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

Mathematical Reasoning on MATH 500 (ACC, Pass@8)

96.4Accuracy

GRPO + SALT

84.33687.46890.693.732Jun 4, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
96.499.2
2026.06
95.499.3
2026.06
95.298.9
2026.06
95.299.1
2026.06
9599.1
2026.06
94.998.8
2026.06
94.999
2026.06
94.499.1
2026.06
94.299
2026.06
8897.3
2026.06
8898
2026.06
87.298.4
2026.06
87.197
2026.06
86.197
2026.06
85.996.8
2026.06
85.696.7
2026.06
85.596.4
2026.06
84.896.8