Pinned
Apple's on-device models give us free, private and low-latency inference in Swift. But their performance is limited by device constraints (memory, compute, thermal considerations), and the context window size (8k as of iOS 27 beta 3).
Learn how to overcome these limits with a
