Bottom line: LM Studio is a desktop app that lets you download open AI models and chat with them entirely on your own computer — and since July 2025 it has been free for both home and work use, with no licence form to fill in. There are no message caps and no per-token charges. The real cost is your hardware: the model runs on your machine, so speed depends on your RAM and graphics card.
What does LM Studio actually do?
LM Studio gives you a chat window, a built-in model catalogue, and a local server in one app for Windows, Mac, and Linux. You browse models such as Llama, Qwen, Gemma, and DeepSeek inside the app, download the one you want, and start chatting. After the download, everything runs locally — you can disconnect from the internet and keep working, and your prompts never leave the device. For developers, the app can also serve the loaded model through a local, OpenAI-compatible API, so existing tools and scripts can point at your computer instead of a cloud endpoint.
What is free, and what is paywalled?
| Plan | Price | What you get |
|---|---|---|
| LM Studio (home and work) | $0 | Full desktop app, model downloads, local chat, local API server |
| Enterprise | Paid, contact sales | Team administration such as single sign-on and model controls, plus priority support |
For an individual, there is effectively nothing to unlock: the free app is the whole product. What you give up versus a cloud chatbot is raw power — a model that fits on a typical laptop is smaller than the frontier models behind paid subscriptions, and large models need serious memory. Plan on a machine with plenty of RAM; small models run acceptably on modest hardware, while bigger ones reward a capable GPU.
Pros and cons
Pros
- Completely free for personal and work use — no subscription, no usage caps
- Private by design: chats run locally and work offline after download
- Beginner-friendly model browser with one-click downloads
- Local API server compatible with OpenAI-style tooling
- Wide model choice: Llama, Qwen, Gemma, DeepSeek, and more
Cons
- Quality and speed are limited by your own hardware
- Model files are large — downloads run to several gigabytes each
- The app itself is closed-source, unlike some local rivals
- No built-in cloud fallback when a model is too big for your machine
Who is LM Studio best for?
LM Studio is best for privacy-conscious users, developers testing models without API bills, and anyone who wants an AI assistant that works on a plane, on a job site, or anywhere else without a connection. If you need the smartest possible answers and do not care where your text goes, a cloud chatbot is still stronger. If you want capable help that stays on your machine and costs nothing to run, this is the easiest place to start.
Frequently asked questions
Is LM Studio really free for work?
Yes. Since 8 July 2025 the makers removed the separate commercial licence step — you can simply download the app and use it at work. Paid Enterprise options only add team administration and support.
Does LM Studio work offline?
Yes. You need the internet to download a model once; after that, chatting and the local server work fully offline.
What computer do I need?
A recent laptop with 16GB of RAM handles small models comfortably. Larger models want more memory and, ideally, a dedicated GPU or an Apple Silicon Mac with unified memory to spare.
Which models can I run?
Any compatible open model in the in-app catalogue, including families such as Llama, Qwen, Gemma, Mistral, and DeepSeek, in sizes that fit your hardware.
Verdict: The most approachable free way to run AI locally. Download it, pick a mid-size model, and you have a private chatbot that never sends you a bill.