On Thursday, Alibaba Group Holding (NYSE:BABA) cloud unit launched a new artificial intelligence model in its “Qwen series.“

The new “Qwen2.5-Omni-7B” is a multimodal model that can process inputs, including text, images, audio, and videos, while generating real-time text and natural speech responses, CNBC cites from Alibaba Cloud’s website. 

The company can deploy the model on edge devices like mobile phones. Now, visually impaired people can navigate their environment through real-time audio descriptions. 

Also Read: In Geopolitical Chess, China’s Latest Energy Norms Block Nvidia’s Chip Strategy Amid US Sanctions

Alibaba open-sourced the new model on the platforms Hugging Face and Github. 

Reports in January indicated that Alibaba’s Qwen2.5-VL AI model surpasses GPT-4o and Claude 3.5 in video analysis, math and document parsing capabilities. Qwen2.5-VL can control devices, analyze charts, and book flights, showcasing advanced AI applications for practical use.

Last week, Baidu Inc (NASDAQ:BIDU) launched a new multimodal foundational model and its debut reasoning-focused model. 

Alibaba has committed $53 billion in its cloud computing and AI infrastructure over the next three years. 

Kai Wang of Morningstar told CNBC that large Chinese tech players like Alibaba are well-positioned to benefit from China’s post-DeepSeek AI boom. 

Alibaba stock surged 85% in the last 12 months.

Price Action: BABA stock is up 1.29% at $133.95 premarket at last check Thursday.

Also Read:

• Amazon Takes Creative Control Of James Bond, Taps Top Producers to Reboot Iconic Franchise

Photo by Poetra.RH via Shutterstock