Macs Just Got Faster AI That Runs Without the Cloud

A new software update (version 0.40.0) is making AI run faster and more privately on Mac computers. On Macs with Apple Silicon chips (the M1, M2, M3, and newer), AI models that support a technology called MLX will now run directly on your computer instead of in the cloud. This means you don't need an internet connection to use them, your data stays on your device, and you won't pay per-use fees. It's like having a mini AI supercomputer in your laptop.

The update adds several new AI models you can run locally, including Gemma 4, Qwen 3.6, and Qwen 3.5. These are general-purpose models that can chat, write, and answer questions. It also adds decision-making models (Nimble, Tev1, Clef, and Clef-Flash) that can help with tasks like choosing between options or making recommendations. Plus, there's a new embedding model called EmbeddingGemma-2, which helps computers understand the meaning of text for search and organization.

Why does this matter to you? Running AI locally means your personal data — emails, documents, photos — never leaves your Mac. That's a big win for privacy. It also means AI features in apps can work offline and respond instantly, without waiting for a server. And you won't rack up cloud computing bills. For everyday users, this could mean smarter, faster, and safer AI in the tools you already use.

The catch? Not all Macs are supported — you need an Apple Silicon chip (from late 2020 onward). Also, not every AI model works with MLX yet, but the list is growing. The developers say they'll keep testing and adding more models. So if you have a recent Mac, expect your AI apps to get a quiet upgrade that makes them more capable and private.