Alibaba Unveils "Happy Horse" Video AI Model, Tops Global Benchmarks
Alibaba's new video model unseats ByteDance's Seedance 2.0 on the global text-to-video leaderboard.
Alibaba Group has made a significant entry into the generative video AI arena with the launch of its 'Happy Horse' model. Developed by the company's Alibaba Token Hub innovation business unit, the model has achieved the top ranking on the widely recognized Artificial Analysis global text-to-video benchmark. This performance notably dethrones ByteDance's previously leading Seedance 2.0 model, signaling a shift in competitive dynamics. The model is currently in a beta testing phase, with Alibaba confirming plans to offer commercial API access in the near future, opening the technology to developers and enterprises.
This strategic move positions Alibaba as a formidable contender in a market segment that has seen intense competition from tech giants like OpenAI with Sora and Google with Veo. The benchmark victory on Artificial Analysis, a key industry evaluator, provides immediate credibility and technical validation for Happy Horse. For Alibaba, this represents more than just a research milestone; it's a direct challenge in a high-value segment of generative AI, aiming to capture market share and establish its technological prowess beyond e-commerce and cloud services.
- Happy Horse model achieved #1 ranking on Artificial Analysis's global text-to-video leaderboard.
- The model surpassed ByteDance's previously leading Seedance 2.0 in benchmark performance.
- Developed by Alibaba Token Hub's innovation unit, with API access planned following beta testing.
Why It Matters
This challenges the current generative video AI hierarchy and signals intensified global competition, offering new options for video creation tools.