List your company

Code Generation on APPS+ Interview

12.14Pass@1WITH GRPO ADVANTAGE
6.80488.18999.57510.9601May 27, 2026

Evaluation results

MethodLinks
WITH GRPO ADVANTAGEBackbone=Qwen-2.5-Coder, Parameters=7B2026.0512.1430.2
SFTBackbone=Qwen-2.5-Coder, Parameters=7B2026.0511.330.3
ADVATNAGE WITH EXP()Backbone=Qwen-2.5-Coder, Parameters=7B2026.0511.0930.3
RLOOBackbone=Qwen-2.5-Coder, Parameters=7B2026.0510.3127.8
BASEBackbone=Qwen-2.5-Coder, Parameters=7B2026.057.0125.7