List your company

Mathematical Reasoning on AIME 2025 (avg@8, pass@8)

40.42Avg@8 ScoreMixed
33.9235.607537.29538.9825Jun 3, 2026

Evaluation results

MethodLinks
MixedFramework=TD-Grokking2026.0640.4260
DynamicFramework=TD-Grokking2026.0639.5863.33
SFTFramework=GRPO2026.0637.550
QuestAFramework=GRPO2026.0637.560
VanillaFramework=GRPO2026.0636.6756.67
Base2026.0635.8353.33
Scaf-GRPOFramework=GRPO2026.0634.5863.33
Sub-onlyFramework=TD-Grokking2026.0634.5850
AblationFramework=GRPO2026.0634.1750