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

Math on AIME 2025 (accuracy (%))

69.9Accuracy (%)

VP2O

Updated 1mo ago

Evaluation Results

Method	Links
VP2O 2026.06		69.9
GRPO-baseline 2026.06		67.8
LLaDA-2.1-flash 2026.06		63.33
VP2O 2026.06		62.5
LLaDA-2.0-flash 2026.06		60
Qwen3.5-9B 2026.06		60
GRPO-baseline 2026.06		59.6
FLARE-9B 2026.06		54.44
FLARE-9B 2026.06		53.33
LLaDA-2.0-mini 2026.06		36.67
LLaDA-2.1-mini 2026.06		36.67
Original (T2T) 2026.04		33.33
T2M 2026.04		33.33
TIES 2026.06		20
RAM 2026.06		16.67
RESMERGE 2026.06		16.67
DARE + TIES 2026.06		13.33
TSV-Merge 2026.06		13.33
SDAR 30B-A3B 2026.06		10.8
SimpleRL 2026.06		10
Zero 2026.06		10
Reasoner 2026.06		10
ISO-C 2026.06		10
ISO-CTS 2026.06		10
TA 2026.06		3.33
Qwen2.5-7B 2026.06		0