List your company

Mathematical Reasoning on Out-domain Aggregate SAT Math, AMC23, AIME24, OLYMPIAD Bench

52.98Avg Acc (A_bar)L1
48.549649.699850.8552.0002Feb 13, 2026

Evaluation results

MethodLinks
L1Base Model=DeepSeek-1.5B2026.0252.989,902.20.120.12
CRTBase Model=DeepSeek-1.5B2026.0252.438,786.40.210.21
O1-PrunerBase Model=DeepSeek-1.5B, Ratio=4.02026.0251.589,1410.120.12
ACPOBase Model=DeepSeek-1.5B2026.0251.338,772.20.140.14
DeepSeek-1.5BModel Type=Base Model2026.0250.849,894.500
O1-PrunerBase Model=DeepSeek-1.5B, Ratio=1.02026.0249.769,042.8-0.01-0.12
ShorterBetterBase Model=DeepSeek-1.5B2026.0249.587,722.40.1-0.03
ThinkPruneBase Model=DeepSeek-1.5B, Constraint=1k2026.0249.468,908.5-0.03-0.17
ThinkPruneBase Model=DeepSeek-1.5B, Constraint=5002026.0248.729,123.2-0.13-0.33