Stateful Agent-User Interaction on tau-Bench Airline
39.1Pass@1EVOTOOL
Evaluation Results
| Method | Links | |
|---|---|---|
| EVOTOOLBackbone=GPT-4o2026.03 | 39.1 | |
| AdaPlannerBackbone=GPT-4o2026.03 | 38.8 | |
| OPROBackbone=GPT-4o2026.03 | 36.8 | |
| Plan-and-SolveBackbone=GPT-4o2026.03 | 36.7 | |
| EvoPromptBackbone=GPT-4o2026.03 | 36.3 | |
| ANYTOOLBackbone=GPT-4o2026.03 | 36.2 | |
| ReActBackbone=GPT-4o2026.03 | 35.9 | |
| PromptBreederBackbone=GPT-4o2026.03 | 31.3 | |
| EASYTOOLBackbone=GPT-4o2026.03 | 29.9 | |
| DRAFTBackbone=GPT-4o2026.03 | 29.2 | |
| CoTBackbone=GPT-4o2026.03 | 25.3 | |
| AdaPlannerBackbone=Qwen3-8B2026.03 | 16.4 | |
| EVOTOOLBackbone=Qwen3-8B2026.03 | 15.7 | |
| Plan-and-SolveBackbone=Qwen3-8B2026.03 | 15 | |
| OPROBackbone=Qwen3-8B2026.03 | 15 | |
| EvoPromptBackbone=Qwen3-8B2026.03 | 14.7 | |
| ANYTOOLBackbone=Qwen3-8B2026.03 | 14.6 | |
| ReActBackbone=Qwen3-8B2026.03 | 14.4 | |
| PromptBreederBackbone=Qwen3-8B2026.03 | 11.3 | |
| EASYTOOLBackbone=Qwen3-8B2026.03 | 10.3 | |
| DRAFTBackbone=Qwen3-8B2026.03 | 8.9 | |
| CoTBackbone=Qwen3-8B2026.03 | 8.2 |