This paper introduces Goku, a state-of-the-art family of joint image-and-video generation models leveraging rectified flow Transformers to achieve industry-leading performance. We detail the foundational elements enabling high-quality visual generation, including the data curation pipeline, model architecture design, flow formulation, and advanced infrastructure for efficient and robust large-scale training. The Goku models demonstrate superior performance in both qualitative and quantitative evaluations, setting new benchmarks across major tasks. Specifically, Goku achieves 0.76 on GenEval and 83.65 on DPG-Bench for text-to-image generation, and 84.85 on VBench for text-to-video tasks. We believe that this work provides valuable insights and practical advancements for the research community in developing joint image-and-video generation models.
@article{arxiv.2502.04896,
title = {Goku: Flow Based Video Generative Foundation Models},
author = {Shoufa Chen and Chongjian Ge and Yuqi Zhang and Yida Zhang and Fengda Zhu and Hao Yang and Hongxiang Hao and Hui Wu and Zhichao Lai and Yifei Hu and Ting-Che Lin and Shilong Zhang and Fu Li and Chuan Li and Xing Wang and Yanghua Peng and Peize Sun and Ping Luo and Yi Jiang and Zehuan Yuan and Bingyue Peng and Xiaobing Liu},
journal= {arXiv preprint arXiv:2502.04896},
year = {2025}
}