List your company

Scientific Reasoning on GPQA-D (Pass@1, #Tok)

68.2Pass@1 AccuracyDyCon
37.20845.25453.361.346Jun 5, 2026

Evaluation results

MethodLinks
DyConbackbone=Qwen3-4B-Thinking-25072026.0668.26,205
Baselinebackbone=QwQ-32B2026.0667.77,732
DyConbackbone=QwQ-32B2026.0667.75,699
Baselinebackbone=Qwen3-4B-Thinking-25072026.0666.29,210
DyConbackbone=Qwen3-14B2026.0665.26,180
Baselinebackbone=Qwen3-14B2026.0665.17,654
DyConbackbone=DeepSeek-R1-Distill-Qwen-7B2026.06475,304
Baselinebackbone=DeepSeek-R1-Distill-Qwen-7B2026.0638.47,518