Inference
Xiaoxuan Du et al.

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.

ModelsModels with a result.20
Top resultBest result on this test.95.0%Claude Opus 4.6
Top-3 spreadPoints from first to third.21.1 pts
YearYear of release.2025

Result and price

020406080100$0.3$1$3$10$30RESULTOUTPUT PRICE PER 1M TOKENS · LOG SCALE

Results

20 results
#ModelResultThe score from the source.
01Claude Opus 4.695.0%
01Claude Sonnet 4.695.0%
03Qwen 3.6 Max (preview)73.9%
04Qwen3.7 Max73.6%
05Qwen3.6 Plus71.6%
06Qwen3.7 Plus71.4%
07Seed 2.1 Pro70.8%
08Claude Opus 4.570.6%
09Qwen3.5 397B70.4%
10Kimi K2.569.2%
11Seed 2.1 Turbo67.4%
12Qwen3.5-122B-A10B67.1%
13GLM-566.8%
14Qwen3.6-27B66.0%
15Qwen3.5-27B65.6%
16Qwen3.6-35B-A3B64.7%
17Qwen3.5-35B-A3B63.4%
18Qwen3 235B 250762.6%
19MiniCPM5-2B40.8%
20MiniCPM5-1B23.1%