List your company

Agent Task and Scenario Completion on AppWorld (train)

91.1Task Goal CompletionTrajectory-Informed Memory Generation
86.52487.71288.990.088Mar 11, 2026

Evaluation results

MethodLinks
Trajectory-Informed Memory GenerationMemory Configuration=+Memory2026.0391.183.3
BaselineMemory Configuration=Baseline2026.0386.773.3