Alibaba Cloud announced the launch of its latest artificial intelligence model in the “Qwen series” on Thursday, intensifying the competition among large language models in China following the recent “DeepSeek moment.”

The new model, named “Qwen2.5-Omni-7B,” is multimodal, enabling it to handle various inputs such as text, images, audio, and video while providing real-time text and natural speech outputs, as stated on Alibaba Cloud’s website.

The company highlighted that the model can be deployed on edge devices, including smartphones, ensuring high efficiency without sacrificing performance.

“This distinctive combination establishes it as an ideal foundation for creating agile, cost-effective AI agents that offer significant value, particularly in intelligent voice applications,” Alibaba remarked.

The company noted that the new model could assist visually impaired individuals by providing real-time audio descriptions to help them navigate their surroundings.

The model has been made open-source on platforms like Hugging Face and GitHub, following a recent trend in China after DeepSeek made its R1 model open-source.

Open-source typically means that the software’s source code is freely accessible online for modification and redistribution. Alibaba Cloud claims to have open-sourced over 200 generative AI models in recent years.

In the wake of China’s AI surge spurred by DeepSeek, Alibaba and other competitors in the generative AI space are rapidly launching new, cost-effective models and products.

Just last week, Baidu introduced a new multimodal foundational model alongside its first reasoning-focused model.

Meanwhile, Alibaba released its updated Qwen 2.5 AI model in late January and unveiled a new version of its AI assistant tool, Quark, earlier this month.

The company is firmly dedicated to its AI strategy, recently announcing a plan to invest $53 billion in cloud computing and AI infrastructure over the next three years—surpassing its spending in the sector over the last decade.

Kai Wang, a senior equity analyst at Morningstar, told CNBC that major Chinese tech firms like Alibaba, which are establishing data centers to support AI computing needs while also developing their own large language models, are well-positioned to capitalize on the post-DeepSeek AI boom in China.

Last month, Alibaba achieved a significant milestone for its AI business by confirming a partnership with Apple to integrate AI features into iPhones sold in China.

On Wednesday, the company also announced an expanded strategic partnership with BMW aimed at speeding up the incorporation of its AI technology into the automaker’s next-generation intelligent vehicles.

Read more

Alibaba makes AI model for video, image generation publicly available

China’s Baidu launches two new AI models as industry competition heats up

Alibaba releases AI model Qwen 2.5, claims superiority over DeepSeek