List your company

Mathematical Reasoning on MATH 500 (pass@k)

45.19Pass@1LP-SFT
3.142814.058924.97535.8911Jul 6, 2026

Evaluation results

MethodLinks
LP-SFTBackbone=Qwen3-14B2026.0745.1977.8
ASFTBackbone=Qwen3-14B2026.0743.2370.6
LP-SFTBackbone=Qwen3-4B2026.0742.7976.8
ASFTBackbone=Qwen3-4B2026.0740.7170
CEBackbone=Qwen3-14B2026.0739.3974.2
EAFTBackbone=Qwen3-14B2026.0739.174.6
DFTBackbone=Qwen3-14B2026.0737.5954.6
GEMBackbone=Qwen3-14B2026.0737.3471.2
EAFTBackbone=Qwen3-4B2026.0735.8873
CEBackbone=Qwen3-4B2026.0735.672.6
DFTBackbone=Qwen3-4B2026.0734.3152.4
GEMBackbone=Qwen3-4B2026.0733.2672.6
CEBackbone=Llama3.1-8B2026.079.9542.6
LP-SFTBackbone=Llama3.1-8B2026.078.4840
EAFTBackbone=Llama3.1-8B2026.078.3540
GEMBackbone=Llama3.1-8B2026.078.1941.4
ASFTBackbone=Llama3.1-8B2026.077.933.4
DFTBackbone=Llama3.1-8B2026.074.7614.6