Model Architecture and Desktop Execution
Meta announced the Muse Spark Open AI model, claiming it is the most capable US open-weight model to date. The company also states it is the first model in its class capable of running on desktop hardware using two Apple M5 Ultra Mac Studios with the latest MLX framework.
Meta and Apple team up to make Muse AI models run better on Macs
The leadership of both companies publicly supported the development and integration of Muse Spark Open.
Hardware Integration
The new Muse Spark Open AI model is configured to operate directly on M5 Ultra Mac Studios.
Despite historical market rivalry, Meta and Apple collaborated closely on the product launch and public promotion.
Executive Statements
Meta's leadership noted on social media that the released model weights function effectively on new Mac Studios following engineering efforts alongside Apple.
Apple executives highlighted the efficiency of running a model of Muse Spark Open's caliber and scale directly from a desktop environment.
Representatives from Meta confirmed that the engineering teams maintained an active collaboration over several months.
Technical Specifications
The system is classified as a natively multimodal model featuring 816 billion total parameters designed for long-horizon tasks, complex software engineering, and multi-step reasoning. Concurrently, Apple released technical documentation regarding MLX, an array framework optimized for machine learning tasks on Apple Silicon.
