List your company

Math Reasoning on AIME 25 (Avg@32, Pass@16)

31.15Avg@32 Scoretraining-time reweighting
4.713211.576618.4425.3034Mar 23, 2026

Evaluation results

MethodLinks
training-time reweightingModel=Qwen3-8B-Base2026.0331.1555.38
DAPOModel=Qwen3-8B-Base2026.0326.6746.76
training-time reweightingModel=Qwen2.5-Math-7B2026.0318.5436.72
DAPOModel=Qwen2.5-Math-7B2026.0317.630.45
BaseModel=Qwen2.5-Math-7B2026.036.6727.84
BaseModel=Qwen3-8B-Base2026.035.7332.8