Multi-step Reasoning on MUSR Teams (Accuracy, Cost)
61AccuracyStatic Workflow
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Static WorkflowWorkflow Type=Static, LLM Model=DeepSeek-V3.12026.05 | 61 | 0.28 | |
| Zero-shot (Default Imp.)Workflow Type=Zero-shot, LLM Model=DeepSeek-V3.12026.05 | 59 | 0.14 | |
| Dynamic Workflow (ReAct)Workflow Type=Dynamic, LLM Model=DeepSeek-V3.12026.05 | 49 | 1.38 |