Alibaba Cloud last week globally released more than 100 new open-source variants of its large language foundation model, Qwen 2.5, to the global open-source community. The company has also revamped its proprietary offering as a full-stack AI-computing infrastructure across cloud products, networking and data center architecture, all aimed at supporting the growing demands of AI computing. Alibaba Cloud’s significant contribution was revealed at the Apsara Conference, the annual flagship event held by the cloud division of China’s e-retail giant, often referred to as the Chinese Amazon.

The new Qwen 2.5 family of models, which includes a text-to-video tool, “are designed for use in applications and sectors ranging from automobiles to gaming and science research” and have advanced capabilities in math and coding, writes CNBC.

“Alibaba’s release of over 100 open-source artificial intelligence models and text-to-video technology signals its growing focus on AI for digital commerce,” PYMNTS opines, adding that “the move could democratize access to advanced AI tools, particularly benefiting small- to medium-sized businesses (SMBs),” by providing “new tools for customer engagement and content creation in online retail and marketing.”

In addition to taking on domestic competitors like Baidu and Huawei, the move is seen as Alibaba’s challenge to the dominant U.S. AI companies OpenAI, Google and Meta.

“Over the past few months, variants of Qwen have been topping the leaderboards of sites that measure an AI model’s performance,” writes the non-profit publication Rest of World, quoting Hugging Face CEO Clem Delangue’s statement that “Qwen 72B is the king, and Chinese models are dominating” when the model topped his company’s open LLM leaderboard in June.

The new open-source Qwen 2.5 models range from 0.5 to 72 billion parameters in size. They include “base models, instruct models, and quantized models of various precision levels and methods, spanning different modalities such as language, audio, and vision, along with specialized code and mathematical models,” Alibaba Cloud notes in a news release.

The Qwen model series has generated more than 40 million downloads across platforms including Hugging Face and Alibaba’s own ModelScope. On Hugging Face alone, the open-source Qwen series has formed the basis for over 50,000 new, individualized models, Alibaba Cloud says.