Guides · Updated 2026-10-17

How to install Ollama on a Mac (Apple Silicon)

Install Ollama on macOS, pick models that fit your unified memory, and know where files live.

Apple Silicon Macs are excellent local AI machines because the GPU shares the full pool of unified memory. A 32 GB Mac can load models that would need a 24 GB graphics card on a PC.

Requirements

• macOS Sonoma (14) or newer.
• Apple Silicon (M1 to M5) for GPU acceleration via Metal. Intel Macs run on CPU only and will be slow.

Install

1. Download Ollama.dmg from ollama.com/download.
2. Drag Ollama into Applications and open it.
3. On first launch it offers to link the ollama command into /usr/local/bin. Accept.

Then in Terminal:

ollama run gemma4:12b

How much memory can models use?

macOS reserves part of unified memory for the system, so the GPU can typically use around two-thirds to three-quarters of it. Rough guide:

Speed depends heavily on memory bandwidth: Pro, Max and Ultra chips are much faster than base chips at the same model size. See the GPU picker for estimates.

Setting options

The Mac app reads environment variables via launchctl, e.g.:

launchctl setenv OLLAMA_CONTEXT_LENGTH 8192

Then quit and reopen Ollama.

Where files live

• Models and config: ~/.ollama
• Logs: ~/.ollama/logs (server.log)

Prefer a graphical model browser? LM Studio is also excellent on Macs because it supports Apple’s MLX format. See LM Studio vs Ollama.

Source: Ollama macOS docs.