List your company

Mathematical Reasoning on AMC23 (Accuracy, Length)

90.83AccuracyStandardp
35.366849.765964.16578.5641Feb 2, 2026

Evaluation results

MethodLinks
StandardpBackbone=DeepSeek-R1-7B2026.0290.836,224.23
S3-CoTBackbone=DeepSeek-R1-7B2026.0290.835,715.74
EffReaBackbone=DeepSeek-R1-7B2026.0288.334,649.27
LASERDEBackbone=DeepSeek-R1-7B2026.0288.332,766.08
AutoTHINKBackbone=DeepSeek-R1-7B2026.0288.333,846.43
C3oTBackbone=DeepSeek-R1-7B2026.0287.56,820.08
S3-CoTscBackbone=DeepSeek-R1-7B2026.0287.55,654.8
LC-R1Backbone=DeepSeek-R1-7B2026.02852,788.19
EfficientpBackbone=DeepSeek-R1-7B2026.0283.336,062.81
ShorterBetterBackbone=DeepSeek-R1-7B2026.0266.671,613.66
CoT-ValveBackbone=DeepSeek-R1-7B2026.02651,940.33
C3oTBackbone=Qwen2.5-7B2026.0251.67788.6
S3-CoTscBackbone=Qwen2.5-7B2026.0246.67687.33
S3-CoTBackbone=Qwen2.5-7B2026.0245.83678.8
TokenSkipBackbone=Qwen2.5-7B2026.0245766.95
EfficientpBackbone=Qwen2.5-7B2026.0244.17573.44
StandardpBackbone=Qwen2.5-7B2026.0243.33846.49
CoT-ValveBackbone=Qwen2.5-7B2026.0237.5900.23