List your company

Scientific Reasoning on GPQA D (Mean@16, Pass@16)

27.28Mean@16ADS
8.986413.735718.48523.2343Jun 21, 2026

Evaluation results

MethodLinks
ADSBackbone=Qwen2.5-Math-1.5B2026.0627.2889.39
DOTS+RRBackbone=Qwen2.5-Math-1.5B2026.0620.7877.61
ADSBackbone=Qwen3-4B-Base2026.0620.6971.55
GRPOBackbone=Qwen2.5-Math-1.5B2026.0620.1487.03
ADSBackbone=Olmo3-7B-SFT2026.0617.8469.36
GRPOBackbone=Olmo3-7B-SFT2026.0616.5669.53
GRPOBackbone=Qwen3-4B-Base2026.0616.1169.7
DOTS+RRBackbone=Olmo3-7B-SFT2026.0613.3463.47
BaseBackbone=Qwen2.5-Math-1.5B2026.0612.5880.13
BaseBackbone=Olmo3-7B-SFT2026.0611.5356.91
DOTS+RRBackbone=Qwen3-4B-Base2026.0610.8959.09
BaseBackbone=Qwen3-4B-Base2026.069.6962.79