List your company

Mathematical Reasoning on GSM8K (Pass@1 Accuracy & Step Count)

86.5Pass@1 AccuracyHEX
71.919275.704679.4983.2754Feb 12, 2026

Evaluation results

MethodLinks
HEXGen. Len.=256, Training Necessity=Training-free2026.0286.53,200
Majority VotingGen. Len.=256, Training Necessity=Training-free2026.0284.69640
DVOTINGGen. Len.=256, Training Necessity=Training-free2026.0284.53223.1
HEXGen. Len.=128, Training Necessity=Training-free2026.0282.411,600
RFGGen. Len.=256, Training Necessity=FT or RL2026.0282.1512
DVOTINGGen. Len.=128, Training Necessity=Training-free2026.0279.15158.8
Majority VotingGen. Len.=128, Training Necessity=Training-free2026.0278.47320
LLaDA-1.5*Gen. Len.=256, Training Necessity=None2026.0278.01256
LLaDA-1.5*Gen. Len.=128, Training Necessity=None2026.0272.48128