Multimodal Audio Understanding on MMAU Mini
86.31Sound ScoreHuman
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| HumanType=Proprietary2026.05 | 86.31 | 78.22 | 82.17 | 82.23 | |
| Step-Audio-2Type=Open-Weights2026.05 | 84.04 | 73.56 | 75.15 | 77.58 | |
| Audio-ThinkerType=Open-Weights2026.05 | 81.98 | 74.25 | 76.88 | 77.7 | |
| Omni-R1Type=Open-Weights2026.05 | 81.7 | 73.4 | 76 | 77 | |
| MiMo-AudioType=Open-Weights2026.05 | 81.68 | 74.25 | 68.17 | 74.7 | |
| Nova 2 OmniType=Proprietary2026.05 | 81.08 | 70.36 | 81.98 | 77.8 | |
| MAPOType=Open-Weights2026.05 | 79.88 | 71.86 | 81.68 | 77.8 | |
| Audio Flamingo 3Type=Open-Weights2026.05 | 79.58 | 66.77 | 66.37 | 73.3 | |
| Covo-AudioType=Open-Weights2026.05 | 78.68 | 76.05 | 71.17 | 75.3 | |
| Qwen3-Omni-InstType=Open-Weights2026.05 | 77.78 | 73.05 | 78.68 | 76.5 | |
| Qwen3-Omni-ThinkType=Open-Weights2026.05 | 77.48 | 70.96 | 76.58 | 75 | |
| Gemini 2.5 ProType=Proprietary2026.05 | 75.08 | 68.26 | 71.47 | 71.6 | |
| GPT-4o AudioType=Proprietary2026.05 | 64.56 | 56.29 | 66.67 | 62.5 |