List your company

Mathematical Reasoning on AIME 24 (Accuracy, Length)

53.33AccuracyEffRea
0.2914.0627.8341.6Feb 2, 2026

Evaluation results

MethodLinks
EffReaBackbone=DeepSeek-R1-7B2026.0253.3310,706.94
StandardpBackbone=DeepSeek-R1-7B2026.0251.1114,061.22
LASERDEBackbone=DeepSeek-R1-7B2026.0251.115,985.25
C3oTBackbone=DeepSeek-R1-7B2026.0251.1112,884.91
S3-CoTBackbone=DeepSeek-R1-7B2026.0251.1112,217.53
S3-CoTscBackbone=DeepSeek-R1-7B2026.025012,167.95
EfficientpBackbone=DeepSeek-R1-7B2026.0247.7811,224.51
AutoTHINKBackbone=DeepSeek-R1-7B2026.0247.788,303.59
LC-R1Backbone=DeepSeek-R1-7B2026.0242.226,371.22
ShorterBetterBackbone=DeepSeek-R1-7B2026.0237.784,701.88
CoT-ValveBackbone=DeepSeek-R1-7B2026.02204,177.39
S3-CoTBackbone=Qwen2.5-7B2026.0212.22800.62
S3-CoTscBackbone=Qwen2.5-7B2026.0211.11831.47
StandardpBackbone=Qwen2.5-7B2026.027.78996.75
EfficientpBackbone=Qwen2.5-7B2026.027.78741.21
CoT-ValveBackbone=Qwen2.5-7B2026.026.681,068.58
C3oTBackbone=Qwen2.5-7B2026.025.56866.8
TokenSkipBackbone=Qwen2.5-7B2026.022.33842.71