Recently, the new generation multimodal video generation model HappyHorse, developed by Alibaba ATH Innovation Department, has officially started a gray-scale test. As a strong competitor in the global AI video field, the model has demonstrated outstanding performance in three core rankings of Arena.ai: text-to-video, image-to-video, and video editing.

Film-Level Visual Quality and Deep Semantic Understanding

HappyHorse focuses on "film-level" visual performance and supports 1080P ultra-high-definition output. It can accurately handle various visual styles, whether it's the nostalgic Hong Kong style, grand historical atmosphere, or new Chinese-style anime effects, all can be recreated with delicate lighting and natural camera movements.

image.png

At the technical level, the model has strong semantic parsing capabilities, capable of handling complex instructions up to 800 words, and supports inputting up to 9 reference images simultaneously, ensuring the stability of character appearance, scene elements, and prop details. According to the information, generating a high-quality 15-second narrative video takes only 2 to 5 minutes.

One-Click Video Creation with Audio-Visual Synchronization

In addition to breakthroughs in the visual aspect, HappyHorse also performs well in multilingual support and audio-visual interaction. It is compatible with multiple languages such as Cantonese, English, and French, enabling characters' mouth movements, expressions, and body language to be highly consistent. More notably, the model can automatically generate ambient sound effects and background music based on the content of the visuals, achieving automated audio-visual synchronization creation workflow.

Chinese Online's "Dimensional Divine Brush" Has Been Integrated, Expanding the AI Toolkit

At the same time as the model starts testing, the domestic digital content leading enterprise Chinese Online also announced that its full-stack AI content creation solution "Dimensional Divine Brush" has been officially integrated with the HappyHorse model.

Previously, "Dimensional Divine Brush" had already integrated several industry-leading models, including Seedance 2.0. By introducing HappyHorse, Chinese Online further built a comprehensive creative ecosystem covering text, images, and videos. This move aims to provide creators with more diverse tool choices, lower the threshold for producing high-quality content, and help high-quality IP achieve rapid transformation across different formats.

As global AI technology continues to evolve, the popularity of such tools may reshape the creation logic of films and short videos.