The model represents a significant shift in local AI deployment, functioning best on a Mac Studio equipped with the M5 Ultra chip. Meta demonstrated the system's utility by pairing it with the Muse Code terminal to execute automated bug fixes and build updates for a Mastodon client. Beyond software development, the AI proved capable of handling creative workflows, including reviewing and cropping footage within DaVinci Resolve.
In section Startups & Technology
Meta Unveils Muse Spark Open, a Heavyweight AI for Mac Hardware
Meta has introduced Muse Spark Open, an open-weight AI model packing 816 billion parameters and a 1-million-token context window. Designed specifically for complex coding and agentic tasks, the model aims to bring high-end performance to local Apple hardware, provided the user can supply the necessary 1 terabyte of RAM.
By releasing Muse Spark Open as an open-weight model, Meta offers developers a transparent alternative to proprietary systems. This architecture avoids the black-box limitations of closed models while maintaining cost-efficiency for large-scale agentic labor. The company has made the model available immediately through various API platforms for those seeking to integrate it into their own technical environments.
Comments (0)
No comments yet. Be the first!