Meta has introduced what it describes as the "most capable" open-weight model available in the United States. Named Muse Spark Open, this model is specifically engineered to operate seamlessly on Mac computers.
Featuring 816 billion parameters and a massive 1 million token context window, the new model claims superior performance compared to other popular open-weight alternatives, according to evaluation data released by Meta on Thursday.
Designed as a natively multimodal system for complex coding and agentic labor, Muse Spark Open supports local execution on Apple hardware. However, running it locally requires substantial power; specifically, at least 1 terabyte of RAM is necessary, per Engadget. The model performs optimally on a Mac Studio equipped with the M5 Ultra chip.
Despite the high memory requirements, Muse Spark Open shows great promise as a productive tool for coding and desktop tasks. Meta has integrated the model with its Muse Code terminal and deployed it into various agentic workflows, such as performing "bug fixes, new features, [and] full build" updates for a Mastodon client.
The company also demonstrated the model's utility in professional environments by using it to review and edit footage for DaVinci Resolve, a high-end video editing software. Open-weight models offer distinct benefits over proprietary systems, including greater cost-efficiency, transparency, and flexibility.
Meta is making Muse Spark Open available immediately through several API platforms.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.