Audio Understanding on MMAR (test)
67.1PerformanceQwen3-Omni-Instruct
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-Omni-InstructType=Omni, Size=30B-A3B2026.02 | 67.1 | |
| Gemini 2.5 Flash2025.12 | 65.6 | |
| MiMo-Audio-7B-Instruct2025.12 | 63.6 | |
| GPT-4o-Audio2025.12 | 63.5 | |
| Step-Audio-2-miniType=Audio, Size=8B2026.02 | 61.57 | |
| Audio Flamingo 3Type=Audio, Size=8B2026.02 | 61 | |
| Audio Flamingo 32025.12 | 58.5 | |
| Kimi-Audio-7B-InstructType=Audio, Size=7B2026.02 | 57.4 | |
| Qwen2.5-Omni2025.12 | 56.7 | |
| Eureka-Audio-Instruct w DataFluxType=Ours, Size=1.7B2026.02 | 56.2 | |
| Step-Audio2-mini2025.12 | 55.8 | |
| Eureka-Audio-Instruct w/o DataFluxType=Ours, Size=1.7B2026.02 | 50.7 | |
| Qwen2.5-Omni-7BType=Omni, Size=7B2026.02 | 49.6 | |
| Kimi-Audio-Instruct2025.12 | 48 | |
| MiniCPM-oType=Omni, Size=9B2026.02 | 47.9 | |
| Eureka-Audio-BaseType=Ours, Size=1.7B2026.02 | 46.8 | |
| Ming-Lite-Omni-1.5Type=Omni, Size=19B-A2.8B2026.02 | 45.4 | |
| Qwen2.5-Omni-3BType=Omni, Size=3B2026.02 | 43.4 | |
| Qwen2-AudioType=Audio, Size=7B2026.02 | 37.9 | |
| GLM-4-Voice2025.12 | 29.5 |