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

Math Reasoning on AIME24

76.7Accuracy

DiffMAS

Updated 1mo ago

Evaluation Results

Method	Links
DiffMAS 2026.04		76.7
DiffMAS 2026.04		73.3
DiffMAS 2026.04		70
LatentMAS 2026.04		66.7
Single 2026.04		66.7
CE-GPPO 2025.09		66
Single 2026.04		63.3
TextMAS 2026.04		63.3
CE-GPPO 2025.09		62.3
DAPO 2025.09		59.7
GRPO 2025.09		55.3
DS-R1-Distill-Qwen-7B 2025.09		54.5
Single 2026.04		50
TextMAS 2026.04		50
LatentMAS 2026.04		50
CE-GPPO 2025.09		43.6
CE-GPPO 2025.09		42
DAPO 2025.09		40
GRPO 2025.09		33.4
LatentMAS 2026.04		33.3
DS-R1-Distill-Qwen-1.5B 2025.09		29.2
TextMAS 2026.04		26.7