Tool Calling on In-domain (unseen)
57.24Exact Match (EM)TInR-U
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TInR-U2026.04 | 57.24 | 74.83 | 62.76 | |
| ToolGen2026.04 | 55.86 | 72.59 | 60.52 | |
| ToolRetriever+ToolRLRetriever=ToolRetriever, Reasoning Model=ToolRL2026.04 | 51.72 | 67.41 | 59.83 | |
| TR-Feedback+ToolRLRetriever=TR-Feedback, Reasoning Model=ToolRL2026.04 | 51.38 | 67.41 | 57.76 | |
| ToolRetriever+Qwen3-8BRetriever=ToolRetriever, Reasoning Model=Qwen3-8B2026.04 | 48.45 | 68.79 | 52.93 | |
| BM25+ToolRLRetriever=BM25, Reasoning Model=ToolRL2026.04 | 46.38 | 59.48 | 51.9 | |
| Ada-embedding+ToolRLRetriever=Ada-embedding, Reasoning Model=ToolRL2026.04 | 45.69 | 60 | 51.9 | |
| ToolRetriever+xLAM-7b-rRetriever=ToolRetriever, Reasoning Model=xLAM-7b-r2026.04 | 36.38 | 51.21 | 40 | |
| ToolRetriever+Hammer2.1-7bRetriever=ToolRetriever, Reasoning Model=Hammer2.1-7b2026.04 | 33.28 | 43.45 | 34.83 | |
| ATU2026.04 | 26.38 | 44.14 | 41.38 |