Support workflow execution on Zoom Support n=200 (eval)
4.92Task SuccessIn-Context Baseline
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| In-Context BaselineProtocol=In-context2026.05 | 4.92 | 4.92 | 5 | 5 | 5 | |
| LangGraph OrchestratorType=Orchestrator2026.05 | 4.62 | 4.75 | 4.55 | 4.52 | 4.64 | |
| 8B Subterranean AgentArchitecture=8B, Protocol=Compiled2026.05 | 4.5 | 4.26 | 4.42 | 4.62 | 4.87 |