Apple Rebuilds Apple Intelligence on Google Gemini Foundation Models

Apple has overhauled Apple Intelligence around a new set of foundation models co-developed with Google using Gemini technology. The models run across on-device hardware and Apple’s Private Cloud Compute servers, and Apple is pitching the partnership as a step-change in reasoning and multimodal capability, including image generation, advanced photo editing, and visual question answering. A subset of unspecified devices will get a more powerful variant that adds speech generation and stronger natural language understanding.

A new system orchestrator coordinates features across apps, adapting responses to whatever the user is doing. Apple paired the announcement with a familiar privacy pitch, contrasting itself with rivals it accused of racing ahead recklessly and reiterating that requests are processed without exposing data to Apple or third parties, with outside auditors able to verify the claims.

The move is notable because Apple is now leaning on a direct competitor’s model lineage for its flagship AI platform, even as Google rebrands its own Android AI as Gemini Intelligence and pushes Gemini deeper into products like Google Maps for CarPlay.