MMLU Pro - Law
General QA · 2024-06-03
The Law split of MMLU Pro. This child benchmark separates a source-reported subtask or subtrack from the parent aggregate so scores at different grains do not share one benchmark_slug.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5 | 86.6 |
| Claude Fable 5 | 86.5 |
| Gemini 3.1 Pro Preview | 84.1 |
| Gemini 3 Pro Preview | 82.5 |
| Opus 4.7 | 82.1 |
| Grok 4.5 | 82.0 |
| Opus 4.8 | 81.9 |
| Gemini 3.5 Flash | 80.4 |
| Gemini 3.6 Flash | 80.1 |
| Opus 4.1 | 79.9 |
| Opus 4.6 | 79.7 |
| Qwen3.7-Max | 79.5 |
| GPT-5.6 Sol | 79.1 |
| MiniMax M2.1 | 78.9 |
| Sonnet 5 | 78.7 |
| Opus 4.5 | 78.7 |
| Gemini 3 Flash Preview | 78.5 |
| Kimi K3 | 77.0 |
| GPT-5.5 | 76.9 |
| Muse Spark 1.1 | 76.5 |
| GPT-5.1 | 75.8 |
| Qwen3.6 Plus (2026-04-02) | 75.7 |
| GPT-5.4 | 75.5 |
| GPT-5 | 75.3 |
| Gemini 3.1 Flash-Lite Preview | 75.0 |
| Muse Spark | 75.0 |
| Sonnet 4.5 | 74.9 |
| Opus 4 | 74.6 |
| GPT-5.6 Terra | 74.4 |
| Grok 4.3 | 74.4 |
| Kimi K2.6 | 74.3 |
| DeepSeek-V4-Pro | 74.0 |
| GLM-5.1 | 74.0 |
| Qwen3.5 Plus (2026-02-15) | 73.9 |
| Sonnet 4.6 | 73.6 |
| GPT-5.2 | 73.5 |
| o3 | 73.3 |
| GPT-5.6 Luna | 73.0 |
| Grok 4.20 | 72.9 |
| Nemotron 3 Ultra 550B A55B | 72.9 |