On Thursday, Alibaba, a prominent Chinese technology firm, announced the release of new open-source artificial intelligence models and text-to-video AI technology, further enhancing its position in the rapidly expanding generative AI sector.
The newly introduced open-source models, exceeding 100 in total, are part of Alibaba's Qwen 2.5 family, which represents its latest foundational large language model launched in May.
Similar to their U.S. counterparts, Chinese technology companies are making significant investments in generative AI, as they strive to create comprehensive product portfolios and diverse offerings.
While rivals like Baidu and OpenAI have largely opted for closed-source strategies, Alibaba has adopted a hybrid approach, focusing on both proprietary and open-source development to expand its AI product lineup.
The new models vary in size from 0.5 to 72 billion parameters—key factors that influence an AI model's capabilities and performance—demonstrating expertise in mathematics, coding, and supporting over 29 languages, according to Alibaba's statement.
These models are designed to serve a broad spectrum of AI applications across multiple industries, including automotive, gaming, and scientific research.
Additionally, Alibaba introduced a new text-to-video model as part of its Tongyi Wanxiang image generation series, aligning itself with a growing number of Chinese tech companies venturing into this emerging field.
This development positions Alibaba in direct competition with global entities like OpenAI, which has also expressed interest in text-to-video technology.
In August, ByteDance, the parent company of TikTok, launched its text-to-video application, Jimeng AI, on Apple's App Store for users in China.