Tool-use Parameter Quality Assessment on ToolBench
76.1TAFC Win RateTAFC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TAFCModel=Llama-3.1-8B2026.01 | 76.1 | 12.8 | 11.1 | |
| TAFCModel=Qwen2.5-7B2026.01 | 74.6 | 14.2 | 11.2 | |
| TAFCModel=Qwen2.5-32B2026.01 | 71.3 | 16.4 | 12.3 | |
| TAFCModel=Llama-3.1-70B2026.01 | 69.7 | 18.3 | 12 | |
| TAFCModel=Average2026.01 | 69.6 | 18.2 | 12.2 | |
| TAFCModel=Qwen2.5-72B2026.01 | 68.2 | 19.1 | 12.7 | |
| TAFCModel=Claude-3.5-Sonnet2026.01 | 64.8 | 22.7 | 12.5 | |
| TAFCModel=GPT-4o2026.01 | 62.4 | 24.2 | 13.4 |