MMLU Pro - Psychology
General QA · 2024-06-03
The Psychology 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 | 92.5 |
| Claude Fable 5 | 92.2 |
| Gemini 3.1 Pro Preview | 92.0 |
| Opus 4.7 | 91.2 |
| Gemini 3 Pro Preview | 91.0 |
| Gemini 3.5 Flash | 90.4 |
| Grok 4.5 | 90.2 |
| Opus 4.8 | 90.0 |
| Qwen3.7-Max | 90.0 |
| Opus 4.6 | 89.8 |
| Gemini 3.6 Flash | 89.5 |
| Opus 4.1 | 89.3 |
| GPT-5.6 Sol | 89.2 |
| Kimi K2.6 | 89.2 |
| Gemini 3 Flash Preview | 88.8 |
| Opus 4.5 | 88.7 |
| Sonnet 5 | 88.7 |
| GPT-5.5 | 88.5 |
| Kimi K3 | 88.3 |
| Opus 4 | 88.2 |
| DeepSeek-V4-Pro | 88.0 |
| MiniMax M2.1 | 88.0 |
| Sonnet 4.5 | 87.8 |
| GPT-5.4 | 87.8 |
| Grok 4.3 | 87.7 |
| GPT-5 | 87.5 |
| Qwen3.6 Plus (2026-04-02) | 87.5 |
| Sonnet 4.6 | 87.3 |
| GLM-5.1 | 87.2 |
| Muse Spark 1.1 | 87.2 |
| GPT-5.1 | 87.1 |
| Qwen3.5 Plus (2026-02-15) | 87.0 |
| Inkling | 86.6 |
| Muse Spark | 86.5 |
| o3 | 86.5 |
| GPT-5.6 Terra | 86.3 |
| GLM-5.2 | 86.2 |
| GPT-5.2 | 86.1 |
| Kimi K2.5 | 86.1 |
| MiMo-V2.5-Pro | 86.1 |