Wikibench

SuperGPQA

math

SuperGPQA is a comprehensive benchmark that evaluates large language models across 285 graduate-level academic disciplines. The benchmark contains 25,957 questions covering 13 broad disciplinary areas including Engineering, Medicine, Science, and Law, with specialized fields in light industry, agriculture, and service-oriented domains. It employs a Human-LLM collaborative filtering mechanism with over 80 expert annotators to create challenging questions that assess graduate-level knowledge and reasoning capabilities.

Leaderboard

Showing 20 of 31 results

  1. Qwen3.7 Max
    73.6%
    i
  2. Qwen3.6 Plus
    71.6%
    i
  3. Qwen3.5-397B-A17B
    70.4%
    i
  4. Qwen3.5-122B-A10B
    67.1%
    i
  5. Qwen3.6-27B
    66.0%
    i
  6. Qwen3.5-27B
    65.6%
    i
  7. Qwen3 Max
    65.1%
    i
  8. Qwen3-235B-A22B-Thinking-2507
    64.9%
    i
  9. Qwen3.6-35B-A3B
    64.7%
    i
  10. Qwen3 VL 235B A22B Thinking
    64.3%
    i
  11. Qwen3.5-35B-A3B
    63.4%
    i
  12. Qwen3-235B-A22B-Instruct-2507
    62.6%
    i
  13. Qwen3-Next-80B-A3B-Thinking
    60.8%
    i
  14. Qwen3 VL 235B A22B Instruct
    60.4%
    i
  15. Qwen3 VL 32B Thinking
    59.0%
    i
  16. Qwen3-Next-80B-A3B-Instruct
    58.8%
    i
  17. Qwen3.5-9B
    58.2%
    i
  18. Kimi K2 Instruct
    57.2%
    i
  19. Kimi K2-Instruct-0905
    57.2%
    i
  20. Qwen3 VL 30B A3B Thinking
    56.4%
    i