Meta and Apple Unite to Run Muse Spark Open on Mac
According to a recent report by MacRumors, Meta and Apple have forged an ice-breaking partnership. This rare collaboration allows the Muse Spark Open model to run locally on two Mac Studio desktop computers. Experts describe this system as the most powerful American open-weight model to date.
A Revolutionary AI Architecture
Meta recently launched the Muse Spark Open AI model. It features a staggering 816 billion total parameters. Out of these, 42 billion function as active parameters. Furthermore, it supports an expansive one-million-token context window. Engineers designed this system primarily for complex workflows. These include advanced coding, autonomous agent tasks, multi-step reasoning, and software engineering.
An Unprecedented Industry Alliance
However, the unexpected alliance between Meta and Apple has captured the most industry attention. For years, these two technology giants maintained a fiercely adversarial stance. They clashed frequently over business models, privacy policies, and product ecosystems. Therefore, this joint technical endeavor represents a remarkably rare public collaboration.
Harnessing the Power of Apple Silicon
The official demonstration utilized two Apple Mac Studio workstations. Each machine featured the formidable M5 Ultra chip and 512 gigabytes of unified memory. By leveraging the latest Apple MLX framework, they successfully executed the massive AI model.
Both Meta CEO Mark Zuckerberg and Apple CEO John Ternus confirmed this partnership on social media. Ternus proudly stated that the Mac Studio continues to push the boundaries of on-device artificial intelligence. Meanwhile, Zuckerberg noted that the new model performs exceptionally well on the latest Apple hardware.
Nevertheless, this powerful local setup carries a premium price tag. Apple has not yet officially released the 512-gigabyte version of the Mac Studio. Consumers expect this high-end configuration to launch later this month. Meanwhile, the current 256-gigabyte model already costs more than $10,000.
