Multi-turn function calling on BFCL Multi Turn
59.4Task Completion Score (TCS)Qwen3-235B-A22B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-235B-A22BAverage Steps (Stp)=5.2, Input Tokens (Tok In)=190.6, Output Tokens (Tok Out)=0.5, Execution Time=36.92026.05 | 59.4 | |
| Gemini-3-flashAverage Steps (Stp)=9.4, Input Tokens (Tok In)=462.0, Output Tokens (Tok Out)=22.0, Execution Time=331.82026.05 | 57.5 | |
| GPT-5-miniAverage Steps (Stp)=19.2, Input Tokens (Tok In)=801.6, Output Tokens (Tok Out)=44.0, Execution Time=1025.82026.05 | 27.3 | |
| DeepSeek-V3.2Average Steps (Stp)=15.3, Input Tokens (Tok In)=680.2, Output Tokens (Tok Out)=1.5, Execution Time=502.32026.05 | 26.3 |