Audio Understanding & Reasoning on MMAU
79.8MMAU ScoreGemini 3 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 3 Pro2026.04 | 79.8 | |
| qwen3.5-omni-plus2026.04 | 79.6 | |
| Step-Audio-R1.52026.04 | 77.9 | |
| qwen3.5-omni-flash2026.04 | 77.2 | |
| Step-Audio-R12026.04 | 77 | |
| Audio ThinkerArchitecture=Native LALM2026.02 | 75.9 | |
| Gemini 3 Flash2026.04 | 75.9 | |
| TACArchitecture=TAC + Text-only Reasoner, Reasoner=Qwen32026.02 | 73.9 | |
| TACArchitecture=TAC + Text-only Reasoner, Reasoner=Gemini32026.02 | 72.2 | |
| Audio Flamingo 3 + CoATBackbone=Audio Flamingo 3, Method=CoAT2026.06 | 70 | |
| Qwen2.5-Omni-7B + CoATBackbone=Qwen2.5-Omni-7B, Method=CoAT2026.06 | 69.4 | |
| Audio Flamingo 3Backbone=Audio Flamingo 3, Method=Baseline2026.06 | 69.4 | |
| Qwen2-Audio + CoATBackbone=Qwen2-Audio, Method=CoAT2026.06 | 66.9 | |
| Qwen2.5-Omni-7BBackbone=Qwen2.5-Omni-7B, Method=Baseline2026.06 | 65.6 | |
| Qwen2-AudioBackbone=Qwen2-Audio, Method=Baseline2026.06 | 52.5 |