TL;DR: Ollama's latest update enables Gemma4 to process images and audio on the MLX engine, expanding local multimodal AI capabilities for developers.
Summary: Ollama v0.33.3 introduces support for multimodal inputs (images and audio) for the Gemma4 model when running on the MLX engine. This update enhances the local inference capabilities of Gemma4, allowing it to handle more complex data types directly on Apple Silicon.
Why it matters: This significantly boosts local multimodal AI development, enabling indie developers to build and experiment with advanced applications without cloud dependencies. Explore integrating Gemma4 with image and audio inputs for local AI projects.
Source: github_releases