-on-skillsbench-v1-1-with-skillsDeepSeek V4 Flash (OpenHands)#6Pass Rate: 44.7
agents-on-skillsbench-v1-1-with-skillsMiniMax M2.7 (OpenHands)#7Pass Rate: 34.9
agents-on-skillsbench-v1-1-with-skillsGPT-5.2 (Codex)–Pass Rate: 51.7046
agents-on-skillsbench-v1-1-with-skillsGPT-5.4 Mini (OpenHands)–Pass Rate: 41.4
agents-on-skillsbench-v1-1-with-skillsGPT-5.5 (OpenHands)–Pass Rate: 67.3
agents-on-skillsbench-v1-1-with-skillsGemini 3 Flash (Gemini CLI)–Pass Rate: 54.6068
agents-on-skillsbench-v1-1-with-skillsGemini 3.1 Flash Lite (OpenHands)–Pass Rate: 20.1
agents-on-skillsbench-v1-1-with-skillsGemini 3.1 Pro (Gemini CLI)–Pass Rate: 60.8
agents-on-skillsbench-v1-1-with-skillsGemini 3.5 Flash (OpenHands)–Pass Rate: 48.2
agents-on-skillsbench-v1-1-with-skillsGrok 4.3 (OpenHands)–Pass Rate: 41.7
agents-on-skillsbench-v1-1-with-skillsHaiku 4.5 (Claude Code)–Pass Rate: 30.0831
agents-on-skillsbench-v1-1-with-skillsOpus 4.5 (Claude Code)–Pass Rate: 49.0395
agents-on-skillsbench-v1-1-with-skillsOpus 4.6 (Claude Code)–Pass Rate: 50.2241
agents-on-skillsbench-v1-1-with-skillsOpus 4.7 (Claude Code)–Pass Rate: 61.2
agents-on-skillsbench-v1-1-with-skillsOpus 4.8 (OpenHands)–Pass Rate: 54.1
agents-on-skillsbench-v1-1-with-skillsSonnet 4.5 (Claude Code)–Pass Rate: 36.1723
agents-on-skillsbench-v1-1-with-skillsSonnet 4.6 (OpenHands)–Pass Rate: 47.2