ClawProBench: leaderboard
OpenClaw agent benchmark measuring model performance on reasoning, planning, tool use, reliability, efficiency, and safety across repeated runs.
Metric: Reported Average Score (%). Source: suyoumo.github.io. Saturation forecast: Around July 2027. 71 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 (xHigh) | 67.9 | #31 (GPT-5.5) |
| 2 | DeepSeek V4 Pro | 64.38 | #102 |
| 3 | Qwen 3.5 Plus | 64.19 | #125 |
| 4 | Qwen 3.5 397B A17B | 64.18 | #146 |
| 5 | MiMo-V2.5-Pro | 63.3 | #158 |
| 6 | GLM-5.1 | 62.93 | #97 |
| 7 | GLM-5-Turbo | 61.92 | #119 |
| 8 | DeepSeek V4 Flash | 61.47 | #137 |
| 9 | Seed 2.0 Pro | 61.07 | #118 |
| 10 | Claude Sonnet 4.6 | 60.5 | #90 |
| 11 | Seed 2.0 Lite | 60.4 | #145 |
| 12 | MiMo-V2.5 | 60.39 | #196 |
| 13 | Qwen 3.6 Plus | 60.2 | #120 |
| 14 | DeepSeek V3.2 | 60.13 | #205 |
| 15 | Kimi K2.6 | 59.31 | #104 |
Interactive version: theaggregate.ai/benchmark?slug=clawprobench · How It Works · Data refreshed daily, snapshot 2026-10-20.