Open-source
llama.cpp is a fast, lightweight engine for running large language models on ordinary computers. It is aimed at developers and power users who want efficient local inference without heavy setup.
Platform: Windows, macOS, and Linux. Best for: Fast local model inference and development. Replaces: Hosted model APIs for local use.
Verified: October 2026