Device models and developer access
Apple makes devices and operating systems that distribute AI to end users. Its filing identifies the California corporation as established in 1977. The June 2024 Apple Intelligence announcement combined on-device models with Private Cloud Compute and a separate ChatGPT integration for selected requests. [1][2]
The Foundation Models framework, announced in June 2025, gave developers an interface to the on-device Apple Intelligence model, including guided generation and tool calling. Access through an operating-system framework is a different distribution mechanism from publishing model weights under an open license. [3]
Siri's Gemini collaboration and beta rollout
Apple and Google's January 2026 joint statement said the next generation of Apple Foundation Models would be based on Gemini models and Google's cloud technology. Apple's September 14 Siri AI release explicitly confirmed the collaboration while retaining on-device and Private Cloud Compute execution. The September announcement made the model collaboration explicit. [4][5]
Siri AI began that September rollout as an English beta on supported devices. At launch it excluded China and EU iPhone, iPad and Watch access, and some server-side features had daily limits with expanded paid access promised later. Those are the release's initial conditions, rather than proof that every announced language, region or paid tier was available immediately. [5]
Apple also announced a leadership transition in April 2026, specifying September 1 for John Ternus to become CEO and Tim Cook executive chairman. That dated corporate announcement is separate from the evidence for which Siri features actually reached users. [6]