List your company

Mathematical Reasoning on AIME 25 (SC Acc, PoLR Diff, Efficiency, Overhead)

0.767SC AccuracySelf-Consistency
0.302120.422810.54350.66419Jan 29, 2026

Evaluation results

MethodLinks
Self-ConsistencyModel=QWQ32B, Sample size N=512026.010.767———
PoLRModel=QWQ32B, Sample size N=512026.010.767-1056.812.3
Self-ConsistencyModel=QWQ32B, Sample size N=312026.010.75———
PoLRModel=QWQ32B, Sample size N=312026.010.75-459.56.9
Self-ConsistencyModel=QWQ32B, Sample size N=112026.010.747———
PoLRModel=QWQ32B, Sample size N=112026.010.747-665.35.3
Self-ConsistencyModel=Phi-4-15B, Sample size N=112026.010.357———
PoLRModel=Phi-4-15B, Sample size N=112026.010.3572.360.52.3
Self-ConsistencyModel=DSQ7B, Sample size N=312026.010.353———
PoLRModel=DSQ7B, Sample size N=312026.010.353048.43.9
Self-ConsistencyModel=DSQ7B, Sample size N=112026.010.337———
PoLRModel=DSQ7B, Sample size N=112026.010.3372.748.92.2
Self-ConsistencyModel=DSQ7B, Sample size N=512026.010.333———
PoLRModel=DSQ7B, Sample size N=512026.010.333048.87.8
Self-ConsistencyModel=Phi-4-15B, Sample size N=512026.010.333———
PoLRModel=Phi-4-15B, Sample size N=512026.010.333054.711
Self-ConsistencyModel=Phi-4-15B, Sample size N=312026.010.32———
PoLRModel=Phi-4-15B, Sample size N=312026.010.32454.85.9