HunyuanVideo: A Systematic Framework For Large Video Generative Models

By Weijie Kong, Qi Tian, Zijian Zhang, Rox Min, Zuozhuo Dai, Jin Zhou, Jia-Liang Xiong, Xin Li, Bo Wu, Jianwei Zhang and 42 more. arXiv.org, 2024

1,493 citations recorded by the source corpus.

Recent advancements in video generation have significantly impacted daily life for both individuals and industries. However, the leading video generation models remain closed-source, resulting in a notable performance gap between industry capabilities and those available to the public. In this report, we introduce HunyuanVideo, an innovative open-source video foundation model that demonstrates performance in video generation comparable to, or even surpassing, that of leading closed-source models. HunyuanVideo encompasses a comprehensive framework that integrates several key elements, including data curation, advanced architectural design, progressive model scaling and training, and an efficient infrastructure tailored for large-scale model training and inference. As a result, we successfully trained a video generative model with over 13 billion parameters, making it the largest among all open-source models. We conducted extensive experiments and implemented a series of targeted designs to ensure high visual quality, motion dynamics, text-video alignment, and advanced filming techniques. According to evaluations by professionals, HunyuanVideo outperforms previous state-of-the-art models, including Runway Gen-3, Luma 1.6, and three top-performing Chinese video generative models. By releasing the code for the foundation model and its applications, we aim to bridge the gap between closed-source and open-source communities. This initiative will empower individuals within the community to experiment with their ideas, fostering a more dynamic and vibrant video generation ecosystem. The code is publicly available at https://github.com/Tencent/HunyuanVideo.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/1fa298e3f745099d39ca5bd088fcb62f87e8cc2f

Citations in the corpus

• Step1X-Edit: A Practical Framework for General Image Editing (2025, neutral): “The extracted embeddings are then fed into a lightweight connector module, such as the token refiner [32, 24].”
• Qwen-Image Technical Report (2025, neutral): “1-VAE (Wan et al., 2025), Hunyuan-VAE (Kong et al., 2024), and Qwen-Image-VAE function as joint image and video tokenizers.”
• Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion (2025, neutral): “Video diffusion models typically adopt bidirectional attention mechanisms to simultaneously denoise all video frames [3, 4, 6, 13, 23–26, 39, 64, 80, 83, 97].”
• VACE: All-in-One Video Creation and Editing (2025, neutral): “With the rapid development of image [2, 7, 16, 18, 58, 59] and video [22, 32, 73, 77] generation models, they are being used to create high-quality visual content and are widely applied in fields such as advertising, film special effects, game development, and animation production [13, 43–45 , 55 ].”
• DanceGRPO: Unleashing GRPO on Visual Generation (2025, neutral): “Our foundation model is HunyuanVideo [22], with reward signals derived from VideoAlign [14].”
• Flow-GRPO: Training Flow Matching Models via Online RL (2025, neutral): “Although this work focuses on T2I tasks, Flow-GRPO has potential for video generation [25, 27], raising several future directions: (1) Reward Design: Simple heuristics, such as using object detectors or trackers as rule-based rewards, can encourage physical realism and temporal consistency, but…”
• Hunyuan3D 2.0: Scaling Diffusion Models for High Resolution Textured 3D Assets Generation (2025, neutral): “…importance of automated 3D generation and rapid development in image and video generation fueled by the rise of diffusion models [33, 74, 24, 50, 43], the field of 3D generation appears to be relatively stagnant in the era of large models and big data, with only a handful of works making…”