SkillsBench: leaderboard
SkillsBench (BenchFlow) scores agentic models on completing tasks across domains using modular Skills (folders of instructions, scripts, and resources), and also reports skill-lift, the with-minus-without-Skills gain.
Metric: With-skills score (%). Source: www.skillsbench.ai. Saturation forecast: Around January 2027. 25 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 67.3 | #31 |
| 2 | GLM-5.1 | 58.4 | #97 |
| 3 | Hy3 | 55.9 | #93 |
| 4 | Gemini 3 Flash | 54.61 | #99 |
| 5 | Claude Opus 4.8 | 54.1 | #37 |
| 6 | Kimi K2.6 | 54 | #104 |
| 7 | Claude Opus 4.7 | 53.1 | #49 |
| 8 | MiniMax-M3 | 53 | #136 |
| 9 | Gemini 3.1 Pro (Preview) | 52.8 | #58 |
| 10 | GPT-5.2 | 51.7 | #110 |
| 11 | Claude Opus 4.6 | 50.22 | #63 |
| 12 | DeepSeek V4 Pro | 50.1 | #102 |
| 13 | Claude Opus 4.5 | 49.04 | #83 |
| 14 | Gemini 3.5 Flash | 48.2 | #60 |
| 15 | Claude Sonnet 4.6 | 47.2 | #90 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=skillsbench · How It Works · Data refreshed daily, snapshot 2026-10-20.