Agentic Skills Evaluation Framework: leaderboard
Agent skills -- structured, reusable knowledge artifacts that augment LLM agent capabilities -- have been rapidly adopted in industry, yet their cross-domain impact and use across commercial and open-source models remain under-studied, and no reusable methodology exists for evaluating an individual skill. In this work, we present an evaluation framework that lets a skill author construct realistic
Metric: Overall Score w/ (self-reported). Source: benchmarklist.com. Saturation forecast: Around August 2027. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 92.7 | #37 |
| 2 | Claude Opus 4.7 | 92.3 | #49 |
| 3 | Claude Sonnet 4.6 | 91.5 | #90 |
| 4 | GLM-5.1 | 91.1 | #97 |
| 5 | Gemini 3.5 Flash | 88.6 | #60 |
| 6 | Gemini 3.1 Pro (Preview) | 88.5 | #58 |
| 7 | GPT-5.4 | 88.2 | #79 |
| 8 | DeepSeek V4 Pro | 86.6 | #102 |
| 9 | Gemini 3 Flash (Preview) | 85.9 | #82 |
| 10 | GPT-5.4 Mini | 84.5 | #212 |
| 11 | Claude Haiku 4.5 | 84.1 | #280 |
| 12 | DeepSeek V4 Flash | 83.9 | #137 |
| 13 | GPT-5.4 Nano | 81.9 | #328 |
| 14 | Kimi K2.6 | 76.1 | #104 |
| 15 | Gemini 3.1 Flash Lite | 71.4 | #191 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=agentic-skills-evaluation-framework · How It Works · Data refreshed daily, snapshot 2026-10-20.