ClawsBench - Capability (OpenClaw; skills off; meta off): leaderboard
Metric: Mean native task score x100 (-100 to 100); OpenClaw, domain skills off, meta prompt off; mean over observed scored trials, including scored error rows; higher is better. Source: huggingface.co. Saturation forecast: Around December 2027. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 11.85 |
| 2 | GLM-5 | 5.9 |
| 3 | GPT-5.4 | 2.71 |
| 4 | Claude Sonnet 4.6 | 2.61 |
| 5 | Claude Opus 4.6 | 1.75 |
| 6 | Gemini 3.1 Flash Lite (Preview) | 1.75 |
Interactive version: theaggregate.ai/benchmark?slug=clawsbench-capability-openclaw-skills-off-meta-off · How It Works · Data refreshed daily, snapshot 2026-10-09.