Inference runtime
llama.cpp
Portable local inference runtime for GGUF models across CPU and accelerator backends.
Hardware
Supported hardware records
Mac Studio M3 Ultra · 512 GB
Apple · Metal backend available.
DGX Spark (GB10)
NVIDIA
Dell Pro Max with GB10 · 128 GB
Dell · NVIDIA CUDA backend
Mac Studio · M4 Max · 128 GB
Apple · Apple silicon / Metal backend
MacBook Pro 16-inch · M5 Max · 128 GB
Apple · Apple silicon / Metal backend
MacBook Pro · M4 Max · 128 GB
Apple · Apple silicon / Metal backend
RTX PRO 6000 Blackwell Workstation Edition
NVIDIA
Mac mini · M4 Pro · 64 GB
Apple · Apple silicon / Metal backend
MacBook Pro 16-inch · M5 Pro · 64 GB
Apple · Apple silicon / Metal backend
RTX 6000 Ada Generation
NVIDIA
GeForce RTX 5090
NVIDIA · CUDA backend available.
MacBook Air · M5 · 32 GB
Apple · Apple silicon / Metal backend
MacBook Pro 14-inch · M5 · 32 GB
Apple · Apple silicon / Metal backend
GeForce RTX 3090
NVIDIA
GeForce RTX 4090
NVIDIA · CUDA backend available.