Alibaba has officially confirmed ownership of HappyHorse, an AI video generation model developed by its innovation arm, ATH (Alibaba Token Hub). The announcement, made on April 10, follows several days of market speculation over the model’s origin and formally ties the fast-rising system to Alibaba’s broader AI push.
Leading the text-to-video leaderboard
HappyHorse has quickly gained visibility in the generative AI space, climbing to the top of the Artificial Analysis text-to-video leaderboard with an Elo score of 1357. That ranking highlights the model’s growing competitive position in a crowded field where performance, consistency, and production efficiency are becoming key differentiators.
Focus on synchronized audiovisual generation
According to the source material, one of HappyHorse’s main strengths is its ability to generate synchronized audiovisual content efficiently. This capability could be especially valuable for commercial video workflows, where aligning visuals and audio often requires labor-intensive editing. If deployed effectively, the model could reduce post-production costs by more than 50%, making it relevant for media, advertising, and digital content production.
Part of Alibaba’s broader AI strategy
The model’s emergence fits into Alibaba’s wider strategic acceleration in AI. The company has recently rolled out multiple models, including Qwen3.6Plus, which the source says set a global usage record. More broadly, Alibaba appears to be shifting from fragmented AI initiatives toward a more systematic approach built on full-stack capabilities.
That transition suggests Alibaba is not treating AI as a side project, but as a foundational layer across businesses ranging from e-commerce to AI-powered solutions. By formally claiming HappyHorse, Alibaba is not only clarifying the model’s origin but also signaling stronger ambitions in multimodal AI and enterprise-scale content generation.