General Multi-modal Understanding on Public Benchmarks (MMAU, MMAR) (test)
76.4MMAU Test Mini (%)Qwen-Audio-AHA
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Qwen-Audio-AHAAlignment=DPO (Aligned)2025.12 | 76.4 | 72.5 | 54.8 | 59.7 | |
| Qwen2.5-OmniAlignment=Base2025.12 | 74.6 | 71.2 | 54.2 | 58.1 | |
| Audio Flamingo 32025.12 | 74.5 | 72.4 | 25 | 54.1 | |
| Gemini 2.52025.12 | 70.5 | 67.1 | 54.9 | 57.3 | |
| GPT-4o2025.12 | 57.8 | 54.5 | 41.7 | 49.2 | |
| Kimi-Audio2025.12 | 54.9 | 56.1 | 50.1 | 49.1 |