Function Calling on BFCL Average v4
63AccuracyDVAO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DVAOBase Model=Qwen2.5-7B-Instruct2026.05 | 63 | 79.21 | |
| GDPOBase Model=Qwen2.5-7B-Instruct2026.05 | 60.13 | 68.12 | |
| RCBase Model=Qwen2.5-7B-Instruct2026.05 | 58.38 | 76.42 | |
| DVAOBase Model=Qwen2.5-3B-Instruct2026.05 | 56.66 | 76.65 | |
| ACBase Model=Qwen2.5-3B-Instruct2026.05 | 53.47 | 64.69 | |
| GDPOBase Model=Qwen2.5-3B-Instruct2026.05 | 52.73 | 65.88 | |
| GRPOBase Model=Qwen2.5-7B-Instruct2026.05 | 52.26 | 0 | |
| RCBase Model=Qwen2.5-3B-Instruct2026.05 | 51.02 | 60.48 | |
| Qwen2.5-7B-InstructType=Base Model2026.05 | 47.83 | 0 | |
| ACBase Model=Qwen2.5-7B-Instruct2026.05 | 44.25 | 68.04 | |
| GRPOBase Model=Qwen2.5-3B-Instruct2026.05 | 36.56 | 5.61 | |
| Qwen2.5-3B-InstructType=Base Model2026.05 | 36.1 | 5.66 |