Claw Bench: leaderboard
Claw Bench is a standardized leaderboard for evaluating AI agent frameworks across task completion, efficiency, security, skills, and UX dimensions.
Metric: Overall (points). Source: github.com. Source: benchmarklist.com. Saturation forecast: Around July 2027. 100 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 92 | #90 |
| 2 | Claude Opus 4.5 | 92 | #83 |
| 3 | Grok 4.20 | 92 | #175 |
| 4 | Claude Sonnet 4.5 | 91 | #142 |
| 5 | Claude Sonnet 4 | 91 | #199 |
| 6 | Gemini 2.5 Pro | 90 | #152 |
| 7 | DeepSeek R1 | 90 | #251 |
| 8 | Claude 3.5 Sonnet | 90 | #350 |
| 9 | DeepSeek V3.2 | 89 | #205 |
| 10 | Llama 4 Maverick | 89 | #464 |
| 11 | GLM-5 | 89 | #141 |
| 12 | Gemini 2.5 Flash | 88 | #245 |
| 13 | Qwen 3.5 Plus | 88 | #125 |
| 14 | DeepSeek Reasoner | 88 | #242 |
| 15 | Qwen 3 Coder Plus | 88 | #338 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=claw-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.