List your company

Mathematical Reasoning on AMC (pass@1, pass@16, maj@16)

61Pass@1DiRL
36.66442.98249.355.618Jun 9, 2026

Evaluation results

MethodLinks
DiRLBase Model=Qwen3-4B-Base2026.066175.694.5
G2RLBase Model=Qwen3-4B-Base2026.0654.168.490.4
Entropy BonusBase Model=Qwen3-4B-Base2026.0653.767.888.8
EVOL-RLBase Model=Qwen3-4B-Base2026.0653.167.187.8
GRPOBase Model=Qwen3-4B-Base2026.065366.386.4
DiRLBase Model=Qwen3-1.7B-Base2026.0643.655.583.4
G2RLBase Model=Qwen3-1.7B-Base2026.0640.553.882.5
Entropy BonusBase Model=Qwen3-1.7B-Base2026.0638.452.974.4
EVOL-RLBase Model=Qwen3-1.7B-Base2026.0638.253.179.4
GRPOBase Model=Qwen3-1.7B-Base2026.0637.65377.2