List your company

Mathematical Reasoning on GSM8K and MATH500 Aggregate

86.02Avg AccuracyL1
76.483278.959181.43583.9109Feb 13, 2026

Evaluation results

MethodLinks
L1base_model=DeepSeek-1.5B2026.0286.023,581.5-0.0022-0.0022
CRTbase_model=DeepSeek-1.5B2026.0285.352,499.20.29010.2901
O1-Pruner-4.0base_model=DeepSeek-1.5B2026.0284.973,006.50.12880.1288
DeepSeek-1.5B2026.0284.813,42800
ACPObase_model=DeepSeek-1.5B2026.0284.112,713.30.16730.1261
ThinkPrune-1kbase_model=DeepSeek-1.5B2026.0283.622,6670.15150.0811
O1-Pruner-1.0base_model=DeepSeek-1.5B2026.0283.062,7230.1022-0.0013
ThinkPrune-500base_model=DeepSeek-1.5B2026.0282.082,7250.0441-0.1168
ShorterBetterbase_model=DeepSeek-1.5B2026.0276.851,749.60.0207-0.4482