Tool-use and Agentic Reasoning on tau-Bench Retail
56AccuracyDynamic Workflow (ReAct)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Dynamic Workflow (ReAct)Workflow Type=Dynamic, LLM Model=DeepSeek-V3.12026.05 | 56 | 7.03 | |
| Static WorkflowWorkflow Type=Static, LLM Model=DeepSeek-V3.12026.05 | 52 | 4.75 | |
| Zero-shot (Default Imp.)Workflow Type=Zero-shot, LLM Model=DeepSeek-V3.12026.05 | 11 | 3.5 |