Apple begins testing Apple Intelligence with Qwen integration in China

Summary

Apple has confirmed that it will integrate Alibaba's Qwen model into Apple Intelligence in mainland China. Developer beta testing is set to begin on November 4, 2026, with a broader iOS 27.3 rollout scheduled for December.

Why it matters

This represents a significant strategic concession. Rather than waiting for its own models to overcome regulatory hurdles in China, Apple is embedding a local AI lab's model into Siri, Writing Tools, and Visual Intelligence. This move finally cracks open the world's second-largest smartphone market. It also highlights the reality that even the most closed ecosystems must localize their AI infrastructure to operate in China.

Highlights

• Qwen will power Siri, Writing Tools, and Visual Intelligence in the Chinese version of Apple Intelligence.
• Beta testing starts on November 4; the broader release arrives in December with iOS 27.3.
• The initial feature set includes live translation, Genmoji, photo cleanup, natural language photo search, and chat summarization.

Critical take

The report remains surprisingly shallow on the most critical details. It is unclear whether Qwen inference will run on-device or on Alibaba's cloud, how user data will be processed under Chinese laws, and exactly what the 'necessary approvals' required for other Apple platforms entail. This leaves the privacy and regulatory landscape largely in the dark.