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

Math Reasoning on AIME 2025 (Acc, ∆Tok)

80Accuracy

Qwen3-Next-80B

Updated 1mo ago

Evaluation Results

Method	Links
Qwen3-Next-80B 2025.10		80	-
Qwen3-4B-Thinking 2025.10		76.2	-
Qwen3-Next-80B + THINKBRAKE 2025.10		73.3	-16.7
Qwen3-14B 2025.10		69.2	-
Qwen3-32B 2025.10		68.3	-
Qwen3-32B + THINKBRAKE 2025.10		68.3	-4.3
Qwen3-4B-Thinking + NoWait 2025.10		66.7	-35.6
Qwen3-14B + THINKBRAKE 2025.10		65.8	-6
Phi-4-Reasoning 2025.10		63.7	-
Qwen3-4B-Thinking + THINKBRAKE 2025.10		62.8	-33.4
Qwen3-4B + THINKBRAKE 2025.10		60.4	-13.5
Qwen3-4B-Thinking + DEER 2025.10		60	-37.9
Qwen3-4B 2025.10		59.6	-
Qwen3-14B + NoWait 2025.10		56.7	-16.6
Qwen3-32B + NoWait 2025.10		56.7	-10.6
Qwen3-4B-Thinking + Dynasor-CoT 2025.10		53.3	-28.9
Phi-4-Reasoning + NoWait 2025.10		53.3	-8.6
Phi-4-Reasoning + THINKBRAKE 2025.10		53.3	-5.2
Qwen3-4B + NoWait 2025.10		50	-27.1
DeepSeek-R1-7B 2025.10		37.6	-
DeepSeek-R1-7B + Dynasor-CoT 2025.10		33.4	-26.4
DeepSeek-R1-7B + DEER 2025.10		33.3	-9.8
DeepSeek-R1-7B + THINKBRAKE 2025.10		33.2	-20.3
Qwen3-4B-Thinking + THINKBRAKE-p 2025.10		30.4	-76
Qwen3-14B + ThinkLess 2025.10		30	-100
Phi-4-Reasoning + THINKBRAKE-p 2025.10		29.2	-66.6
Qwen3-32B + THINKBRAKE-p 2025.10		26.7	-66.4
DeepSeek-R1-7B + NoWait 2025.10		26.7	-36.5
Qwen3-14B + THINKBRAKE-p 2025.10		26.2	-72
Qwen3-4B + THINKBRAKE-p 2025.10		25	-67.8
Qwen3-4B + ThinkLess 2025.10		23.3	-100
Phi-4-Reasoning + ThinkLess 2025.10		23.3	-100
DeepSeek-R1-7B + THINKBRAKE-p 2025.10		22.9	-73.5
Qwen3-32B + ThinkLess 2025.10		20	-100
DeepSeek-R1-7B + ThinkLess 2025.10		20	-100
Qwen3-4B-Thinking + ThinkLess 2025.10		3.3	-100