Alibaba officially claims the video generation model HappyHorse-1.0, and will release another multimodal model soon

Gate News message: On April 10, Alibaba officially confirmed that its video generation model HappyHorse-1.0 is one of its in-house products. The model comes from the Taotian Group’s former Future Life Lab team. In Alibaba’s latest organizational restructuring, that team was placed under the “AI Innovation Department” within the newly formed Alibaba Token Hub (ATH) business group. In anonymous voting on the third-party evaluation platform Artificial Analysis, HappyHorse-1.0 received significantly more votes than a video generation model from one internet company and a model from another internet company on the pure video generation task. For audio-and-visual integrated generation, it performed on par with the former. According to people close to Alibaba, HappyHorse-1.0 is only one of several multimodal models developed by this team. Recently, Alibaba is also expected to launch another different multimodal model. As of now, HappyHorse-1.0 is not open-sourced, which is consistent with Alibaba’s recent overall shift toward closed-source model strategies. Since the end of March, multiple new models Alibaba has released have not been open-sourced. The push to develop multimodal models in quick succession is against the backdrop of an internet company’s video generation model delivering impressive results during the 2026 Spring Festival period, which surprised some inside Alibaba. At the same time, multimodal generation will significantly increase Token consumption, thereby affecting the market share of MaaS (model as a service). According to IDC data, in the first half of 2025, a certain cloud service provider had already captured 49.2% of the market share, while Alibaba Cloud accounted for 27%.