Apple Intelligence will launch in China using an Apple‑trained LLM, not Google Gemini or си Qwen, according to Reuters. Apple partnered with Alibaba for training, marking a shift from relying on external models. The Chinese version will still use Qwen but Apple’s own model will be primary.

In the U.S. and other regions, Apple Intelligence relies on Google Gemini, while its Foundation Models remain fully Apple‑owned. Apple has also offered ChatGPT extensions for Siri, but those and Gemini or Anthropic’s Claude are unavailable in China.

China’s AI regulations require local partnerships; thus, Apple will integrate Baidu technology alongside its own model and Qwen. Censorship is expected, with Alibaba potentially filtering content. Despite regulatory hurdles, Apple is poised to be the first company to release a proprietary AI model in China, following a brief accidental launch in March.

The iPhone 17 series has seen renewed popularity in China, reversing past declines, and Apple’s entry into the local AI market may further strengthen its position.