List your company

Mathematical Reasoning on GSM8K 776 questions (test)

0Cost ($)SLM
-0.004160.023920.0520.08008Jan 29, 2026

Evaluation results

MethodLinks
SLMStrategy Type=Baseline2026.01073——
Reactive ShepherdingStrategy Type=Cascading2026.010.03489.167.41.97
Oracle ShepherdingStrategy Type=Oracle2026.010.0349867.43.1
Proactive ShepherdingStrategy Type=Routing2026.010.03680.965.90.93
GraphRouterStrategy Type=Routing2026.010.04182.260.60.93
FrugalGPTStrategy Type=Cascading2026.010.04786.754.81.21
ABCStrategy Type=Cascading2026.010.04894.953.71.89
RouteLLMStrategy Type=Routing2026.010.06588.137.50.98
LLMStrategy Type=Baseline2026.010.10498——