Harvey HLAB - Criteria Pass Rate - Corporate M&A
Professional Work · 2026-05-06
The Corporate M&A practice-area split of Harvey's Legal Agent Benchmark held-out evaluation criteria pass rate. This reports pooled rubric criteria passed within that split and is distinct from the all-pass task-score split.
Top models (higher is better)
| Model | Score |
|---|---|
| Muse Wonder Full | 97.1 |
| Gemini 4 Argon | 96.2 |
| Muse Spark 1.1 | 96.2 |
| Muse Spark 1.3 | 96.2 |
| Muse Spark 1.2 | 95.8 |
| Grok 4.5 | 95.2 |
| Grok 4.6 | 95.2 |
| MiMo-V2.6-Flash | 94.9 |
| GLM-5.3 | 94.6 |
| Qwen3.8-Max | 94.6 |
| DeepSeek-V4-Flash-0731 | 94.2 |
| Claude Opus 5 | 93.9 |
| Ember-1 | 93.9 |
| Hy4 Preview | 93.9 |
| Mistral Large 4 Preview | 93.9 |
| Grok 4.7 | 93.6 |
| Kimi K3 | 93.6 |
| MiMo-V2.6-Pro | 93.6 |
| Muse Wonder | 93.6 |
| Qwen3.8 27B | 93.6 |
| Opus 4.7 | 93.3 |
| Step 5 Preview | 93.3 |
| Claude Haiku 5.5 | 92.9 |
| Opus 4.8 | 92.9 |
| DeepSeek V4.1 Flash 1017 | 92.9 |
| GPT-6.1 Sol | 92.9 |
| Claude Opus 5.5 | 92.3 |
| DeepSeek V4 Pro 0813 | 92.3 |
| Gemini 3.7 Flash | 92.3 |
| DeepSeek V4.1 Flash | 92.0 |
| Nemotron 3 Ultra 550B A55B (NVFP4) | 91.7 |
| Gemini 3.8 Flash | 91.3 |
| Sonnet 4.6 | 91.0 |
| Kimi K2.6 | 91.0 |
| GPT-6 Astra | 90.7 |
| GLM-5.3 Flash | 90.4 |
| MiniMax M3 | 90.1 |
| Qwen3.6 Plus (2026-04-02) | 90.1 |
| GPT-6 Sol | 89.7 |
| GPT-6 Luna | 89.1 |