Audio Understanding on MMAU Text instruction
71.47Sound ScoreAudio Flamingo 2
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Audio Flamingo 2Size=3B, Stream.=false, Multi-turn=false2026.06 | 71.47 | 70.96 | 44.74 | 62.4 | |
| Qwen2.5-OmniSize=7B, Stream.=false, Multi-turn=true2026.06 | 67.87 | 69.16 | 59.76 | 65.6 | |
| Baichuan-Omni-1.5Size=7B, Stream.=false, Multi-turn=true2026.06 | 65.47 | 58.98 | 55.26 | 59.9 | |
| Qwen2.5-OmniSize=3B, Stream.=false, Multi-turn=true2026.06 | 65.36 | 48.94 | 57.78 | 57.81 | |
| Audio-InteractionSize=3B, Stream.=true, Multi-turn=true2026.06 | 64.12 | 47.8 | 55.13 | 55.68 | |
| Phi-4-multimodalSize=5.6B, Stream.=false, Multi-turn=true2026.06 | 60.97 | 52.87 | 52.83 | 55.56 | |
| Audio-ReasonerSize=8.4B, Stream.=false, Multi-turn=false2026.06 | 60.06 | 64.3 | 60.7 | 61.71 | |
| Voxtral-MiniSize=3B, Stream.=false, Multi-turn=true2026.06 | 58.56 | 49.7 | 43.53 | 50.6 | |
| Qwen2-AudioSize=7B, Stream.=false, Multi-turn=true2026.06 | 54.95 | 50.98 | 42.04 | 49.2 |