Local LLM Field Guide
Search

Beginner10 min read

Run your first model with Ollama

Install Ollama, download a small model, and get a reply from your terminal.

Install and start Ollama

You’ll need a terminal, internet access for the first download, and spare disk and memory. The starter model here is qwen2.5:1.5b. Ollama lists this particular download at about 1 GB. Running it also takes memory for the app and context.

Choose your operating system on the official Ollama download page.

macOS
Install and open Ollama, then follow its prompt to make the CLI available. Use Terminal.
Windows
Run the official Windows installer, then open a new PowerShell window. Start the Ollama app from Start if it is not already running.
Linux
Follow the official installation instructions for your distribution. If an Ollama service is not already active, run ollama serve in a separate terminal and keep it open.

Check that the command is available in your new terminal.

ollama --version

Send a prompt

Run the command below. The first run downloads the model. When the chat prompt appears, type “Give me two short names for a plant-watering reminder.” Type /bye to leave the chat.

ollama run qwen2.5:1.5b
You’re ready when

You see two suggested names and the chat prompt returns for another message. After leaving the chat, ollama list should show qwen2.5:1.5b. While the model is loaded, ollama ps shows its processor placement.

Check your setup

  1. Once the download is complete, disconnect from the internet and run the same model again.
  2. Try a fresh sample prompt. A generated reply confirms that local inference worked offline with this setup.
  3. Reconnect before downloading another model or opening an online source.
If the command fails

For “command not found,” check that the installer configured the CLI and reopen the terminal. For a connection error, confirm that the Ollama app or Linux service is running. For a loading error, close memory-heavy apps or try qwen2.5:0.5b and use that exact tag in later examples.

Check your understanding

Which command lists files on disk, and which command shows currently loaded models?

ollama list lists downloaded models. ollama ps shows models currently loaded.

Sources

Search the guide