SuperGPQA — Results

Rank	Model	Score
1	deepseek-v4-1-flash	53.1
2	minicpm5-2b	40.8
3	qwen3-7-max	0.736
4	qwen3-6-plus	0.716
5	qwen3-7-plus	0.714
6	seed-2-1-pro	0.708
7	qwen3-5-397b-a17b	0.704
8	seed-2-1-turbo	0.674
9	qwen3-5-122b-a10b	0.671
10	qwen3-6-27b	0.66
11	qwen3-5-27b	0.656
12	qwen3-235b-a22b-thinking-2507	0.649
13	qwen3-6-35b-a3b	0.647
14	qwen3-vl-235b-a22b-thinking	0.643
15	qwen3-5-35b-a3b	0.634
16	qwen3-235b-a22b-instruct-2507	0.626
17	qwen3-next-80b-a3b-thinking	0.608
18	qwen3-vl-235b-a22b-instruct	0.604
19	qwen3-vl-32b-thinking	0.59
20	qwen3-next-80b-a3b-instruct	0.588
21	qwen3-5-9b	0.582
22	kimi-k2-instruct-0905	0.572
23	kimi-k2-instruct	0.572
24	qwen3-vl-30b-a3b-thinking	0.564
25	qwen3-vl-32b-instruct	0.546

S

SuperGPQA — Evaluations

1 phaseActive

No evaluations yet for this benchmark.