Apple begins long-delayed Apple Intelligence trial in China, with Alibaba’s Qwen inside
Apple has begun a trial rollout of Apple Intelligence for Chinese users, more than two years after the AI features launched elsewhere, with Alibaba’s Qwen powering some features in place of models unavailable in China.
Apple has begun a long-delayed trial rollout of its Apple Intelligence tools for users in China, opening the features to developers on Wednesday ahead of a public launch planned for December.
The on-device features, first released in the United States and other countries in autumn 2024, had been held back in China by strict regulatory requirements for artificial intelligence and by the fact that models from Google and OpenAI, which Apple uses elsewhere to boost its own capabilities, are not available in the country.
Apple said in a statement at the weekend that Wednesday’s rollout is restricted to developers, with public access to Apple Intelligence services in China to follow in December alongside iOS 27.3. The initial package includes 18 features, from live translation of phone calls to tools for summarising and proofreading messages, according to Channel News Asia’s report of the launch.
Among the features arriving first are Writing Tools, Genmoji, Clean Up in Photos, Notification Summaries, Priority Notifications, Visual Intelligence and Live Translation when paired with AirPods. Apple’s own models will power the core package, while Alibaba’s Qwen model will be integrated into selected experiences, including Siri, Writing Tools and Visual Intelligence, where users can ask questions about photos, PDFs and camera scenes.
The partnership is central to how Apple is navigating China’s rules. Chinese regulations require AI service providers to register with the country’s internet watchdog, and there are requirements to use local data centres and process data domestically. Apple only secured regulatory approval in July, after reportedly holding talks with Chinese firms including Alibaba, Baidu, ByteDance and Tencent about partnerships on Apple Intelligence.
The rollout is deliberately partial. The rebuilt Siri AI assistant shown at Apple’s Worldwide Developers Conference in June 2026 is not included in this phase, and Image Playground is also unavailable in China at launch. Support for Mac and iPad will depend on receiving separate regulatory approvals, and the features require an iPhone 15 Pro or newer model, including the iPhone 16 series, iPhone Air and iPhone Duo, with the device region set to China and the language set to English or Simplified Chinese.
For Apple, China is one of its most important markets, and the delay has left it playing catch-up to homegrown smartphone makers whose AI features have been available for much longer. Apple has largely stayed out of the global AI investment frenzy, and its approach in China — a local partner model, built around Alibaba, rather than its usual global stack — shows how far it has had to adapt to operate there.
The timing also matters for Alibaba. Qwen, its family of language models, gains a high-profile showcase inside one of the world’s most popular phone ecosystems, and reporting on the announcement said Qwen would be used only after user approval, with requests and responses not used to train its models. That opt-in design is likely to be watched closely by regulators and by privacy-conscious users alike.
Developers in China will now test the features in the beta before the December public release, and their feedback could still change details of the final package. Apple has not given an exact December date, and features shown in a developer beta can change before general release.
If the December launch goes ahead as planned, Chinese iPhone users will finally get the core Apple Intelligence package more than two years after it arrived elsewhere — but in a distinctly Chinese form, with a local model inside, a narrower feature set, and the next generation of Siri still to come.
More in Technology & AI
Photo: Tencent headquarters (file image)
Technology & AITencent releases Hy4 AI model with cybersecurity training limited to defensive uses
Chinese technology company Tencent has released Hy4, a new open-source artificial intelligence model whose cybersecurity training was deliberately limited to defensive work, in an effort to reduce the risk that it could help carry out cyberattacks.
Photo: India Gate in New Delhi (file image)
Technology & AIIndia launches three homegrown AI models, EKA-1 and Dhir 2B and 5B, under the IndiaAI Mission
India’s electronics and IT ministry has released three indigenous foundation models under the IndiaAI Mission: EKA-1, a 26.3-billion-parameter model for Indian languages, banking and cybersecurity, and the compact Dhir 2B and 5B models designed to run offline on mobile phones.
Photo: A locked computer, illustrating cybersecurity (file image)
Technology & AICrowdStrike moves to buy Fiddler AI and Dreadnode in reported $495 million bet on AI agent security
CrowdStrike has signed agreements to acquire AI agent observability firm Fiddler AI and AI security testing startup Dreadnode, adding monitoring and pre-deployment testing to its Falcon platform alongside a new AI kill switch.