The Chinese tech giant said it is open-sourcing four models under Wan2.1, the latest version of Alibaba’s AI platform that can generate images and videos from text commands and image inputs.
Open-source AI technology has been in the spotlight since Chinese company DeepSeek shook up global markets in January. DeepSeek claims its AI models are built at a fraction of the cost of the world's leading AI companies while relying on less advanced Nvidia chips.
DeepSeek's model is open source, like Alibaba's, meaning it can be downloaded and modified by others. Chinese companies in particular have pushed open source models. Alibaba's and DeepSeek's products are now among the most widely used globally.
Alibaba announced its first open source model in August 2023. In the US, Meta is leading the open source race with its Llama models.
In another development, Alibaba on February 24 announced plans to invest more than 380 billion yuan (about 53 billion USD) in developing cloud computing and AI infrastructure over the next 3 years.
Alibaba previously reported an 8% increase in fourth-quarter revenue to 280 billion yuan, a result that exceeded expectations and sent its shares up 14% on the Hong Kong Stock Exchange on February 21.
Alibaba's shares have surged this year, driven by factors including improved financial performance and recent signals that Chinese President Xi Jinping is advocating further support for the domestic private sector.
Comment (0)