List your company

Mathematical Reasoning on AIME 2024 (Accuracy, ECE, Brier, AUROC)

20.8AccuracyCAR (MSCR)
18.09618.79819.520.202Jan 12, 2026

Evaluation results

MethodLinks
CAR (MSCR)Model=Qwen2.5-3B2026.0120.80.5730.4850.548
Vanilla SimpleTIRModel=Qwen2.5-3B2026.0118.20.6920.630.489