How to Run LLMs Using LM Studio on Linux (Beginner’s Guide)

Learn what LM Studio is, why it’s ideal for Linux, and how to install and run large language models (LLMs) on any Linux distribution. A complete, evergreen guide for beginners.

7 February 2026 by

Suraj Barman

What is LM Studio?

LM Studio is an open‑source desktop application that simplifies the download, management, and inference of large language models (LLMs) on local hardware. It provides a graphical interface, model catalog, and built‑in inference engine, removing the need for complex command‑line setups.

Why Use LM Studio on Linux?

Full control over hardware resources (CPU, GPU, RAM).
Better privacy – data never leaves your machine.
Linux offers robust package managers and driver support for NVIDIA/AMD GPUs.
Community‑driven updates and extensive documentation.

How to Install LM Studio on Linux

Update your system: sudo apt update && sudo apt upgrade -y (Debian/Ubuntu) or the equivalent for your distro.
Install required dependencies:
- Git: sudo apt install git
- Node.js (for the UI): sudo apt install nodejs npm
- GPU drivers (optional but recommended): NVIDIA driver sudo apt install nvidia-driver-525 or AMD driver package.

Download the latest LM Studio release from the official GitHub releases page or use the script:

curl -L  -o lmstudio.tar.gz
mkdir -p $HOME/lmstudio && tar -xzf lmstudio.tar.gz -C $HOME/lmstudio --strip-components=1

Add LM Studio to your PATH (optional):

echo 'export PATH="$HOME/lmstudio:$PATH"' >> ~/.bashrc
source ~/.bashrc

Launch LM Studio:
```
lmstudio
```
or run the binary directly from the installation folder.

How to Run an LLM with LM Studio

Open LM Studio and sign in (or use offline mode).
Navigate to the “Model Catalog” tab.
Select a model that matches your hardware (e.g., LLaMA‑2‑7B for 16 GB VRAM).
Click “Download” – LM Studio handles model files and verification.
After download, go to “My Models” and click “Load”.
Configure inference settings:
- Device: CPU, CUDA, or ROCm.
- Batch size, temperature, max tokens, etc.
Press “Start Inference” and use the built‑in chat window or connect via the API endpoint (default ).

Troubleshooting Common Issues

GPU not detected: Verify driver installation and that nvidia-smi (or rocminfo) returns your GPU.
Out‑of‑memory errors: Reduce the model size, enable 8‑bit quantization, or increase swap space.
Model download stalls: Check your internet connection, disable VPNs, or use a mirror URL.
API connection refused: Ensure the LM Studio server is running and the firewall allows port 1234.