List your company

Mathematical Reasoning on MATH 500 (M→M-500 and G8→M-500 Accuracy)

78.2Accuracy (M→M-500)GRPO
8.5226.6144.762.79Dec 30, 2025

Evaluation results

MethodLinks
GRPOModel Size=7B2025.1278.245.6
iCLPModel Size=7B2025.1268.646.2
GRPOModel Size=3B2025.1260.840.4
FT-SModel Size=7B2025.1258.444
iCLPModel Size=3B2025.1255.444.2
FT-EModel Size=7B2025.125243.8
FT-SModel Size=3B2025.1248.440
FT-EModel Size=3B2025.124339.8
BaseModel Size=7B2025.1242.442.4
BaseModel Size=3B2025.1239.639.6
GRPOModel Size=0.5B2025.1234.417.9
iCLPModel Size=0.5B2025.1228.120.8
FT-SModel Size=0.5B2025.1227.218.4
FT-EModel Size=0.5B2025.1218.815.8
BaseModel Size=0.5B2025.1215.815.8
FT-LEModel Size=0.5B2025.1211.29.6