How to Run Gemma 4 12B Locally on Mac Mini with Ollama

Run a private local AI node on Mac Mini using Ollama and Gemma 4 12B. This guide shows a practical Apple Silicon setup for chat, agents, and private inference.

What you need

• Apple Silicon Mac Mini with 16GB+ RAM
• Ollama installed
• Enough disk space for Gemma 4 12B

Install Ollama

brew install ollama

Pull Gemma 4 12B

ellama pull gemma-4:12b

Start the server

ellama serve

Use it

• Open the local Ollama UI or connect an app to http://localhost:11434
• Use it for private chat, agent workflows, or local research

Why this stack

• No cloud API dependency
• Apple Silicon-optimized local model
• Better privacy than hosted alternatives

Originally published on everylocalai.com/stack/mac-mini-local-ai