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 |
|---|---|
| Claude Opus 5 | 90.6 |
| GPT-5.6 Sol | 87.8 |
| Claude Fable 5 | 87.8 |
| Kimi K3 | 87.0 |
| Opus 4.8 | 86.1 |
| Sonnet 4.6 | 85.3 |
| GPT-5.6 Luna | 85.0 |
| Muse Spark 1.1 | 84.6 |
| Sonnet 5 | 84.6 |
| Grok 4.5 | 84.4 |
| GPT-5.6 Terra | 84.0 |
| GPT-5.5 | 84.0 |
| Gemini 3.5 Flash | 83.9 |
| GLM-5.2 | 82.9 |
| MiniMax M3 | 81.1 |
| DeepSeek-V4-Pro | 80.7 |
| Gemini 3.6 Flash | 80.5 |
| Inkling | 80.2 |
| Inkling-Small | 78.9 |
| Qwen3.7-Max | 78.4 |
| GLM-5.1 | 77.8 |
| Gemini 3.1 Pro Preview | 76.6 |
| Kimi K2.6 | 74.7 |
| Grok 4.3 | 69.5 |
| Gemini 3.5 Flash-Lite | 68.9 |
| GPT-5.4 Mini | 67.3 |
| Laguna M.1 | 62.3 |
| Laguna XS.2 | 56.3 |