GoKawiilApple's MLX update lets Muse Spark Open's ~1 trillion-parameter model run on Mac Studio clusters
Apple released a new version of its MLX framework that can run the frontier-class Muse Spark Open model, which has nearly a trillion parameters, locally on Mac Studio hardware. Apple reports that a cluster of four Mac Studios with M5 Ultra chips achieves over 36 tokens per second using tensor parallelization, and that single machines with up to 512GB of unified memory can also run the model, scaling to multiple units via an RDMA over Thunderbolt driver.
GoKawiil's interpretation of the reporting above, not reported fact.
The release suggests Apple is positioning its Mac Studio hardware, built on unified memory architecture, as a viable local alternative to cloud-based infrastructure for running very large AI models. This could appeal to researchers and developers who want privacy and local control when experimenting with frontier-scale models, though the benchmarks are self-reported by Apple and real-world performance may vary by workload.
- Apple's updated MLX framework can run the ~816B-parameter Muse Spark Open model locally on Mac Studio computers
- A four-Mac-Studio cluster with M5 Ultra chips reportedly generates over 36 tokens per second via tensor parallelization
- Multiple Mac Studios can be linked using an RDMA over Thunderbolt driver, with single units supporting up to 512GB of unified memory
Apple Mac Studio (M5 Ultra) — This is literally the machine the article describes running the trillion-parameter Muse Spark Open model locally via MLX. Its unified memory architecture (up to 512GB) is exactly what enables these frontier-class models to run and even scale across clusters with Thunderbolt RDMA. If you want to experiment with cutting-edge local AI like this, a Mac Studio is the hardware foundation.
See Apple Mac Studio (M5 Ultra) on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: machinelearning.apple.com, 2026-10-16
Published there as: “Apple ML Research: Muse Spark Open with MLX on Mac Studio”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.