Meta and Apple Collaborate to Optimize Muse Spark Open AI Model for Mac Studios

Image: Company official website / logo · Source

Meta and Apple jointly announced that Meta's Muse Spark Open AI model can efficiently run on dual Apple M5 Ultra Mac Studios using Apple's MLX framework, marking a rare collaboration between the two tech rivals.

In a notable partnership, Meta and Apple have collaborated to enhance the performance of Meta's latest AI model, Muse Spark Open, on Apple's Mac Studio hardware. Meta unveiled the Muse Spark Open AI model, describing it as the most capable open-weight AI model developed in the United States to date. This model distinguishes itself by being the first of its class to run locally on desktop machines using two Apple M5 Ultra Mac Studios, facilitated by Apple's new MLX framework.

Traditionally competitors, Meta and Apple have publicly supported each other's roles in this launch. Meta CEO Mark Zuckerberg highlighted that Muse Spark Open operates exceptionally well on the new Mac Studios thanks to collaboration with Apple’s team. Apple CEO John Ternus expressed enthusiasm for the ability to run a model of Muse Spark Open’s size and sophistication efficiently on desktop hardware, emphasizing the Mac Studio's advancement in enabling on-device AI computations.

Meta’s chief AI officer Alexandr Wang remarked on the renewed cooperative spirit between the companies, noting that their collaboration has been ongoing for several months. The Muse Spark Open model is characterized as a natively multimodal AI with 816 billion parameters, designed to excel in long-duration tasks, complex software engineering challenges, and multi-step reasoning processes.

Complementing the announcement, Apple released technical documentation describing MLX, an array programming framework tailored for machine learning workloads on Apple Silicon. MLX plays a crucial role in optimizing the deployment and execution of Muse Spark Open and other open AI models on Apple's hardware.

This cooperation marks a significant moment in AI development and hardware optimization, enabling users to leverage powerful AI capabilities directly on consumer-grade desktop computers without reliance on cloud infrastructure. Meta's approach with Muse Spark Open, combined with Apple's hardware and software advancements, points toward more accessible and efficient AI applications in personal computing environments.