List your company

Multi-task Evaluation Summary on Aggregate (MATH-500, AIME, MBPP+, HE+)

63.5Average Performance (Aggregate)LP-SFT
19.97631.275542.57553.8745Jul 6, 2026

Evaluation results

MethodLinks
LP-SFTModel=Qwen3-14B2026.0763.5
CEModel=Qwen3-14B2026.0761.53
EAFTModel=Qwen3-14B2026.0761.42
GEMModel=Qwen3-14B2026.0761.38
ASFTModel=Qwen3-14B2026.0760.47
LP-SFTModel=Qwen3-4B2026.0760.15
DFTModel=Qwen3-14B2026.0758.51
EAFTModel=Qwen3-4B2026.0758.3
GEMModel=Qwen3-4B2026.0758.3
CEModel=Qwen3-4B2026.0757.66
ASFTModel=Qwen3-4B2026.0755.31
DFTModel=Qwen3-4B2026.0753.05
LP-SFTModel=Llama3.1-8B2026.0735.49
GEMModel=Llama3.1-8B2026.0735.24
CEModel=Llama3.1-8B2026.0734.88
EAFTModel=Llama3.1-8B2026.0733.35
ASFTModel=Llama3.1-8B2026.0730.55
DFTModel=Llama3.1-8B2026.0721.65