Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Mathematical Reasoning on AIME 25 (ACC, Pass@8)

46.5ACC

GRPO + SALT

21.0227.63534.2540.865Jun 4, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
46.575.5
2026.06
46.368.3
2026.06
45.374.5
2026.06
44.866.6
2026.06
4364
2026.06
42.872.1
2026.06
4271.3
2026.06
4262.7
2026.06
36.764
2026.06
2847
2026.06
27.847.8
2026.06
27.147.2
2026.06
2645
2026.06
25.643.3
2026.06
25.646.3
2026.06
23.944.3
2026.06
23.644.2
2026.06
2242.6