List your company

Agentic Task Completion and Tool Usage on Complex Multi-step Operational Tasks

83.1Planning EfficiencySupervisor-Specialist
52.62860.53968.4576.361May 24, 2026

Evaluation results

MethodLinks
Supervisor-SpecialistArchitecture=Supervisor-Specialist2026.0583.191.78510099.797.3100
Plan-ExecuteArchitecture=Plan-Execute2026.0553.879.161.797.592.791.790