ActBench - OpenClaw - Benign Utility Grading Score: leaderboard
Metric: Mean utility grading score on matched benign cases (0-1). Source: arxiv.org. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kimi K3 | 0.94 |
| 2 | Claude Opus 4.8 | 0.94 |
| 3 | Grok 4.5 | 0.94 |
| 4 | GLM-5.2 | 0.93 |
| 5 | GPT-5.5 | 0.93 |
| 6 | Claude Sonnet 4.6 | 0.93 |
| 7 | DeepSeek V4 Pro | 0.92 |
| 8 | MiniMax-M3 | 0.92 |
| 9 | Qwen 3.7 Max | 0.92 |
| 10 | Qwen 3.7 Plus | 0.92 |
| 11 | GPT-5.4 Mini | 0.9 |
| 12 | DeepSeek V4 Flash | 0.9 |
| 13 | MiniMax-M2.7 | 0.88 |
| 14 | Kimi K2.6 | 0.87 |
Interactive version: theaggregate.ai/benchmark?slug=actbench-openclaw-benign-utility-grading-score · How It Works · Data refreshed daily, snapshot 2026-09-19.