List your company

Mathematical Problem Solving on AIME 2024 (Avg@8)

84.17Avg@8 ScoreEvoTrainer
77.233279.034180.83582.6359Jun 2, 2026

Evaluation results

MethodLinks
EvoTrainerConfiguration=Proposed Method2026.0684.17
RAGEN v2 SNR FilteringConfiguration=Algorithmic Baseline2026.0681.67
Human-engineered RLConfiguration=Reference2026.0680.83
RAGEN v1 FilteringConfiguration=Algorithmic Baseline2026.0680.42
GRPO + Clip-HigherConfiguration=Algorithmic Baseline2026.0680
GSPO + Clip-HigherConfiguration=Algorithmic Baseline2026.0679.17
Group-Variance FilteringConfiguration=Algorithmic Baseline2026.0678.75
AutoResearchConfiguration=Reference2026.0678.33
Base model (no RL)RL=None2026.0677.5