Legal Research Bench Weighted Pass
Professional Work · 2026-06-23
This row reports the practice-area score for Weighted Pass tasks in Vals AI Legal Research Bench.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 4 Carbon | 91.6 |
| Gemini 4 Argon | 91.0 |
| Claude Opus 5 | 90.6 |
| Muse Wonder Full | 89.8 |
| Grok 4.6 | 88.3 |
| GPT-5.6 Sol | 87.8 |
| GLM-5.3 | 87.7 |
| Muse Wonder | 87.7 |
| GLM-5.3 Flash | 87.5 |
| Gemini 3.8 Flash | 87.5 |
| Muse Spark 1.3 | 86.8 |
| Grok 4.7 | 86.6 |
| DeepSeek V4.1 Flash 1017 | 86.3 |
| Qwen3.8-Max | 86.1 |
| Opus 4.8 | 86.1 |
| Ember-1 | 86.0 |
| Muse Spark Open | 86.0 |
| MiMo-V2.6-Pro | 85.5 |
| Sonnet 4.6 | 85.3 |
| Mistral Large 4 Preview | 85.3 |
| Muse Spark 1.2 | 85.3 |
| Hy4 Preview | 85.2 |
| GPT-5.6 Terra | 85.1 |
| GPT-5.6 Luna | 85.0 |
| GPT-6 Astra | 84.8 |
| Claude Haiku 5.5 | 84.8 |
| Step 5 Preview | 84.8 |
| Kimi K3 | 84.6 |
| Muse Spark 1.1 | 84.6 |
| Sonnet 5 | 84.6 |
| DeepSeek V4.1 Flash | 84.5 |
| Grok 4.5 | 84.4 |
| GPT-6.1 Sol | 84.3 |
| Gemini 3.7 Flash | 84.1 |
| Opus 4.7 | 84.1 |
| GPT-5.5 | 84.0 |
| Gemini 3.5 Flash | 83.9 |
| DeepSeek V4 Pro 0813 | 83.8 |
| MiniMax M3.1 Flash | 83.1 |
| GLM-5.2 | 82.9 |