End-to-end Tool-use on ToolBench I1-Cat v1
61.76SoPRToolLlama*
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ToolLlama*Setting=Retrieval2026.01 | 61.76 | 50.98 | |
| ToolWeaverSetting=Direct Generation2026.01 | 57.41 | 43.14 | |
| ToolWeaver2026.01 | 57.41 | 43.14 | |
| ToolGenSetting=Direct Generation2026.01 | 55.56 | 42.48 | |
| ToolGen2026.01 | 55.56 | 42.48 | |
| GPT-3.5*Setting=Retrieval2026.01 | 53.05 | 54.25 | |
| GPT-4o-miniSetting=Retrieval (Re.)2026.01 | 50.11 | — | |
| GPT-4o-miniSetting=Retrieval2026.01 | 50.11 | 50.33 | |
| ToolGen*2026.01 | 49.46 | 39.87 | |
| ToolLlama-2Setting=Retrieval (Re.)2026.01 | 36.93 | 27.45 | |
| ToolLlama-2Setting=Retrieval2026.01 | 36.93 | 27.45 |