Local LLM Field Guide
Search

Beginner8 min read

Get started with LM Studio

Download, load, and chat with a local model through a desktop interface.

Install and download

  1. Check the linked system requirements for your OS and chip, then download LM Studio.
  2. In LM Studio, open Discover and search for Qwen2.5-1.5B-Instruct-GGUF. The publisher’s GGUF repository provides a Q4_K_M variant around 1.1 GB on disk.
  3. Select that variant and download it. Give the app time to finish before switching to Chat.

Load and chat

Open Chat and use the model picker to select the download. Keep the default loading settings for your first run. Send this prompt.

Rewrite this as a friendly reminder: The library books are due Friday.
You’re ready when

The Chat panel shows a generated reminder. Check that the reply still says the books are due Friday.

Try it

Ask the same model to rewrite the reminder in under twelve words. Count the words and verify that “Friday” is still present. Then try a less specific prompt. Compare how clearly each answer communicates the due date.

If loading fails

Check the model’s file format and the app’s hardware support. Try a smaller model or reduce the context allocation. Use a downloaded model for an offline test. Model discovery and downloading require connectivity.

Check your understanding

Why does the model still need to load after its download finishes?

The download puts the weights on disk. Loading allocates memory and prepares them for inference.

Sources

Search the guide