Alibaba announced on Tuesday that it will release an open-source version of its video and image-generating artificial intelligence model, Wan 2.1, as competition intensifies in China’s AI sector.
The Chinese tech giant will provide full details in a recorded video scheduled for release at 11 p.m. (8:00 p.m. PKT), a company spokesperson said.
The move follows the recent launch of open-source AI models by DeepSeek, which have gained global attention for their performance rivaling industry leaders like OpenAI. Alibaba originally introduced its latest video and image-generating AI model in January before renaming it Wan from Wanx, highlighting its ability to produce highly realistic visuals.
Alibaba previously noted that its AI model leads on VBench, a video generation benchmark, particularly in multi-object interactions. On Tuesday, the company also released a preview of its reasoning model, QwQ-Max, which it intends to make open source upon the full version’s official release.
The announcement comes as Alibaba ramps up investments in AI and cloud computing. Earlier this week, the company pledged at least 380 billion yuan ($52 billion) over the next three years to strengthen its AI infrastructure and cloud services.
