Skip to content
AI Stack Nest
macOS1 min read · updated 17 Nov 2026

Install Ollama on macOS and run your first model

Step-by-step: check requirements, install Ollama on a Mac, run a first model, find where models are stored and fix common problems.

Before you start

The official Ollama documentation says macOS needs Sonoma (version 14) or newer. On a Mac with an Apple M-series chip Ollama can use both the CPU and the GPU. On an Intel Mac it runs on the CPU only, which is much slower.

Models take memory and disk space. As a rough guide from the Ollama library, Gemma 2 9B is about 5.4 GB and Llama 3.1 8B about 4.9 GB. Close heavy apps before you run a model, because the model has to fit in memory.

Step 1: Install the app

  1. Download the disk image from ollama.com/download.
  2. Open ollama.dmg and drag the Ollama app into your Applications folder.
  3. Open the app. It will offer to link the ollama command into /usr/local/bin; accept so the command works in Terminal.

Step 2: Run a model

Open Terminal and run:

ollama run llama3.1

The first run downloads the model. When you see the >>> prompt, type a question. Type /bye to exit.

Step 3: Useful checks

ollama list            # models installed
ollama ps              # models loaded in memory, on GPU or CPU
ollama stop llama3.1   # unload a model now

By default a model stays loaded for 5 minutes after the last request. You can change this with the OLLAMA_KEEP_ALIVE setting.

Where files are stored

Models are kept in ~/.ollama in your home folder.

Troubleshooting

  • command not found: ollama. Open the app once and accept the command-line link, then open a new Terminal window.
  • Answers feel cut off or forget earlier text. The default context length is 4096 tokens. You can raise it with the OLLAMA_CONTEXT_LENGTH setting, or inside a chat with /set parameter num_ctx 8192. A larger context uses more memory.
  • The Mac gets slow. Choose a smaller model.

Sources

Facts on this page come from the official Ollama documentation for macOS and its FAQ, read on 2026-11-17.

Models used here

More setup guides

Not sure your computer can run a model? Use the will-it-run checker.