Details
The predictive code completion feature uses a machine learning model trained specifically on Swift programming code and Apple's software development frameworks. The model is roughly 2GB in size and requires macOS Sequoia, Apple Silicon hardware, and at least 8GB of memory to operate. Unlike competitors such as GitHub Copilot (which sends code to cloud servers), Apple's tool processes everything locally. Apple separately announced Swift Assist, a chat-based coding assistant that runs in the cloud using Private Cloud Compute, but as of early 2026 Swift Assist had not shipped publicly. Xcode 16 also supports third-party large language models such as Claude and GPT-4 via plug-in integrations.
Have evidence about Apple's AI practices? Submit a report.
Report a Sighting →