Meta Launches Muse Spark Open for Local Mac Execution
AI-generated image · Bay Street Wire
The new open-weight model targets complex coding and agentic labor, offering a local alternative to proprietary cloud systems.
Meta has introduced Muse Spark Open, which the company describes as the most capable open-weight model in the U.S. As TechCrunch first reported, the model was engineered to operate locally on Apple hardware. According to the report, the system utilizes 816 billion parameters and a 1 million token context window.
Designed for agentic labor and complex coding, Meta reports that the model has already been utilized for full builds and bug fixes for a Mastodon client via its Muse Code AI coding terminal. Additionally, Meta noted the model's ability to review and crop footage within the high-end video editing software DaVinci Resolve.
While the model is available through various API platforms, running it locally requires significant hardware. TechCrunch notes that, per Engadget, the model requires at least 1 terabyte of RAM and performs optimally on the Mac Studio equipped with the M5 Ultra chip.
Meta positions open-weight models as a more cost-efficient alternative to closed, proprietary models, providing creators and developers with increased flexibility and transparency.