Audio Reasoning on MMAU mini 1.0 (test)
83.48Sound ScoreStep-Audio 2
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Step-Audio 2Source=Closed Source2025.10 | 83.48 | 73.65 | 76.88 | 78 | |
| Omni-R1Source=Closed Source2025.10 | 81.7 | 73.4 | 76 | 77 | |
| AudioToolAgentSource=Closed Source2025.10 | 81.68 | 72.46 | 78.38 | 77.5 | |
| Audio Flamingo 3Source=Open Source2025.10 | 79.58 | 66.77 | 66.37 | 73.3 | |
| AudioToolAgent-OpenSource=Open Source2025.10 | 78.38 | 69.46 | 75.08 | 74.3 | |
| Qwen2.5-OmniSource=Open Source2025.10 | 78.1 | 65.9 | 70.6 | 71.5 | |
| Step-Audio 2 miniSource=Open Source2025.10 | 76.28 | 71.56 | 71.47 | 73.2 | |
| Kimi-AudioSource=Open Source2025.10 | 75.68 | 66.77 | 62.16 | 68.2 | |
| Gemini 2.5 ProSource=Closed Source2025.10 | 75.08 | 68.26 | 71.47 | 71.6 | |
| Gemini 2.5 FlashSource=Closed Source2025.10 | 73.27 | 65.57 | 76.58 | 71.8 | |
| Gemini 3 ProSource=Closed Source2025.10 | 71.17 | 65.57 | 75.08 | 70.6 | |
| DeSTA2.5-AudioSource=Open Source2025.10 | 70.27 | 56.29 | 71.47 | 66 | |
| Audio ReasonerSource=Open Source2025.10 | 67.87 | 69.16 | 66.07 | 67.7 | |
| Phi-4-multimodalSource=Open Source2025.10 | 65.47 | 64.37 | 67.27 | 65.7 | |
| GPT-4o AudioSource=Closed Source2025.10 | 64.56 | 56.29 | 66.67 | 62.5 |