List your company

Critique on CriticBench (test)

94.1Math ScoreQwen3-32B
89.638490.796791.95593.1133May 14, 2026

Evaluation results

MethodLinks
Qwen3-32BModel=Qwen3-32B2026.0594.167.3694.4156.1278
CIPOBase Model=Qwen3-4B, Training=+CIPO2026.059461.7490.4255.9475.53
GRPOBase Model=Qwen3-4B, Training=+GRPO2026.0593.8761.5590.5154.7975.18
InitialBase Model=Qwen3-4B2026.0592.8258.2283.6850.7271.36
Qwen2.5-72B-InstructModel=Qwen2.5-72B-Instruct2026.0589.8164.9491.9963.5877.58