Audio-Visual Video Understanding on Video-MME truncated <= 5min v2
15.33Overall ScoreOurs(Qwen3-Omni)
Evaluation Results
| Method | Links | |
|---|---|---|
| Ours(Qwen3-Omni)Size=30B2026.06 | 15.33 | |
| MiniCPM-o 4.5Size=9B2026.06 | 14.48 | |
| Qwen3-OmniSize=30B2026.06 | 14.31 | |
| OmniVinciSize=9B2026.06 | 11.59 | |
| video-SALMONN 2+Size=7B2026.06 | 11.5 | |
| Qwen2.5-OmniSize=7B2026.06 | 10.28 | |
| Uni-MoE-2.0-OmniSize=30B2026.06 | 8.82 | |
| Ours(Qwen2.5-Omni)Size=7B2026.06 | 8.5 | |
| Ours(VITA-1.5)Size=8B2026.06 | 7.42 | |
| VITA-1.5Size=8B2026.06 | 5.95 |