Alibaba Group Holding on Thursday released a new artificial intelligence (AI) multimodal model it claims can process text, pictures, audio and video, as well as run on smartphones and laptops.
The Qwen2.5-Omni-7B model, publicly available on GitHub and Hugging Face and packed with 7 billion parameters, can also be used to build AI agents. The open-source model follows the release of Qwen 2.5 in January, adding a new competitive layer to an increasingly booming AI market in China.
The e-commerce and cloud-computing giant said it expects its latest model to support cost-effective AI agents, including applications for visually impaired users via real-time audio guidance.
“This unique combination makes it the perfect foundation for developing agile, cost-effective AI agents that deliver tangible value, especially intelligent voice applications,” Alibaba said in announcing its latest AI entrant.
Alibaba is hardly unique in the multimodal model market. Alphabet Inc.’s Google and OpenAI both offer Generative AI tools that process different types of input including text and audio. On Tuesday, OpenAI added to its capabilities by adding advanced image-generation features to ChatGPT.
Qwen2.5-Omni-7B marks the latest, but certainly not the last, AI product from Alibaba in what has been a whirlwind year of activity as it attempts to keep pace with rival DeepSeek. Earlier this month, Alibaba introduced a new version of Quark, its AI assistant tool. Additionally, Alibaba said it planned to spend $53 billion on AI infrastructure and cloud computing over the next three years.
Last month, Alibaba said it was teaming with Apple Inc. to roll out AI integration for iPhones sold in China.
Alibaba has had little choice but to pump out product in a market operating at a frenetic pace.
In recent months, Chinese companies Baidu and Tencent expanded their AI portfolios after DeepSeek roiled the industry with its low-cost, high-performance R1 model in late December. Earlier this month, Baidu upgraded its AI reasoning models.
Indeed, tech leaders in China have bombarded the market with low-cost AI models as alternatives to products from American rivals OpenAI, Google, and Meta Platforms Inc. The ferocious competition has prompted an escalating arms race among major AI players, as well as political skirmishes by federal agencies and countries to ban the use of DeepSeek for security reasons, based on its alleged ties to the Chinese Communist Party.
