List your company

Audio-Language Understanding (MCQ) on MMAU-Pro 1.0 (test)

59.5AccuracyQwen3-Omni (I)
18.4229.08539.7550.415Apr 27, 2026

Evaluation results

MethodLinks
Qwen3-Omni (I)Size=30B†, Evaluation Protocol=Full2026.0459.572.6
Qwen2.5-Omni-7BSize=10.7B, Evaluation Protocol=Full2026.0457.768.2
Qwen3-Omni (T)Size=30B†, Evaluation Protocol=Full2026.0456.571.7
Audio-Flamingo-3Size=8.4B, Evaluation Protocol=Full2026.0452.783.7
Qwen2-Audio-7BSize=8.2B, Evaluation Protocol=Full2026.0444.870.1
Audio-Flamingo-3Size=8.4B, Evaluation Protocol=None2026.0444.1—
Phi-4-MultimodalSize=5.6B, Evaluation Protocol=Full2026.0443.765.5
DeSTA-2.5Size=8.8B, Evaluation Protocol=Full2026.0443.572
Qwen3-Omni (I)Size=30B†, Evaluation Protocol=None2026.0443.2—
Voxtral-Mini-3BSize=4.7B, Evaluation Protocol=Full2026.0441.771.9
Qwen3-Omni (I)Size=30B†, Evaluation Protocol=TB2026.0441—
Qwen3-Omni (T)Size=30B†, Evaluation Protocol=None2026.0440.5—
Qwen2.5-Omni-7BSize=10.7B, Evaluation Protocol=None2026.0439.3—
Qwen3-Omni (T)Size=30B†, Evaluation Protocol=TB2026.0433.8—
Qwen2-Audio-7BSize=8.2B, Evaluation Protocol=None2026.0431.4—
DeSTA-2.5Size=8.8B, Evaluation Protocol=None2026.0431.3—
Audio-Flamingo-3Size=8.4B, Evaluation Protocol=TB2026.0431.2—
Qwen2.5-Omni-7BSize=10.7B, Evaluation Protocol=TB2026.0431.2—
Voxtral-Mini-3BSize=4.7B, Evaluation Protocol=None2026.0430—
Phi-4-MultimodalSize=5.6B, Evaluation Protocol=TB2026.0429.9—
Phi-4-MultimodalSize=5.6B, Evaluation Protocol=None2026.0428.6—
Qwen2-Audio-7BSize=8.2B, Evaluation Protocol=TB2026.0428.2—
Chance LevelEvaluation Protocol=Full2026.0425.9—
Chance LevelEvaluation Protocol=None2026.0425.9—
Chance LevelEvaluation Protocol=TB2026.0425.9—
DeSTA-2.5Size=8.8B, Evaluation Protocol=TB2026.0420.3—
Voxtral-Mini-3BSize=4.7B, Evaluation Protocol=TB2026.0420—