Apple introduces Core AI for on-device generative models

Apple has introduced Core AI, a new framework aimed at running generative AI directly on Apple silicon. The framework is positioned as the successor to Core ML and supports both converted PyTorch models and pre-optimized open-source models.

The move gives developers a more direct path to building AI features that run locally instead of depending entirely on cloud inference. That could matter for latency, privacy, and cost-sensitive apps, especially on devices with capable neural and GPU hardware.

Core AI also signals that Apple wants its developer ecosystem to treat on-device generative models as a first-class platform capability, not a niche optimization.