Meta and Apple Collaborate on Muse Spark Open, an AI Model Designed to Run on Mac Studios
Meta and Apple collaborated for several months to optimize the 816-billion-parameter model for local inference. It runs on two Apple M5 Ultra Mac Studios through Apple’s MLX framework, with CEOs Mark Zuckerberg and John Ternus publicly endorsing the partnership.
Meta announced the Muse Spark Open AI model in October 2026, describing it as the “most capable US open-weight model to date.” The launch marks a notable collaboration between Meta and Apple — companies that have typically been rivals — with both publicly supporting each other around the release.
The model is notable for being, according to Meta, “the first model in its class” capable of running locally “on your desktop using two Apple M5 Ultra Mac Studios with the latest MLX framework.” MLX is Apple’s array framework designed for machine learning tasks on Apple Silicon hardware.
Meta CEO Mark Zuckerberg credited the partnership directly, writing on X: “Muse Spark Open — weights releasing today — runs great on new Mac Studios thanks to work with @johnternus and Apple.” Apple CEO John Ternus responded: “Amazing to run a model of Muse Spark Open’s caliber and scale this efficiently, right from your desk. Mac Studio keeps pushing the boundaries of on-device AI!”
Meta chief AI officer Alexandr Wang told Axios that the two companies had been collaborating for several months. “We’re excited for mommy and daddy to be playing nice again,” Wang said.
Muse Spark Open is described by Meta as a “natively multimodal model” with 816 billion total parameters. The company says it excels at “long-horizon tasks, complex software engineering, and multi-step reasoning.” Apple also published a technical post covering MLX, Muse Spark Open, and two other open models.
The partnership suggests that major AI model releases may increasingly depend on close coordination between software developers and hardware manufacturers, particularly as on-device AI performance becomes a competitive focus.