Function Calling on BFCL Multi-turn
42.3AccuracyEVOTOOL
Evaluation Results
| Method | Links | |
|---|---|---|
| EVOTOOLBackbone=GPT-4o2026.03 | 42.3 | |
| AdaPlannerBackbone=GPT-4o2026.03 | 41.1 | |
| EvoPromptBackbone=GPT-4o2026.03 | 39.9 | |
| ANYTOOLBackbone=GPT-4o2026.03 | 39.3 | |
| ReActBackbone=GPT-4o2026.03 | 37.2 | |
| EVOTOOLBackbone=Qwen3-8B2026.03 | 36.4 | |
| PromptBreederBackbone=GPT-4o2026.03 | 36.2 | |
| EASYTOOLBackbone=GPT-4o2026.03 | 36 | |
| AdaPlannerBackbone=Qwen3-8B2026.03 | 36 | |
| EvoPromptBackbone=Qwen3-8B2026.03 | 35.9 | |
| OPROBackbone=GPT-4o2026.03 | 35.3 | |
| PromptBreederBackbone=Qwen3-8B2026.03 | 34.4 | |
| OPROBackbone=Qwen3-8B2026.03 | 33.5 | |
| ReActBackbone=Qwen3-8B2026.03 | 33.3 | |
| DRAFTBackbone=GPT-4o2026.03 | 32 | |
| ANYTOOLBackbone=Qwen3-8B2026.03 | 30.3 | |
| EASYTOOLBackbone=Qwen3-8B2026.03 | 28.2 | |
| DRAFTBackbone=Qwen3-8B2026.03 | 27.1 | |
| Plan-and-SolveBackbone=GPT-4o2026.03 | 25.3 | |
| Plan-and-SolveBackbone=Qwen3-8B2026.03 | 24 | |
| CoTBackbone=GPT-4o2026.03 | 21.4 | |
| CoTBackbone=Qwen3-8B2026.03 | 20.3 |