LLM Stats (WildClawBench): leaderboard
End-to-end AI agent benchmark with 60 original tasks in a live OpenClaw environment spanning productivity, code intelligence, social interaction, search, creative synthesis, and safety alignment workflows.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen3.8 Omni Flash | 71 | #28 |
| 2 | Seed 2.1 Turbo | 62.8 | #86 |
| 3 | Seed 2.1 Pro | 61.7 | #67 |
| 4 | Hy3 | 53.6 | #93 |
| 5 | Muse Glimmer 30B | 47.6 | #178 |
| 6 | MiMo-V2.5-Pro | 43 | #158 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-wildclawbench · How It Works · Data refreshed daily, snapshot 2026-10-20.