Apple Intelligence Opens Up — iOS 27 Lets You Pick Claude, Gemini, or ChatGPT
At WWDC 2026, Apple opened Apple Intelligence to third-party models — iOS 27 lets users choose Claude, Gemini, or ChatGPT as the engine behind Siri and Writing Tools.
Every so often a platform decision reshapes the whole landscape around it. At WWDC 2026 on June 8, Apple made one: Apple Intelligence is opening to third-party AI models. With iOS 27, iPadOS 27, and macOS 27, users will be able to designate Anthropic's Claude, Google's Gemini, or OpenAI's ChatGPT as the default provider behind Apple's built-in intelligence features. For a company long associated with tightly integrated, single-vendor experiences, this is a notable and genuinely exciting move toward openness.
Apple Intelligence Becomes a Platform, Not a Walled Garden
The headline for anyone tracking the convergence of large language models and personal computing is simple: Apple is treating the AI model as a user-selectable component rather than a fixed part of the operating system. The same way you can already choose a default browser or email client, iOS 27 will let you choose which model powers Apple Intelligence. Each option carries its own distinct voice and strengths, and the choice is yours to make and change.
How the Multi-Model Extensions System Works
The mechanism is a new Extensions framework. When a supported AI app — Claude, Gemini, or ChatGPT — is installed on the device, the rebuilt Siri can hand questions off to it through a clean routing layer. That means a request can be answered by Apple's on-device intelligence for privacy-sensitive tasks, or routed to a frontier model when a query benefits from deeper reasoning. Features such as Writing Tools and Image Playground can likewise draw on the model you prefer, so the choice extends across the system rather than living only inside Siri.
A Rebuilt Siri as a Standalone App
Apple also previewed a substantially rebuilt Siri that arrives as a standalone app on iPhone, iPad, and Mac. It surfaces inside the Dynamic Island with a quick "search or ask" prompt, supports multi-step commands, keeps a persistent conversation history synced across devices via iCloud, and lets you attach images and documents directly to a request. The result is an assistant that feels less like a one-shot voice command and more like a continuous, context-aware collaborator.
Why Letting Users Choose Their AI Model Matters
From an AI-strategy standpoint, this is the most consequential detail of the keynote. Model choice turns Apple Intelligence into a marketplace of capabilities: developers and users are no longer betting on a single provider's roadmap, and the models themselves are incentivized to compete on quality, voice, and trust. It also respects that different people want different things from an assistant — some prioritize Claude's careful, thorough style, others Gemini's breadth, others ChatGPT's familiarity. Designing for that diversity, rather than against it, is the right instinct for a device used by more than a billion people.
The iOS 27 Rollout Timeline
Developer betas for the new operating systems launched immediately after the June 8 keynote, with public betas expected in July and full releases scheduled for the fall. The Apple Intelligence model-choice feature and the rebuilt Siri will roll out incrementally rather than all at once — a measured approach that fits the scale of the change. For everyone watching AI move from novelty to everyday infrastructure, Apple's embrace of an open, multi-model future is one of the clearest signals yet of where personal computing is headed.
Sources: Apple WWDC 2026 keynote (June 8, 2026); TechTimes, "Apple WWDC 2026" (June 8, 2026); MacRumors WWDC 2026 guide (June 2026); Tom's Guide WWDC 2026 live coverage (June 8, 2026).