Multimodal Understanding on MMSU (MMSU Score)
79.36MMSU ScoreMAPO
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MAPO2026.05 | 79.36 | — | 76.36 | 82.56 | |
| Qwen3-Omni-InstructTraining mode=Instruction-tuned2026.05 | 77.76 | — | 73.95 | 81.82 | |
| Qwen3-Omni-ThinkingTraining mode=Thinking baseline2026.05 | 76.3 | — | 74.19 | 78.55 | |
| Covo-Audio2026.05 | 66.64 | — | 58.95 | 74.83 | |
| Audio Flamingo3Model Category=Specialized Models2026.04 | 61.4 | — | — | — | |
| Gemini 1.5 Pro2026.05 | 60.68 | — | 46.1 | 76.16 | |
| Qwen2.5-OmniModel size=7B2026.05 | 60.57 | — | 42.5 | 79.83 | |
| Qwen2.5-Omni-3BModel Category=Specialized Models2026.04 | 56.83 | — | — | — | |
| Audio-OmniModel Category=Unified Models2026.04 | 56.83 | — | — | — | |
| GPT-4o Audio2026.05 | 56.38 | — | 39.67 | 71.96 | |
| Qwen2-Audio-InstructModel Category=Specialized Models2026.04 | 53.27 | — | — | — | |
| Ming-OmniModel Category=Unified Models2026.04 | 47.53 | — | — | — | |
| Unified-IO2Model Category=Unified Models2026.04 | 30.74 | — | — | — | |
| MuMuLLaMAModel Category=Unified Models2026.04 | 6.32 | — | — | — | |
| AF-Next-Captioner2026.04 | — | 63.3 | — | — | |
| AF-Next-Instruct2026.04 | — | 59.4 | — | — | |
| AF-Next-Think2026.04 | — | 61.2 | — | — | |
| Gemini-2.5-Flash2026.04 | — | 66.1 | — | — |