Tool Calling on ToolACE 500 tools
82.28TSAJTPRO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| JTPROBackbone=GPT-52026.04 | 82.28 | 90 | 74.38 | |
| GEPABackbone=GPT-52026.04 | 77.17 | 85.75 | 66.12 | |
| JTPROBackbone=o3-mini2026.04 | 76.19 | 88.52 | 65.29 | |
| JTPROBackbone=GPT-4o mini2026.04 | 75.25 | 88.12 | 69.42 | |
| GEPABackbone=GPT-4o mini2026.04 | 73.33 | 85.27 | 61.98 | |
| GEPABackbone=o3-mini2026.04 | 73.33 | 85.27 | 61.98 | |
| BaseBackbone=GPT-52026.04 | 73.02 | 84.785 | 62.73 | |
| BaseBackbone=o3-mini2026.04 | 70.78 | 84.994 | 59.454 | |
| BaseBackbone=GPT-4o mini2026.04 | 63.832 | 86.996 | 60 |