Tool Use on StableToolBench G2 Instruction
68.8SL ScoreTrace-Based
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Trace-BasedEvaluation Protocol=Trace-based2026.02 | 68.8 | 45.4 | |
| D2Evaluation Protocol=Trace-based2026.02 | 68.2 | 43.2 | |
| DRAFTEvaluation Protocol=Trace-based2026.02 | 68.2 | 44.9 | |
| Play2PromptEvaluation Protocol=Trace-based2026.02 | 68 | 43.4 | |
| D0Evaluation Protocol=Trace-based2026.02 | 67.8 | 43.4 | |
| D1Evaluation Protocol=Trace-based2026.02 | 67.4 | 42.9 |