Ollama v0.40.0 Defaults to MLX Runtime for Models on Apple Silicon

Core summary

One sentence to understand this update

Ollama `v0.40.0` introduces a significant change where models with supported architectures will automatically run on the MLX runtime when deployed on Apple Silicon devices.

Impact & opportunity

What this could mean

Apple Silicon users can expect improved performance and efficiency for local LLM inference with Ollama due to the default MLX integration.