Alibaba Group Holding Ltd. has quietly launched an internal initiative code-named “Qianwen” to develop a personal AI assistant app based on its most advanced large language model, Qwen, according to people familiar with the matter.

The app is designed to compete head-on with OpenAI’s ChatGPT, signaling the Chinese tech giant’s push into the global consumer AI race. Senior executives see the project as “the battle for the future in the AI era,” the people said, as Alibaba seeks to leverage Qwen’s open-source capabilities to gain an advantage over Western rivals.

The move marks a major escalation in Alibaba’s artificial intelligence ambitions, following its pledge earlier this year to invest 380 billion yuan (US$52 billion) over three years in AI infrastructure. Until now, the company has largely focused on enterprise AI services, offering model APIs through Alibaba Cloud to industries such as finance, retail and manufacturing.

With Qwen’s strong technical performance and rising global profile, Alibaba now sees an opportunity to expand into the consumer market. More than 100 engineers have been reassigned to work on the project under tight secrecy, and two floors at the company’s Hangzhou headquarters have been dedicated to the effort, the people said.

Alibaba is also developing an international version of the Qianwen app aimed at global users, positioning Qwen to directly challenge ChatGPT in overseas markets.

Qwen, first launched three years ago, has become one of the world’s most widely used open-source large models, with more than 600 million downloads to date. Its latest flagship version, Qwen3-Max, released in September, features over one trillion parameters and was trained on 36 trillion tokens of data. Alibaba says Qwen3-Max’s performance surpasses OpenAI’s GPT-5 and Anthropic’s Claude Opus 4, ranking it among the top three AI models worldwide.