Ollama v0.40.8 runs Recursive Flint, a 40B coding model, on macOS (October 21, 2026)

What you get

Ollama v0.40.8, released October 21, 2026, adds Recursive Flint. It is a 40B mixture-of-experts coding model with about 2.5B active parameters and a 262K token context window.

Prerequisites

• A Mac (Ollama says other GPUs come in a later release)
• Ollama v0.40.8 or newer
• No API key needed: it runs on your computer

Steps

1. Update Ollama to v0.40.8 or newer from the Ollama download page, then check with ollama --version.
2. Run the model: ollama run flint
3. The first run downloads the model, so give it a few minutes.

Test it

ollama run flint "Write a Python function that checks if a string is a palindrome."

You should get a short Python function back.

Not confirmed

• The release notes do not give the download size or the memory the model needs.