List your company

LLM Routing on LogicGrid, MGSM, and CommonGen (Phase-2 test)

78Average AccuracyGraphPlanner
36.447.25868.8Apr 26, 2026

Evaluation results

MethodLinks
GraphPlannerEvaluation Mode=Zero-shot2026.0478
RouterDCEvaluation Mode=Zero-shot2026.0458
GraphRouterEvaluation Mode=Zero-shot2026.0446
Router-R1Evaluation Mode=Zero-shot2026.0438