Alibaba releases over 100 open- source AI models to boost competition

Beijing: Alibaba unveiled more than 100 open-source artificial intelligence models on Thursday, along with enhancements to its proprietary technology, as part of its efforts to intensify competition with domestic and international rivals.

The newly launched models, branded as Qwen 2.5, cater to a wide range of industries, from automotive to gaming and scientific research. According to Alibaba, the models boast improved capabilities in mathematics and coding. The Chinese tech giant aims to challenge competitors such as Baidu, Huawei, and U.S.-based firms like Microsoft and OpenAI.

Alibaba’s AI models, trained on massive datasets, can comprehend prompts and generate both text and images. The company emphasized that open-sourcing these models will enable global researchers, academics, and businesses to leverage them for developing generative AI applications, without needing to train their own systems—resulting in significant time and cost savings. By offering these models to the broader community, Alibaba hopes to increase adoption and expand its influence in the AI sector.

First introduced last year, Alibaba's Tongyi Qianwen, or Qwen, model has undergone several updates, with the open-source versions being downloaded over 40 million times to date. In addition to the open-source Qwen 2.5, Alibaba upgraded its proprietary flagship model, Qwen-Max, which remains available through its cloud computing services for businesses. The new Qwen-Max 2.5 reportedly surpasses competitors like Meta's Llama and OpenAI’s GPT-4 in areas such as reasoning and language comprehension.
Additionally, Alibaba launched a new text-to-video tool powered by its AI models, enabling users to input a prompt and generate a video—a feature similar to OpenAI’s Sora.

“Alibaba Cloud is investing with unprecedented intensity in AI research and development and the expansion of its global infrastructure,” said Eddie Wu, CEO of Alibaba, in a statement, signaling the company’s commitment to advancing its position in the competitive AI landscape.