List your company

Code Generation on LiveCodeBench (OOD)

19.2AccuracyRL-PLUS
7.65610.65313.6516.647Jul 31, 2025

Evaluation results

MethodLinks
RL-PLUSBase model=Qwen2.5-Math-7B2025.0719.2
GRPOBase model=Qwen2.5-Math-7B2025.0715.3
Base ModelBase model=Qwen2.5-Math-7B2025.0714.9
SFT+GRPOBase model=Qwen2.5-Math-7B2025.079.7
SFTBase model=Qwen2.5-Math-7B2025.078.1