Tool Calling on StableToolBench I3-Inst
48.3Solvable Pass RateTOOLQP
Evaluation Results
| Method | Links | |
|---|---|---|
| TOOLQPBase Model=Qwen3-30B, Inference Pipeline=ReAct2026.01 | 48.3 | |
| Q2PBase Model=Qwen3-30B, Inference Pipeline=ReAct2026.01 | 41 | |
| Q2DBase Model=Qwen3-30B, Inference Pipeline=ReAct2026.01 | 40.4 | |
| Base RetrieverBase Model=Qwen3-30B, Inference Pipeline=ReAct2026.01 | 40.3 | |
| Q2EBase Model=Qwen3-30B, Inference Pipeline=ReAct2026.01 | 38.6 | |
| Re-InvokeBase Model=Qwen3-30B, Inference Pipeline=ReAct2026.01 | 28.6 |