Run LLMs and VLMs natively on Apple Silicon with Swift
Run frontier LLMs and VLMs locally on GPU, NPU, and CPU with day-0 model support
