Zhixiang releases native omni-modal video generation model HiDream-O1-Video-1.0

On September 15, HiDream.ai officially released its first native omni-modal video generation model, HiDream-O1-Video-1.0 (model abbreviation: HD-V1). HD-V1 adopts HiDream.ai's self-developed native omni-modal architecture, supports multiple modal inputs such as text, images, and videos, and can generate 520 second 1080p high-fidelity videos in one click. It also features upgrades in dimensions including intent understanding, comprehension of physical laws, autonomous narrative planning, and integrated audio-video generation. At the time of release, HD-V1 ranked fourth globally on the Artificial Analysis Image to Video Leaderboard, an AI benchmarking and analysis platform, and eighth on Arena.ai Image-to-Video, an AI model blind evaluation platform.