Alibaba has lifted the lid on one of the AI industry’s more curious recent success stories. In doing so, they’ve shown just how quickly China’s video-generation race is heating up. The company confirmed it is behind HappyHorse 1.0, a text-to-video model that climbed to the top of global benchmark rankings before anyone knew of its true origins.
The model had appeared anonymously on the Artificial Analysis leaderboard, where it quickly took the No.1 spot in blind tests for video generation, sparking speculation over whether it came from a major tech player or an independent lab.
Developed within Alibaba’s newly formed AI unit under its Token Hub (ATH) division, HappyHorse 1.0 is still in closed beta, with API access expected to roll out in the near future. The model supports both text-to-video and image-to-video generation, and is designed to produce relatively realistic, human-centric footage – a hot battleground in commercial AI content creation.
What makes the launch notable isn’t just performance, but positioning. HappyHorse 1.0 outperformed or matched leading models from rivals including ByteDance’s Seedance 2.0 – software that’s also been making a big splash recently – and Kuaishou’s Kling on benchmark rankings, putting Alibaba right back into a race it had previously lagged in.
But why drop something so powerful anonymously? By releasing the model without attribution, Alibaba was effectively able to benchmark performance without brand bias. They let the product speak first, and then stepped in to cover themselves in glory.
This wasn’t a poorly timed move. As mentioned above, video generation is a hot topic now. It’s also emerging as one of the few AI segments with clear routes monetisation. Those routes mostly seem to be coming from across advertising, entertainment and short-form content. And competition in China is fierce. Firms are moving aggressively to grab up a slice of the pie. At the same time, some western players are pull back or refocus efforts, leaving even more of the market up for grabs.
HappyHorse’s longer-term impact probably won’t depend on the leaderboard rankings that made it such a splash. What matters now is execution. Can they offer pricing, compute efficiency and iteration speed to beat the competition. Its sudden rise does suggest something worth noting though. Alibaba is no longer playing catch-up in AI.
