Spoken Task-Oriented Dialogue on SpokenWOZ (test)
100GAHuman
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| HumanAudio (hrs)=–2026.03 | 100 | 100 | 3.58 | 3.64 | 3.8 | 3.67 | — | 84 | 88 | |
| Qwen3-Omni-30B-A3BAudio (hrs)=∼20M2026.03 | 93 | 99 | 3.03 | 3.61 | 2.91 | 3.18 | 15.05 | 95 | 95 | |
| SpokenUS (3B)Audio (hrs)=∼1K2026.03 | 82 | 99 | 3.92 | 4.14 | 4.11 | 4.06 | 11.36 | 93 | 92 | |
| Qwen2.5-Omni-7BAudio (hrs)=∼3M2026.03 | 80 | 97 | 2.86 | 3.29 | 2.94 | 3.03 | 10.53 | 97 | 96 | |
| InteractiveOmni-4BAudio (hrs)=∼465K2026.03 | 59 | 89 | 2.83 | 2.44 | 2.62 | 2.63 | 15.68 | 95 | 94 | |
| Qwen2.5-Omni-3BAudio (hrs)=∼3M2026.03 | 36 | 93 | 1.97 | 2.77 | 2.28 | 2.34 | 15.09 | 90 | 92 | |
| GLM-4-Voice-9BAudio (hrs)=∼700K2026.03 | 33 | 79 | 2.51 | 2.06 | 2.38 | 2.32 | 15.33 | 94 | 96 |