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

Mathematical Problem Solving on AIME 2025 (Top-1 Accuracy)

91.67Top-1 Accuracy (%)

Majority Vote

Updated 23d ago

Evaluation Results

Method	Links
Majority Vote 2026.04		91.67	-	-	-	-
Self Refine 2026.04		90	-	-	-	-
ParaManager 2026.04		90	-	-	-	-
Majority Vote 2026.04		80	-	-	-	-
Base 2026.04		78.75	-	-	-	-
ToolOrchestra 2026.04		78.75	-	-	-	-
Self Refine 2026.04		76.67	-	-	-	-
Majority Vote 2026.04		76.67	-	-	-	-
Self Refine 2026.04		76.67	-	-	-	-
Puppeteer 2026.04		76.67	-	-	-	-
Base 2026.04		70	-	-	-	-
Base 2026.04		67.08	-	-	-	-
CIKA 2026.05		50	-	-	-	15
ePF w/ LaM 2025.10		26.97	-	-	-	-
ePF w/ LaM 2025.10		25.16	-	-	-	-
ePF 2025.10		23.13	-	-	-	-
ePF w/ LaM 2025.10		22.99	-	-	-	-
ePF 2025.10		22.06	-	-	-	-
ePF 2025.10		20.8	-	-	-	-
PF 2025.10		20.25	-	-	-	-
ePF w/ LaM 2025.10		19.6	-	-	-	-
Best-of-N 2025.10		19.22	-	-	-	-
PF 2025.10		19.1	-	-	-	-
PF 2025.10		19.09	-	-	-	-
ePF w/ LaM 2025.10		19.03	-	-	-	-
PF 2025.10		18.6	-	-	-	-
Best-of-N 2025.10		18.45	-	-	-	-
Best-of-N 2025.10		18.4	-	-	-	-
ePF 2025.10		18.4	-	-	-	-
ePF 2025.10		18.38	-	-	-	-
Best-of-N 2025.10		17.83	-	-	-	-
PF 2025.10		16.4	-	-	-	-
Best-of-N 2025.10		15.8	-	-	-	-
ePF w/ LaM 2025.10		13.55	-	-	-	-
ePF 2025.10		13.35	-	-	-	-
PF 2025.10		12.61	-	-	-	-
Best-of-N 2025.10		11.54	-	-	-	-
Vanilla-GRPO 2026.05		11.46	-	-	-	-
VHG (Soft) 2026.05		11.46	-	-	-	-
Vanilla-GRPO 2026.05		10.83	-	-	-	-
Base Sampling 2025.10		9.8	-	-	-	-
Qwen3-4B-Base 2026.05		8.12	-	-	-	-
R-Zero 2026.05		7.71	-	-	-	-
R-Zero 2026.05		7.08	-	-	-	-
R-Zero 2026.05		4.17	-	-	-	-
Base Sampling 2025.10		2.6	-	-	-	-
Base Sampling 2025.10		-	-	3.33	-	-
Best-of-N 2025.10		-	5.13	3.6	2.52	-
Beam-Search 2025.10		-	9.45	7.4	4.32	-
PF 2025.10		-	7.32	4.5	2.87	-
ePF 2025.10		-	10.82	7.28	3.42	-
Base Sampling 2025.10		-	-	6.66	-	-
Best-of-N 2025.10		-	17.41	15.8	14.9	-
Beam-Search 2025.10		-	14.19	16.2	16.81	-
PF 2025.10		-	21.61	19.8	17.61	-
ePF 2025.10		-	28.83	25.1	21.96	-