List your company

Mathematical Reasoning on AIME 2026 (val)

70.7AccuracyDeepConf
50.9456.0761.266.33Jun 7, 2026

Evaluation results

MethodLinks
DeepConfSample size (N)=32, Max length=16,000 tokens, Selection variant=best over each dataset2026.0670.7
Self-ConsistencySample size (N)=32, Max length=16,000 tokens2026.0665.8
iS@1Sample size (N)=32, Max length=16,000 tokens, Top-k=12026.0664.2
Self-CertaintySample size (N)=32, Max length=16,000 tokens, Selection variant=best over each dataset2026.0651.7