LiveClawBench - Hard: leaderboard
Metric: Average reward (0-100) over three runs on the 19 Hard cases (DeepSeek-V4-Pro anchor reward 0.3 or less), OpenClaw-style personal-assistant tasks run in reproducible full-stack mock services (22 stateful web applications, browser, APIs and shell), three runs per case, rewards from case verifiers and a DeepSeek-V3.2 rubric judge for open-ended cases; higher is better. Source: arxiv.org. Saturation forecast: Around July 2028. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GLM-5.1 | 27.7 |
| 2 | Kimi K2.7 Code | 25.9 |
| 3 | MiniMax-M3 | 25.7 |
| 4 | GLM-5.2 | 24.3 |
| 5 | GPT-5.5 | 22 |
| 6 | MiniMax-M2.7 | 19.5 |
| 7 | Kimi K2.6 | 18.3 |
| 8 | MiMo-V2.5-Pro | 17.2 |
| 9 | Qwen 3.6 Flash | 14.3 |
| 10 | Qwen 3.6 Plus | 13.2 |
| 11 | DeepSeek V4 Flash | 13 |
| 12 | Qwen 3.5 Plus | 12.8 |
| 13 | Claude Opus 4.8 | 10.2 |
| 14 | Qwen 3.5 Flash | 5 |
| 15 | Qwen 3.6 27B | 4.7 |
Interactive version: theaggregate.ai/benchmark?slug=liveclawbench-hard · How It Works · Data refreshed daily, snapshot 2026-10-07.