End-to-end Tool-use on ToolBench I3 v1
52.19SoPRToolWeaver
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ToolWeaverSetting=Direct Generation2026.01 | 52.19 | 59.02 | |
| ToolWeaver2026.01 | 52.19 | 59.02 | |
| ToolLlama*Setting=Retrieval2026.01 | 51.37 | 31.15 | |
| ToolGen*2026.01 | 47.54 | 34.42 | |
| ToolGenSetting=Direct Generation2026.01 | 36.34 | 49.18 | |
| ToolGen2026.01 | 36.34 | 49.18 | |
| ToolLlama-2*Setting=Retrieval2026.01 | 34.7 | 21.31 | |
| GPT-3.5*Setting=Retrieval2026.01 | 34.43 | 37.7 | |
| GPT-4o-miniSetting=Retrieval (Re.)2026.01 | 24.86 | — | |
| GPT-4o-miniSetting=Retrieval2026.01 | 24.86 | 44.26 | |
| ToolLlama-2Setting=Retrieval (Re.)2026.01 | 10.93 | 24.59 | |
| ToolLlama-2Setting=Retrieval2026.01 | 10.93 | 24.59 |