List your company

Mathematical Reasoning on Brumo 2025 (Pass@1, Efficiency Ratio, MV Budget)

96.4Pass@1PC-cubic
65.7273.68581.6589.615May 8, 2026

Evaluation results

MethodLinks
PC-cubicModel=Nemotron3-30B, Token Budget (B)=5M2026.0596.4———
Standard MVModel=Nemotron3-30B, Token Budget (B)=250k2026.0589.6———
PC-cubicModel=GPT-OSS-120B, Token Budget (B)=5M2026.0585.7———
PC-cubicModel=Nemotron3-30B, Target accuracy (alpha)=90%2026.0581.60.393.2651
Standard MVModel=GPT-OSS-120B, Token Budget (B)=250k2026.0580.1———
PC-cubicModel=Ministral3-14B, Token Budget (B)=5M2026.0575.9———
AC sweepModel=GPT-OSS-120B, Target accuracy (alpha)=99%2026.0571.30.283.33.3
Standard MVModel=Ministral3-14B, Token Budget (B)=250k2026.0566.9———