LM Studiorecommended
A polished desktop app for local LLMs: GUI model search, one-click downloads, a chat UI and a local server.
Official site ↗LLM Runner Picker
New to running models locally? Pick your OS and experience level — and we will recommend a runner that matches.
4 of 4 runners support your setup
Ollama
The simplest way to run local LLMs: one install, one command, an OpenAI-compatible API. Our default recommendation for beginners.
Official site ↗llama.cpp
C/C++ inference engine for LLMs: the best performance per gigabyte of VRAM, full GGUF quantization tooling, and the backbone of Ollama and LM Studio.
Official site ↗vLLM
High-throughput LLM serving engine with PagedAttention and continuous batching — the standard for multi-user and API workloads, with an OpenAI-compatible server.
Official site ↗