- tech360.tv
Alibaba Unveils New AI Models and Text-to-Video Technology
Alibaba has released over 100 new open-source AI models from its Qwen 2.5 family. The models range from 0.5 to 72 billion parameters in size and support over 29 languages. Alibaba has also unveiled a new text-to-video model, joining the emerging market for this technology.
This move, announced on Thursday, marks a substantial acceleration in Alibaba's AI efforts and positions the company as a formidable competitor in the rapidly evolving field of generative AI.
The newly released open-source models, numbering over 100, are part of Alibaba's Qwen 2.5 family, which was initially introduced in May as the company's latest foundational large language model. These models boast an impressive range of capabilities, with sizes varying from 0.5 to 72 billion parameters. This diversity allows for a wide spectrum of AI applications across various industries, including automotive, gaming, and scientific research.
Alibaba's approach to AI development sets it apart from some of its competitors. While firms like Baidu and OpenAI have largely focused on closed-source strategies, Alibaba has adopted a hybrid model. This approach involves investing in both proprietary and open-source development, allowing the company to expand its AI product range more effectively.
One of the standout features of the new models is their versatility. According to Alibaba, these AI tools demonstrate proficiency in mathematics and coding, and can support over 29 languages. This multilingual capability could prove crucial in Alibaba's efforts to expand its AI offerings on a global scale.
In addition to the open-source models, Alibaba has also unveiled a new text-to-video model as part of its Tongyi Wanxiang image generation family. This development puts Alibaba in direct competition with international players like OpenAI, which has also shown interest in text-to-video technology. It also aligns Alibaba with other Chinese tech firms entering this emerging market, such as ByteDance, which launched its text-to-video app Jimeng AI on Apple's App Store for Chinese users in August.
• Alibaba has released over 100 new open-source AI models from its Qwen 2.5 family
• The models range from 0.5 to 72 billion parameters in size and support over 29 languages
• Alibaba has also unveiled a new text-to-video model, joining the emerging market for this technology
Source: REUTERS