Legal Research Bench Health
Professional Work · 2026-06-23
This row reports the practice-area score for Health tasks in Vals AI Legal Research Bench.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.6 Sol | 80.0 |
| Claude Opus 5 | 76.7 |
| GPT-5.6 Terra | 76.7 |
| GPT-5.6 Luna | 70.0 |
| Claude Fable 5 | 66.7 |
| GPT-5.5 | 66.7 |
| Grok 4.5 | 66.7 |
| Sonnet 5 | 63.3 |
| Kimi K3 | 63.3 |
| Opus 4.8 | 53.3 |
| Sonnet 4.6 | 53.3 |
| GLM-5.2 | 50.0 |
| DeepSeek-V4-Pro | 46.7 |
| MiniMax M3 | 46.7 |
| Gemini 3.5 Flash | 43.3 |
| Inkling | 43.3 |
| Grok 4.3 | 40.0 |
| Muse Spark 1.1 | 40.0 |
| Qwen3.7-Max | 40.0 |
| Gemini 3.6 Flash | 36.7 |
| GLM-5.1 | 36.7 |
| Gemini 3.1 Pro Preview | 33.3 |
| Inkling-Small | 30.0 |
| Kimi K2.6 | 30.0 |
| GPT-5.4 Mini | 16.7 |
| Gemini 3.5 Flash-Lite | 13.3 |
| Laguna M.1 | 3.3 |
| Laguna XS.2 | 0.0 |