Benchgen

SuperGPQA — Results

RankModelScore
1qwen3-7-max0.736
2qwen3-6-plus0.716
3qwen3-7-plus0.714
4seed-2-1-pro0.708
5qwen3-5-397b-a17b0.704
6seed-2-1-turbo0.674
7qwen3-5-122b-a10b0.671
8qwen3-6-27b0.66
9qwen3-5-27b0.656
10qwen3-235b-a22b-thinking-25070.649
11qwen3-6-35b-a3b0.647
12qwen3-vl-235b-a22b-thinking0.643
13qwen3-5-35b-a3b0.634
14qwen3-235b-a22b-instruct-25070.626
15qwen3-next-80b-a3b-thinking0.608
16qwen3-vl-235b-a22b-instruct0.604
17qwen3-vl-32b-thinking0.59
18qwen3-next-80b-a3b-instruct0.588
19qwen3-5-9b0.582
20kimi-k2-instruct-09050.572
21kimi-k2-instruct0.572
22qwen3-vl-30b-a3b-thinking0.564
23qwen3-vl-32b-instruct0.546
24qwen3-vl-30b-a3b-instruct0.531
25qwen3-5-4b0.529
S

SuperGPQA — Phases

1 phaseActive

Phase 1·

SuperGPQA

Aug 3, 2026
Current