llama.cpp
The C++ runtime that made GGUF and Metal inference normal.
When to use it
Use this when you want the widest model and format compatibility (GGUF) plus cross-platform support — on Mac alone, MLX is usually faster.
Alternatives in Runtimes
MLX
Apple's own array framework, built so unified memory stops being a compromise.
MLX LM
The package behind every `mlx_lm.generate` command on this site.
MLX Swift
MLX's tensor core, ported so Swift apps don't need a Python bridge.
MLC LLM
A compiler-first runtime that targets Metal, CUDA and mobile from one graph.
Candle
Hugging Face's minimalist Rust tensor library, Metal backend included.