Ollama Now Runs Faster on Macs Thanks to Apple's MLX Framework

MacRumors • March 31, 2026 • technology

Key Points:

Ollama has released an update for its AI model app that leverages Apple's MLX machine learning framework, significantly improving performance on Macs with Apple silicon.
The update boosts prompt processing speed by approximately 1.6 times and nearly doubles response generation speed, with the biggest gains seen on Macs featuring M5-series chips.
Enhanced memory management in the update aims to improve responsiveness during prolonged use of AI-powered coding tools and chat assistants.
Users running personal assistants like OpenClaw or coding agents such as Claude Code, OpenCode, or Codex on macOS are expected to benefit most from these performance improvements.