List your company

Mathematical Reasoning on AIME 25 (P@4, #Tok)

56.7P@4 AccuracyMPD
46.34951.754.4May 9, 2026

Evaluation results

MethodLinks
MPD2026.0556.712,900
CRISP2026.0556.616,300
Vanilla LLM2026.0553.317,700
Direct Comp.2026.055015,200
Chain-of-Draft2026.055013,000
LiteCoT2026.0546.714,600