MOVA addresses the limitations of proprietary systems like Sora 2 and Veo 3 by offering a fully open-source framework for Image-to-Video-Audio (IT2VA) and Text-to-Video-Audio (T2VA) tasks. The model employs an asymmetric dual-tower architecture fused via a bidirectional cross-attention mechanism, leveraging a Mixture-of-Experts (MoE) design with 32B total parameters (18B active during inference) to ensure high-quality synthesis with efficient deployment. Alongside the model weights, we provide a fine-grained bimodal data pipeline and support for LoRA fine-tuning, empowering the community to advance research in synchronized cinematic synthesis.
Please refer to the
Quick Start section on the GitHub page for model usage and inference scripts.
We evaluate our model through both objective benchmarks and subjective human evaluations. Below are the Elo scores and win rates comparing MOVA to existing open-source models.
1@article{yu2026mova,
2 title={MOVA: Towards Scalable and Synchronized Video-Audio Generation},
3 author={Yu, Donghua and Chen, Mingshu and Chen, Qi and Luo, Qi and Wu, Qianyi and Cheng, Qinyuan and Li, Ruixiao and Liang, Tianyi and Zhang, Wenbo and Tu, Wenming and others},
4 journal={arXiv preprint arXiv:2602.08794},
5 year={2026}
6}
