Apple Intelligence could be coming soon to China as Alibaba releases new Qwen3 AI models

Alibaba has released a new family of open-source Qwen3 AI models designed to work with Apple’s MLX machine learning framework. According to a report by South Morning China Post, this move could pave the way for Apple Intelligence to launch in mainland China.

In a post on X, the team behind Alibaba Cloud’s Qwen project said the Qwen3 models have been optimised for Apple’s MLX framework, allowing them to run efficiently across a range of Apple devices—from iPhones and iPads to high-performance Macs. MLX is Apple’s in-house framework designed to train and deploy AI models directly on Apple silicon, such as M-series chips.

These models are now available on Hugging Face and Alibaba’s own ModelScope platform, potentially giving Chinese developers a homegrown option for building apps and services that take advantage of Apple’s AI infrastructure, as per the report.

Furthermore, the Qwen3 rollout appears tied to regulatory realities in China. Apple’s global AI strategy hinges on OpenAI’s GPT models, but Chinese regulations require foreign tech to partner with local firms for AI services. A February report from the South China Morning Post — owned by Alibaba — claimed Apple had struck a deal to use Qwen models for Apple Intelligence inside China.