WildClawBench: leaderboard
60 hand-built agent tasks run inside the OpenClaw assistant with real shells, files, browsers, email and calendar, spanning productivity, coding, social, search, creative and safety; InternLM, 2026.
Metric: Overall Score (self-reported). Source: benchmarklist.com. Status: saturation imminent. 66 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 67.2 |
| 2 | Seed 2.1 Turbo | 62.8 |
| 3 | Claude Opus 4.7 | 62.2 |
| 4 | GPT-5.5 | 58.2 |
| 5 | Kimi K3 | 54.5 |
| 6 | Nex N2 Pro | 53.5 |
| 7 | Claude Opus 4.6 | 51.6 |
| 8 | GPT-5.4 | 50.3 |
| 9 | GLM-5.1 | 48.2 |
| 10 | Muse Glimmer 30B | 47.6 |
| 11 | Kimi K2.7 Code | 46.9 |
| 12 | DeepSeek V4 Pro | 43.7 |
| 13 | GLM-5 | 42.6 |
| 14 | Gemini 3.1 Pro (Preview) | 40.8 |
| 15 | MiMo-V2-Pro | 40.2 |
Interactive version: theaggregate.ai/benchmark?slug=wildclawbench · How It Works · Data refreshed daily, snapshot 2026-09-05.