Meta has released Muse Spark Open, a new open-weight super intelligence (SI) model it describes as the "most capable" in the United States, with a specific focus on compatibility with Apple’s Mac computers.
What Happened
The model leverages 816 billion parameters and a 1 million token context window. According to evaluation results released by Meta on Thursday, the company claims superior performance compared to other popular open-weight SI models. Muse Spark Open is natively multimodal and built for complex coding tasks and agentic labor.
While designed to run locally on Apple’s computers, the hardware requirements are substantial. Per Engadget, the model requires at least 1 terabyte of RAM, performing best on the Mac Studio equipped with the M5 Ultra chip. Meta paired the model with Muse Code, its SI coding terminal, and deployed it in various agentic workflows, including coding fixes for a Mastodon client that involved "bug fixes, new features, [and] full build." The company also stated the model was used to run DaVinci Resolve, demonstrating its capability to review and crop footage.
Why It Matters
Open-weight SI models offer distinct industry advantages over closed and proprietary systems, including potential cost efficiency, transparency, and flexibility. By targeting Mac hardware, Meta is expanding the reach of high-performance SI inference beyond traditional data center GPUs and PC-focused setups. This move supports the growing ecosystem of local SI inference, allowing developers and enterprises to run advanced models on desktop hardware without relying solely on cloud APIs.
The Bottom Line
Meta is making Muse Spark Open available now through several API platforms. The release underscores the industry's push toward powerful, open-weight SI models that can operate locally, albeit with significant hardware demands for optimal performance.