End-to-end Tool-use on ToolBench I1 v1
56.13SoPRToolLlama-2*
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ToolLlama-2*Setting=Retrieval2026.01 | 56.13 | 50.92 | |
| ToolGen*2026.01 | 56.13 | 50.92 | |
| ToolWeaverSetting=Direct Generation2026.01 | 54.85 | 36.08 | |
| ToolLlama*Setting=Retrieval2026.01 | 54.6 | 49.08 | |
| ToolWeaverSetting=Direct Generation2026.01 | 53.17 | 40.49 | |
| ToolWeaver2026.01 | 53.17 | 40.49 | |
| GPT-4o-miniSetting=Retrieval (Re.)2026.01 | 53.16 | — | |
| ToolGenSetting=Direct Generation2026.01 | 52.97 | 36.2 | |
| ToolGen2026.01 | 52.97 | 36.2 | |
| GPT-4o-miniSetting=Retrieval (Re.)2026.01 | 52.25 | — | |
| GPT-4o-miniSetting=Retrieval2026.01 | 52.25 | 47.24 | |
| GPT-3.5*Setting=Retrieval2026.01 | 51.43 | 53.37 | |
| ToolGenSetting=Direct Generation2026.01 | 45.36 | 32.91 | |
| ToolLlama-2Setting=Retrieval (Re.)2026.01 | 28.94 | 25.15 | |
| ToolLlama-2Setting=Retrieval2026.01 | 28.94 | 25.15 | |
| ToolLlama-2Setting=Retrieval (Re.)2026.01 | 28.48 | 26.58 |