LLM Stats (Claw-Eval): leaderboard
Agentic task evaluation testing LLM performance on complex workflows requiring multi-step reasoning, tool use, and legal/contractual document analysis.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around 2028. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MiniMax-M3 | 74.5 | #136 |
| 2 | Hy3 | 68.5 | #93 |
| 3 | Qwen 3.7 Max | 65.2 | #75 |
| 4 | MiMo-V2.5-Pro | 64 | #158 |
| 5 | MiMo-V2.5 | 63.2 | #196 |
| 6 | LFM2.5-2.6B | 62.85 | #822 |
| 7 | Qwen 3.7 Plus | 62.7 | #92 |
| 8 | MiMo-V2-Pro | 61.5 | #168 |
| 9 | Qwen 3.6 27B | 60.6 | #170 |
| 10 | Qwen 3.6 Plus | 58.7 | #120 |
| 11 | MiMo-V2-Omni | 54.8 | #253 |
| 12 | Qwen 3.6 35B A3B | 50 | #226 |
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-claw-eval · How It Works · Data refreshed daily, snapshot 2026-10-20.