List your company

Mathematical Reasoning on AIME 2025 (Performance %)

9.58AIME 2025 PerformanceGT
3.084.76756.4558.1425May 25, 2026

Evaluation results

MethodLinks
GTBackbone=Phi4-mini-instruct2026.059.58
CAREBackbone=Phi4-mini-instruct2026.057.81
TTRLBackbone=Phi4-mini-instruct2026.056.67
OracleBackbone=Phi4-mini-instruct2026.055.94
RandomBackbone=Phi4-mini-instruct2026.055.83
EntropyBackbone=Phi4-mini-instruct2026.055.83
ProbBackbone=Phi4-mini-instruct2026.054.9
VanillaBackbone=Phi4-mini-instruct2026.053.33