Harvey's Legal Agent Benchmark - Corporate M&A
Professional Work · 2026-05-06
The Corporate M&A split of Vals' archived v1 Harvey Legal Agent Benchmark held-out evaluation. Its task score averages two judges' all-criteria-pass decisions rather than counting binary task successes.
Top models (higher is better)
| Model | Score |
|---|---|
| Opus 4.7 | 20.0 |
| Claude Opus 5 | 20.0 |
| DeepSeek V4.1 Flash 1017 | 20.0 |
| DeepSeek V4.1 Flash | 20.0 |
| DeepSeek-V4-Flash-0731 | 20.0 |
| DeepSeek V4 Pro 0813 | 20.0 |
| DeepSeek-V4-Pro | 20.0 |
| Gemini 4 Argon | 20.0 |
| GLM-5.2 | 20.0 |
| GLM-5.3 Flash | 20.0 |
| GPT-6.1 Sol | 20.0 |
| GPT-6 Astra | 20.0 |
| GPT-6 Luna | 20.0 |
| Grok 4.6 | 20.0 |
| Grok 4.7 | 20.0 |
| Hy4 Preview | 20.0 |
| MiMo-V2.6-Flash | 20.0 |
| MiMo-V2.6-Pro | 20.0 |
| Mistral Large 4 Preview | 20.0 |
| Muse Spark 1.1 | 20.0 |
| Muse Spark 1.2 | 20.0 |
| Muse Spark 1.3 | 20.0 |
| Muse Spark Open | 20.0 |
| Muse Wonder Full | 20.0 |
| Muse Wonder | 20.0 |
| Qwen3.8 27B | 20.0 |
| Qwen3.8-Max | 20.0 |
| Step 5 Preview | 20.0 |
| Haiku 4.5 | 0.0 |
| Claude Haiku 5.5 | 0.0 |
| Opus 4.8 | 0.0 |
| Claude Opus 5.5 | 0.0 |
| Sonnet 4.6 | 0.0 |
| Claude Sonnet 5.5 | 0.0 |
| Sonnet 5 | 0.0 |
| Ember-1 | 0.0 |
| Gemini 3.1 Flash-Lite Preview | 0.0 |
| Gemini 3.1 Pro Preview | 0.0 |
| Gemini 3.5 Flash | 0.0 |
| Gemini 3.5 Flash-Lite | 0.0 |