LM Studio

What it does

LM Studio is a desktop application for discovering, downloading, and running LLMs locally. It wraps llama.cpp under the hood but exposes a clean GUI — a ChatGPT-style chat interface, a model browser connected to Hugging Face, and a developer mode that serves an OpenAI-compatible local API.

3.8 million+ downloads make it one of the most popular ways to run local models without touching the command line.

Installation

Download from lmstudio.ai for macOS, Windows, or Linux. No package manager required.

Key features

Model browser

LM Studio connects to Hugging Face’s model hub and surfaces GGUF-format models with recommended quantization levels based on your hardware. Select a model, click Download — it handles the rest.

Chat interface

A clean ChatGPT-style chat window. Multiple conversation threads, system prompt editor, temperature/top-p sliders, token count display.

Local API server

Enable Developer Mode and LM Studio runs a local server on port 1234:

from openai import OpenAI

client = OpenAI(base_url="http://localhost:1234/v1", api_key="lm-studio")
response = client.chat.completions.create(
    model="local-model",  # use the model currently loaded in LM Studio
    messages=[{"role": "user", "content": "Hello!"}]
)

The API is compatible with any OpenAI SDK — Python, TypeScript, curl.

Multi-model loading

LM Studio 0.3+ supports loading multiple models simultaneously (hardware permitting) and switching between them without reloading.

vs Ollama

LM Studio is the right choice if you want a GUI-first experience. Ollama is better for server/headless setups and scripted automation.