MLX
Apple's open-source array and machine-learning framework, built for Apple Silicon's unified memory. With MLX-LM it runs language models efficiently on Macs, and is often the fastest path to good local performance on Apple hardware.
Strengths
- Excellent performance on Apple Silicon, using unified memory well
- Active community porting current models to the MLX format
- MIT licensed, and developed by Apple
Weaknesses
- Apple Silicon only, so no use on other hardware
- A framework and toolkit rather than a turnkey app
- A smaller model and tooling ecosystem than llama.cpp or CUDA
At a glance
- Licence
- MIT
- Pricing
- Free
- Platforms
- macOS
- GPU required
- No
Links
Works with
Entry last verified 30 July 2026.