Meta Releases Muse Spark Open, an Advanced AI Model Optimized for Macs

Image: TechCrunch · Source

Meta unveiled Muse Spark Open, an 816-billion-parameter open-weight AI model designed for Apple silicon Macs, supporting complex coding and multimodal tasks locally with substantial memory requirements.

Meta has introduced Muse Spark Open, a powerful new open-weight artificial intelligence model designed specifically to run on Apple silicon Macs. According to Meta’s announcement and a report by TechCrunch dated October 15, 2026, this model represents one of the most capable open-weight AI systems currently available in the United States. Featuring 816 billion parameters and a one million token context window, Muse Spark Open is engineered to excel at complex coding projects and multimodal tasks.

Uniquely, Muse Spark Open is optimized to operate locally on Macs, leveraging their hardware capabilities. However, it requires substantial computing power, including at least one terabyte of RAM to function effectively. This makes the Mac Studio with the M5 Ultra chip the recommended platform to utilize the model’s full potential.

In their deployment examples, Meta integrated Muse Spark Open with Muse Code, their AI coding terminal, to perform agentic workflows. These included automatic code fixes, enhancements, and entire builds for software such as the Mastodon client. Additionally, the model demonstrated utility in video editing contexts by assisting with reviewing and cropping footage within DaVinci Resolve.

The open-weight nature of Muse Spark Open provides significant industry advantages over proprietary models. These include greater transparency, flexibility for customization, and potentially lower costs for deployment. Meta is currently making the model accessible through various API platforms, aiming to support developers and professionals pursuing advanced AI-driven tasks on Apple hardware.

This launch highlights a growing partnership between Meta and Apple, focusing on local AI capabilities without needing cloud dependency. While the memory demands limit its availability to high-end systems, Muse Spark Open could mark a transformative step in bringing powerful AI tools directly to personal desktop environments, particularly for coding and multimedia workflows.