MMLU Pro - History
General QA · 2024-06-03
The History 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 |
| Gemini 3.1 Pro Preview | 85.6 |
| Claude Fable 5 | 85.3 |
| Gemini 3 Pro Preview | 84.8 |
| Gemini 3.6 Flash | 83.5 |
| Opus 4.7 | 83.2 |
| Opus 4.6 | 82.7 |
| Opus 4.8 | 82.7 |
| Opus 4.1 | 82.4 |
| Qwen3.7-Max | 82.2 |
| Sonnet 4.5 | 81.9 |
| GPT-5.6 Sol | 81.9 |
| Gemini 3.5 Flash | 81.6 |
| GPT-5.5 | 81.6 |
| Grok 4.5 | 81.4 |
| Sonnet 4.6 | 81.1 |
| Qwen3.6 Plus (2026-04-02) | 81.1 |
| Gemini 3 Flash Preview | 80.8 |
| Muse Spark | 80.8 |
| Opus 4 | 80.3 |
| Opus 4.5 | 80.3 |
| Sonnet 5 | 80.3 |
| Muse Spark 1.1 | 79.8 |
| GPT-5.6 Terra | 79.5 |
| MiniMax M2.1 | 79.5 |
| Sonnet 4 | 79.3 |
| GPT-5.4 | 79.3 |
| GPT-5 | 79.3 |
| Kimi K3 | 79.3 |
| Qwen3.5 Plus (2026-02-15) | 78.7 |
| Gemini 3.5 Flash-Lite | 78.5 |
| Kimi K2.6 | 78.5 |
| GLM-5.1 | 78.4 |
| Gemini 3.1 Flash-Lite Preview | 78.2 |
| Grok 4.3 | 78.0 |
| o3 | 78.0 |
| DeepSeek-V4-Pro | 77.7 |
| GPT-5.1 | 77.7 |
| GLM-5 | 77.4 |
| Grok 4.20 | 77.4 |