Meta and Apple team up to make Muse AI models run better on Macs

Meta announced the Muse Spark Open AI model, positioning it as the most capable US open-weight model released so far. The headline claim is that it can run locally on a desktop setup consisting of two Apple M5 Ultra Mac Studios using the latest MLX framework, marking a notable collaboration between Meta and Apple on local AI inference.

The significance lies in the shift toward running large open-weight models entirely on consumer-grade hardware rather than relying on cloud GPUs. If a model of this class can operate on dual Mac Studios, it lowers the barrier for developers and enterprises that want on-device or on-premise AI without sending data to external servers.

The MLX framework, which is optimized for Apple silicon, plays a central role in this achievement. Its ability to efficiently handle large model inference on M-series chips has been growing, and this announcement suggests that the framework is now mature enough to support frontier-class open models.

For the open-source AI community, a capable US open-weight model that runs locally could accelerate adoption among researchers, startups, and privacy-conscious organizations. It also intensifies competition among open-weight releases, pushing other labs to optimize their models for local deployment.

The partnership angle between Meta and Apple is worth watching. While Apple has historically kept its AI efforts largely internal, collaborating with Meta on model optimization for Apple silicon could signal a broader trend of cross-company cooperation on local AI infrastructure.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.