What the context limit of apple intellinge?

Apple Intelligence, as discussed in the Platforms State of the Union, is designed to run powerful generative models on Apple devices, leveraging the capabilities of Apple Silicon. The focus is on running as much as possible on-device to ensure low latency, better user experience, and enhanced privacy. The on-device foundation model is a large language model that is powerful yet small enough to run locally.

For more advanced features that require larger models, Apple extends its capabilities to the cloud using private cloud compute, ensuring that users' personal data remains private. This approach allows Apple Intelligence to handle complex data while maintaining privacy and security.

If you are interested in the technical details of how Apple Intelligence was developed and its integration into Apple's platforms, you can watch the Platforms State of the Union session.