Single-turn Tool Use on ToolBench
19.6ScoreToolGrad
Evaluation Results
| Method | Links | |
|---|---|---|
| ToolGradModel Version/Size=12B2025.08 | 19.6 | |
| ToolGradModel Version/Size=4B2025.08 | 17.6 | |
| Claude 4.5Model Version/Size=opus2025.08 | 15.4 | |
| GPT-5Model Version/Size=nano2025.08 | 15.4 | |
| GPT-5Model Version/Size=mini2025.08 | 14.7 | |
| ToolGradModel Version/Size=1B2025.08 | 14.1 | |
| Claude 4.5Model Version/Size=sonnet2025.08 | 13.5 | |
| Claude 4.5Model Version/Size=haiku2025.08 | 12.8 | |
| GPT-5Model Version/Size=base2025.08 | 12.7 | |
| Gemini 2.5Model Version/Size=pro2025.08 | 11.4 | |
| Gemini 2.5Model Version/Size=flash2025.08 | 8.5 | |
| Gemini 2.5Model Version/Size=lite2025.08 | 6.9 |