List your company

Mathematical Reasoning on GSM8K (P@4, #Tok)

94.5Pass@4Vanilla LLM
40.73254.69168.6582.609May 9, 2026

Evaluation results

MethodLinks
Vanilla LLM2026.0594.52,100
CRISP2026.0594.31,700
MPD2026.0594.21,400
Direct Comp.2026.0594.11,200
LiteCoT2026.05521,200
Chain-of-Draft2026.0542.8600