Audio Understanding on MMAU mini original (test)
80.78Accuracy (Sound Domain)MiMo-Audio
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MiMo-Audio2026.02 | 80.78 | 70.96 | 63.06 | 71.6 | |
| Audio-Flamingo-3Evaluator=GPT-52026.02 | 79.58 | 72.46 | 60.36 | 70.8 | |
| MPAR2-7B2026.02 | 77.98 | 71.3 | 74.47 | 74.59 | |
| MPAR2-7BAblation Components=w/o R_spr2026.02 | 77.74 | 70.68 | 73.78 | 74.08 | |
| MPAR2-7BAblation Components=w/o R_spr, R_perception and R_rea, Reward Strategy=simple accuracy reward2026.02 | 77.54 | 67.32 | 73.27 | 72.81 | |
| Audio-ThinkerReproduction Status=reproduced2026.02 | 77.48 | 70.36 | 73.37 | 73.7 | |
| MPAR2-7BAblation Components=w/o R_spr and R_perception2026.02 | 77.19 | 70.66 | 71.43 | 73.11 | |
| MPAR2-7BAblation Components=w/o R_perception2026.02 | 75.84 | 72.9 | 71.34 | 73.36 | |
| Omni-R12026.02 | 73.6 | 74.3 | 66.1 | 71.3 | |
| Step-Audio-R12026.02 | 72.97 | 61.68 | 74.47 | 69.7 | |
| SARIReproduction Status=reproduced2026.02 | 72.75 | 67.22 | 61.26 | 67.08 | |
| Step-Audio-R1.12026.02 | 72.67 | 72.46 | 77.18 | 74.1 | |
| Qwen2.5-Omni-7B2026.02 | 69.4 | 66.8 | 61.6 | 65.9 | |
| R1-AQAReproduction Status=reproduced2026.02 | 68.77 | 64.37 | 63.66 | 65.6 | |
| Gemini-2.5-Flash2026.02 | 67.56 | 61.48 | 63.76 | 64.27 | |
| Kimi-Audio2026.02 | 61.68 | 73.27 | 60.66 | 65 | |
| GPT-4o-Audio2026.02 | 60.36 | 57.19 | 67.27 | 61.6 | |
| Audio ReasonerReproduction Status=reproduced2026.02 | 60.06 | 64.3 | 60.7 | 61.71 | |
| DeSTA2.5-AudioEvaluator=GPT-52026.02 | 58.26 | 44.61 | 58.26 | 53.7 | |
| Phi-4-Multimodal2026.02 | 57.96 | 51.8 | 48.35 | 52.7 | |
| Qwen2-Audio-Instruct2026.02 | 49.85 | 52.4 | 42.94 | 48.4 |