List your company

Mathematical Reasoning on MATH 500 (Accuracy and Thinking Token Usage)

88.5AccuracyDART
36.55063.577Jun 22, 2026

Evaluation results

MethodLinks
DARTModel Scale=32B, Thinking Strategy=DART2026.0688.569
DARTModel Scale=8B, Thinking Strategy=DART2026.0688.267
ATModel Scale=14B, Thinking Strategy=AT2026.0687.6—
DARTModel Scale=14B, Thinking Strategy=DART2026.0687.637
ATModel Scale=32B, Thinking Strategy=AT2026.0686.2—
ATModel Scale=8B, Thinking Strategy=AT2026.0685.6—
NTModel Scale=32B, Thinking Strategy=NT2026.0682.2—
NTModel Scale=14B, Thinking Strategy=NT2026.0681.2—
NTModel Scale=8B, Thinking Strategy=NT2026.0676.6—
ATModel Scale=4B, Thinking Strategy=AT2026.0670.5—
NTModel Scale=4B, Thinking Strategy=NT2026.0666.5—
ATModel Scale=1.7B, Thinking Strategy=AT2026.0665.5—
DARTModel Scale=4B, Thinking Strategy=DART2026.0665.530
DARTModel Scale=1.7B, Thinking Strategy=DART2026.065829
ATModel Scale=0.6B, Thinking Strategy=AT2026.0656—
DARTModel Scale=0.6B, Thinking Strategy=DART2026.065423
NTModel Scale=1.7B, Thinking Strategy=NT2026.0652.5—
NTModel Scale=0.6B, Thinking Strategy=NT2026.0638.5—