Robotics & Physical AIOct 7, 2026

Ollama 0.40 runs models on Apple's MLX by default on Apple Silicon

Ollama 0.40.0, released on 6 October 2026, automatically runs supported model architectures on Apple's MLX runtime on Apple Silicon, including Qwen 3.8, Gemma 4, Qwen 3.6 and Qwen 3.5. The decision models Nimble, tev1, Clef and Clef Flash also run on MLX, and the MLX runner gains its first embedding model, EmbeddingGemma 2, with multimodal embeddings through /api/embed.

What it means The runtime under existing Mac setups changes on upgrade, so re-check speed and memory use for the models you already run.

Where it came from Ollama

Back to the Stream