Meta Launches Muse Spark Open Model for Macs
Meta has launched an open-weight artificial intelligence model named Muse Spark Open, which the company describes as its most capable open-weight release in the United States, according to TechCrunch. Featuring 816 billion parameters and a 1 million token context window, the natively multimodal model is built for complex coding tasks and agentic labor workflows.
Unlike models restricted entirely to cloud data centers, Muse Spark Open has been designed to run locally on Apple computers. However, running the system locally requires substantial hardware resources, including a minimum requirement of one terabyte of RAM. Meta notes that the architecture performs best on a Mac Studio equipped with an M5 Ultra chip.
Meta paired the new model with its Muse Code terminal and deployed it across agentic workflows, such as executing bug fixes and building features for a Mastodon client. The company also tested the system running high-end video editing software like DaVinci Resolve, where it handled tasks including footage review and cropping.
Open-weight models offer developers transparency, flexibility, and cost efficiency that proprietary systems often lack. Meta is making Muse Spark Open available immediately through multiple API platforms for builders and enterprise users.
Based on reporting by techcrunch.com.
