Audio-Text Alignment Evaluation on BRACE AudioCaps-Main 1.0 (test)
70.3HHPrompting
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| PromptingModel=AF3-Chat2026.01 | 70.3 | 63.3 | 65.5 | 66.3 | |
| PromptingModel=Qwen2.5-Omni-3B2026.01 | 67.6 | 73 | 69.2 | 70.5 | |
| PromptingModel=AF32026.01 | 64 | 58.7 | 69.2 | 64.5 | |
| CLAPScoreModel=Music+AS2026.01 | 61.9 | 87.6 | 63.5 | 72.7 | |
| CLAPScoreModel=Music+Speech+AS2026.01 | 61.9 | 85.6 | 65.1 | 72.7 | |
| CLAPScoreModel=630k-best2026.01 | 61.9 | 88.1 | 63.8 | 73 | |
| CLAPScoreModel=630k-AS2026.01 | 61.7 | 86.3 | 65.1 | 72.8 | |
| CLAPScoreModel=Music+Speech2026.01 | 59.7 | 80.7 | 66.3 | 71.1 | |
| PromptingModel=AF3-Think2026.01 | 59.7 | 49 | 64 | 57.6 | |
| PromptingModel=Qwen2.5-Omni-7B2026.01 | 58.3 | 73.7 | 69.7 | 69.9 | |
| CascadeModel=Qwen2.5-Omni-7B2026.01 | 57.6 | 69.4 | 62.9 | 64.8 | |
| CascadeModel=AF32026.01 | 56.8 | 52.6 | 67.9 | 60.6 | |
| CascadeModel=AF3-Chat2026.01 | 54.7 | 58.2 | 57.8 | 57.6 | |
| CascadeModel=AF3-Think2026.01 | 52.5 | 49.9 | 57.6 | 54 | |
| CascadeModel=Qwen2.5-Omni-3B2026.01 | 51.8 | 64.5 | 63.3 | 62.4 | |
| AQAScoreModel=Qwen2.5-Omni-3B2026.01 | 51.1 | 79.6 | 66 | 69.4 | |
| AQAScoreModel=AF3-Chat2026.01 | 48.9 | 55.3 | 61.5 | 57.6 | |
| AQAScoreModel=Qwen2.5-Omni-7B2026.01 | 47.5 | 79.8 | 66.8 | 69.5 | |
| AQAScoreModel=AF32026.01 | 46 | 64.9 | 60.3 | 60.4 | |
| AQAScoreModel=AF3-Think2026.01 | 38.9 | 60 | 57.8 | 56.3 |