Unified multimodal models (UMMs) aim to jointly perform multimodal understanding and generation within a single framework. We present TUNA, a native UMM that builds a unified continuous visual representation by cascading a VAE encoder with a representation encoder. This unified representation space allows end-to-end processing of images and videos for both understanding and generation tasks. Compared to prior UMMs with decoupled representations, TUNA's unified visual space avoids representation format mismatches introduced by separate encoders, outperforming decoupled alternatives in both understanding and generation. Moreover, we observe that stronger pretrained representation encoders consistently yield better performance across all multimodal tasks, highlighting the importance of the representation encoder. Finally, in this unified setting, jointly training on both understanding and generation data allows the two tasks to benefit from each other rather than interfere. Our extensive experiments on multimodal understanding and generation benchmarks show that TUNA achieves state-of-the-art results in image and video understanding, image and video generation, and image editing, demonstrating the effectiveness and scalability of its unified representation design.
@article{arxiv.2512.02014,
title = {TUNA: Taming Unified Visual Representations for Native Unified Multimodal Models},
author = {Zhiheng Liu and Weiming Ren and Haozhe Liu and Zijian Zhou and Shoufa Chen and Haonan Qiu and Xiaoke Huang and Zhaochong An and Fanny Yang and Aditya Patel and Viktar Atliha and Tony Ng and Xiao Han and Chuyan Zhu and Chenyang Zhang and Ding Liu and Juan-Manuel Perez-Rua and Sen He and Jürgen Schmidhuber and Wenhu Chen and Ping Luo and Wei Liu and Tao Xiang and Jonas Schult and Yuren Cong},
journal= {arXiv preprint arXiv:2512.02014},
year = {2025}
}