SkillsBench with Skills
Agents · 2026-02-13
SkillsBench resolution rate for the curated-Skills condition on the 87-task v1.1 suite. The current site leaderboard reports 25 model-harness configurations: 24 paired with a no-Skills condition and one audited with-Skills-only result.
Top models (higher is better)
| Model | Score |
|---|---|
| DeepSeek V4.1 Flash | 69.8 |
| GPT-5.5 | 67.3 |
| Grok 4.5 | 66.0 |
| Gemini 3.7 Flash | 65.9 |
| Opus 4.7 | 61.2 |
| Gemini 3.1 Pro Preview | 60.8 |
| GPT-5.6 Luna | 60.4 |
| Opus 4.8 | 59.2 |
| Muse Spark 1.1 | 59.2 |
| GPT-5.6 Terra | 58.9 |
| GLM-5.1 | 58.4 |
| Gemini 3.8 Flash | 58.0 |
| Hy3 | 55.9 |
| Grok 4.6 | 55.8 |
| Gemini 3 Flash Preview | 54.6 |
| Qwen3.7-Plus (source unspecified) | 54.3 |
| GPT-5.6 Sol | 54.1 |
| Kimi K2.6 | 54.0 |
| DeepSeek V4 Pro 0813 | 53.8 |
| Muse Spark 1.2 | 53.0 |
| MiniMax M3 | 53.0 |
| Gemini 3.5 Flash | 52.7 |
| GPT-5.4 | 51.7 |
| GPT-5.2-Codex | 51.7 |
| DeepSeek-V4-Pro | 51.3 |
| DeepSeek-V4-Flash-0731 | 50.7 |
| Opus 4.6 | 50.2 |
| Kimi K2.7 Code | 50.0 |
| Sonnet 4.6 | 49.1 |
| Opus 4.5 | 49.0 |
| GLM-5.3 | 47.5 |
| Sonnet 5 | 46.5 |
| GLM-5.2 | 45.1 |
| DeepSeek-V4-Flash | 44.7 |
| Qwen3.8-Max | 42.0 |
| Grok 4.3 | 41.7 |
| GPT-5.4 Mini | 41.4 |
| GLM-5.3 Flash | 40.2 |
| Qwen3.8 27B | 38.1 |
| Sonnet 4.5 | 36.2 |