Text-Audio Alignment on Baton-Pair two-sound-event
71Pair AccuracyAQAScore
Evaluation Results
| Method | Links | |
|---|---|---|
| AQAScoreBackbone=Qwen2.5-Omni-7B2026.01 | 71 | |
| AQAScoreBackbone=Qwen2.5-Omni-3B2026.01 | 70.3 | |
| CLAPScoreVariant=630k-best2026.01 | 69 | |
| AQAScoreBackbone=AF32026.01 | 66.3 | |
| CLAPScoreTraining Data=Music+Speech2026.01 | 66.2 | |
| AQAScoreBackbone=AF3-Think2026.01 | 65.8 | |
| CLAPScoreTraining Data=Music+AS2026.01 | 65.8 | |
| CLAPScoreTraining Data=Music+Speech+AS2026.01 | 64.7 | |
| CLAPScoreTraining Data=630k-AS2026.01 | 64.6 | |
| AQAScoreBackbone=AF3-Chat2026.01 | 63.1 | |
| PromptingBackbone=Qwen2.5-Omni-7B2026.01 | 55.8 | |
| PromptingBackbone=AF3-Chat2026.01 | 54.6 | |
| PromptingBackbone=AF32026.01 | 50.6 | |
| PromptingBackbone=Qwen2.5-Omni-3B2026.01 | 50.3 | |
| PromptingBackbone=AF3-Think2026.01 | 49.8 | |
| CascadeBackbone=Qwen2.5-Omni-7B2026.01 | 46.1 | |
| CascadeBackbone=AF3-Chat2026.01 | 45.5 | |
| CascadeBackbone=AF32026.01 | 43.3 | |
| CascadeBackbone=Qwen2.5-Omni-3B2026.01 | 37.3 | |
| CascadeBackbone=AF3-Think2026.01 | 15.6 |