Skill execution on SkillsBench
56.9Overall Success Rate (avg@5)Evolver
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| EvolverRefinement iterations (R)=22026.05 | 56.9 | 50.4 | 63.8 | 83.9 | 72 | 40 | 60 | 84 | 64 | — | |
| Revision v3Model=GPT-5.5, Harness=Codex2026.05 | 53 | — | — | — | — | — | — | — | — | 25.58 | |
| EvolverRefinement iterations (R)=12026.05 | 48.2 | 18 | 58 | 75 | 40 | 23.3 | 20 | 45 | 46.7 | — | |
| Revision v3Model=Opus-4.7, Harness=Claude Code2026.05 | 44 | — | — | — | — | — | — | — | — | 32.56 | |
| Human-curated skill2026.05 | 43.6 | 38.8 | 50 | 43.3 | 40 | 30 | 52.5 | 32 | 62 | — | |
| Revision v3Model=DeepSeek-V4-Pro, Harness=Claude Code2026.05 | 41 | — | — | — | — | — | — | — | — | 31.4 | |
| Revision v3Model=Kimi-2.6, Harness=Claude Code2026.05 | 37 | — | — | — | — | — | — | — | — | 30.23 | |
| Skill v0Model=GPT-5.5, Harness=Codex2026.05 | 35 | — | — | — | — | — | — | — | — | 4.65 | |
| Skill-CreatorModel=GPT-5.5, Harness=Codex2026.05 | 34 | — | — | — | — | — | — | — | — | 3.49 | |
| SkillCreator-SkillsBench2026.05 | 33.9 | 8.6 | 50 | 46.4 | 25 | 20 | 26.7 | 63.8 | 30 | — | |
| Self-Gen2026.05 | 32 | — | — | — | — | — | — | — | — | — | |
| No skillModel=GPT-5.5, Harness=Codex2026.05 | 31 | — | — | — | — | — | — | — | — | — | |
| No skill2026.05 | 29.9 | 35 | 24.3 | 38.3 | 29.1 | 15 | 27.5 | 32 | 40 | — | |
| Skill-CreatorModel=Opus-4.7, Harness=Claude Code2026.05 | 28 | — | — | — | — | — | — | — | — | 13.95 | |
| Revision v3Model=Qwen-3.6plus, Harness=Claude Code2026.05 | 27 | — | — | — | — | — | — | — | — | 24.42 | |
| Skill v0Model=DeepSeek-V4-Pro, Harness=Claude Code2026.05 | 23 | — | — | — | — | — | — | — | — | 10.47 | |
| Skill-CreatorModel=DeepSeek-V4-Pro, Harness=Claude Code2026.05 | 21 | — | — | — | — | — | — | — | — | 8.14 | |
| Skill v0Model=Kimi-2.6, Harness=Claude Code2026.05 | 19 | — | — | — | — | — | — | — | — | 9.3 | |
| Skill v0Model=Opus-4.7, Harness=Claude Code2026.05 | 17 | — | — | — | — | — | — | — | — | 1.16 | |
| Skill-CreatorModel=Kimi-2.6, Harness=Claude Code2026.05 | 17 | — | — | — | — | — | — | — | — | 6.98 | |
| No skillModel=Opus-4.7, Harness=Claude Code2026.05 | 16 | — | — | — | — | — | — | — | — | — | |
| No skillModel=DeepSeek-V4-Pro, Harness=Claude Code2026.05 | 14 | — | — | — | — | — | — | — | — | — | |
| No skillModel=Kimi-2.6, Harness=Claude Code2026.05 | 11 | — | — | — | — | — | — | — | — | — | |
| Skill v0Model=Qwen-3.6plus, Harness=Claude Code2026.05 | 9 | — | — | — | — | — | — | — | — | 3.49 | |
| Skill-CreatorModel=Qwen-3.6plus, Harness=Claude Code2026.05 | 8 | — | — | — | — | — | — | — | — | 2.33 | |
| No skillModel=Qwen-3.6plus, Harness=Claude Code2026.05 | 6 | — | — | — | — | — | — | — | — | — |