Alibaba CEO eyes greater AI research and development, as firm open sources 100 new LLMs

Alibaba Group Holding’s cloud computing services unit on Thursday unveiled its biggest open-source initiative involving large language models (LLMs) – the deep-learning technology underpinning generative artificial intelligence (AI) services like ChatGPT – and a new text-to-video model, as the company shows its rapid progress in this field.

Alibaba Cloud, the digital technology backbone of the e-commerce giant, announced the release of more than 100 of its newly launched Qwen 2.5 LLMs to the global open-source community at the opening of the three-day Apsara Conference, its annual flagship event that runs until Saturday. Alibaba owns the South China Morning Post.

“Alibaba Cloud is investing, with unprecedented intensity, in the research and development of AI technology and the building of its global infrastructure,” Alibaba chief executive Eddie Wu Yongming, who also serves as chairman and chief executive of Alibaba Cloud Intelligence, said at the event. “We aim to establish an AI infrastructure of the future to serve our global customers and unlock their business potential.”

Open source gives public access to a program’s source code, allowing third-party software developers to modify or share its design, fix broken links or scale up its capabilities. Open-source technologies have been a huge contributor to China’s tech industry over the past few decades.

Qwen 2.5 is the latest iteration of Alibaba Cloud’s open-source Tongyi Qianwen family of LLMs. These models range from 0.5 to 72 billion parameters in size, feature enhanced knowledge and stronger capabilities in maths and coding, and are able to support more than 29 languages. Alibaba Cloud said these cater to a wide array of AI applications across various sectors from automobiles and video gaming to science research.

Alibaba Cloud’s latest announcements reflect the Hangzhou-based company’s bid to keep in step with developments leading to the path of artificial general intelligence (AGI), a field of research that attempts to create software with humanlike intelligence and the ability to self-teach.