Harvey HLAB - Criteria Pass Rate - Emerging Companies Venture Capital
Professional Work · 2026-05-06
The Emerging Companies Venture Capital 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 Spark 1.3 | 95.9 |
| Muse Wonder | 95.6 |
| Muse Spark 1.2 | 95.2 |
| Gemini 4 Argon | 94.9 |
| Muse Spark Open | 94.0 |
| MiMo-V2.6-Pro | 93.9 |
| Step 5 Preview | 93.5 |
| Muse Wonder Full | 93.2 |
| Grok 4.7 | 92.8 |
| DeepSeek V4.1 Flash 1017 | 92.2 |
| GLM-5.3 | 92.2 |
| Kimi K3 | 92.2 |
| Grok 4.6 | 91.8 |
| Muse Spark 1.1 | 91.8 |
| Hy4 Preview | 91.5 |
| Qwen3.8-Max | 90.8 |
| DeepSeek-V4-Flash-0731 | 90.4 |
| DeepSeek V4 Pro 0813 | 90.4 |
| MiMo-V2.6-Flash | 90.4 |
| Mistral Large 4 Preview | 90.4 |
| Ember-1 | 90.1 |
| Gemini 3.8 Flash | 89.8 |
| Qwen3.8 27B | 89.8 |
| Opus 4.8 | 89.1 |
| Claude Opus 5 | 89.1 |
| GPT-6.1 Sol | 89.1 |
| Claude Haiku 5.5 | 88.7 |
| Claude Sonnet 5.5 | 88.7 |
| GPT-6 Astra | 88.7 |
| GPT-6 Sol | 88.7 |
| DeepSeek V4.1 Flash | 87.7 |
| Opus 4.7 | 87.4 |
| DeepSeek-V4-Pro | 87.4 |
| GLM-5.2 | 87.4 |
| GPT-5.6 Luna | 87.0 |
| Gemini 3.7 Flash | 86.7 |
| GPT-6 Luna | 86.3 |
| Grok 4.5 | 86.3 |
| Sonnet 5 | 85.7 |
| GPT-5.6 Sol | 85.3 |