China Version of Apple Intelligence Technology Roadmap Revealed! Core Model Shifts from Gemini to Apple's In-House AI

Apple is adopting a different technical route for Apple Intelligence in the Chinese market compared to overseas markets. According to a Reuters report, the China version of Apple Intelligence will be centered on a large language model trained in-house by Apple, rather than directly using Google Gemini or Alibaba's Tongyi Qianwen (Qwen).

This model is being co-trained by Apple and Alibaba. While Apple leads the development of the core large language model, Qwen is expected to be integrated into Apple (AAPL-US) devices sold in China in some form, and Baidu's related technologies may also be included.

This arrangement differs from Apple's strategy in markets like the United States.

The international version of Apple Intelligence uses Apple's own foundational models, but some advanced features rely on Google Gemini technology, and Siri can connect to OpenAI's ChatGPT via plug-ins. Since Gemini, ChatGPT, and Anthropic's Claude are currently unavailable in China, Apple must build a separate technical framework that complies with local regulations.

China's complex AI regulations require Apple to partner with local firms

China enforces strict regulations on generative AI services, and foreign tech companies typically need to collaborate with Chinese firms to launch such products locally. Apple has therefore partnered with Alibaba and is incorporating Qwen and Baidu technologies to meet requirements for model training, content moderation, and regulatory compliance.

Earlier reports suggested that Alibaba may handle the content moderation mechanism for the China version of Apple Intelligence, with these restrictions likely applying to both Apple's in-house model and Qwen-powered features.

Chinese regulators have already approved Apple Intelligence for local use, but the service has not yet officially launched and may still face additional regulatory procedures. If successfully rolled out, Apple could become one of the first foreign tech companies to deploy its own AI model in China.

Launch preparations appear nearly complete, but no official date announced

Apple has not yet announced the exact launch date for the China version of Apple Intelligence, but multiple signs indicate that preparations are in the final stages. In March this year, the service briefly became available in the Chinese market—an incident widely speculated to be an accidental early rollout—suggesting that technical deployment is largely complete.

Additionally, Apple's support website recently published a guide on using Qwen on Mac, which was later removed. This has further fueled speculation that the collaboration is nearing implementation.

China has long been a crucial market for Apple. Although iPhone sales in the region declined significantly in recent years, the trend has improved recently, with the iPhone 17 series performing better in China than Apple initially expected. A successful launch of Apple Intelligence could become a key driver for Apple to sustain growth momentum in the Chinese market.