Skip to content

SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines

An expanded version of GPQA that evaluates graduate-level knowledge and reasoning capabilities across 285 disciplines, providing comprehensive coverage of academic domains.

20 models

Higher is better.

  1. 1Claude Opus 4.6Anthropic95.0
  2. 2Claude Sonnet 4.6Anthropic95.0
  3. 3Qwen 3.6 Max (preview)Alibaba73.9
  4. 4Qwen3.7 MaxAlibaba73.6
  5. 5Qwen3.6 PlusAlibaba71.6
  6. 6Qwen3.7 PlusAlibaba71.4
  7. 7Seed 2.1 ProByteDance70.8
  8. 8Claude Opus 4.5Anthropic70.6
  9. 9Qwen3.5 397BAlibaba70.4
  10. 10Kimi K2.5Moonshot AI69.2
  11. 11Qwen3.5-122B-A10BAlibaba67.1
  12. 12GLM-5Z.AI66.8
  13. 13Qwen3.6-27BAlibaba66.0
  14. 14Qwen3.5-27BAlibaba65.6
  15. 15Qwen3.6-35B-A3BAlibaba64.7
  16. 16Qwen3.5-35B-A3BAlibaba63.4
  17. 17Qwen3 235B 2507Alibaba62.6
  18. 18Kimi K2Moonshot AI57.2
  19. 19MiniCPM5-2BOpenBMB40.8
  20. 20MiniCPM5-1BOpenBMB23.1
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Multiple choice questions
Tasks
285 disciplines
Authors
Xiaoxuan Du, Yao Yao, Kexin Ma, Bowen Wang, Tianyu Zheng, Kaiyan Zhu, Yiming Zhang, Yutao Zhu, Jiawei Zhou, Jingren Zhou (2025)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.