Audio Understanding on MMAU mini (test)
78.5AccuracyAudio-DeepThinker
Evaluation results
| Method | Links | |
|---|---|---|
| Audio-DeepThinkerModel Category=Open-source Models Finetuned with RL2026.04 | 78.5 | |
| AudioMCQModel Category=Open-source Models Finetuned with RL2026.04 | 78.2 | |
| Audio-ThinkerModel Category=Open-source Models Finetuned with RL2026.04 | 78 | |
| Qwen3-Omni-30B-A3B-InstructModel Category=Open-source Models2026.04 | 77.8 | |
| Step-Audio 2Model Category=Open-source Models2026.04 | 77.58 | |
| CESARModel Category=Open-source Models Finetuned with RL2026.04 | 77.1 | |
| Omni-R1Model Category=Open-source Models Finetuned with RL2026.04 | 77 | |
| Qwen3-Omni-30B-A3B-ThinkingModel Category=Open-source Models2026.04 | 75.8 | |
| Covo-AudioModel Category=Open-source Models2026.04 | 75.3 | |
| Ke-Omni-RModel Category=Open-source Models Finetuned with RL2026.04 | 74.6 | |
| Audio Flamingo 3Model Category=Open-source Models2026.04 | 73.3 | |
| Gemini 2.5 FlashModel Category=Closed-source Models2026.04 | 71.8 | |
| Qwen-2.5-OmniModel Category=Open-source Models2026.04 | 71.5 | |
| Gemini 2.0 FlashModel Category=Closed-source Models2026.04 | 70.5 | |
| Kimi-AudioModel Category=Open-source Models2026.04 | 68.2 | |
| Audio-ReasonerModel Category=Open-source Models Finetuned with RL2026.04 | 67.7 | |
| GPT-4o AudioModel Category=Closed-source Models2026.04 | 62.5 | |
| Audio Flamingo 2Model Category=Open-source Models2026.04 | 62.4 | |
| Qwen2-Audio-InstructModel Category=Open-source Models2026.04 | 59.6 | |
| GPT-4o mini AudioModel Category=Closed-source Models2026.04 | 53 |