Apple ML Research: Muse Spark Open with MLX on Mac Studio
Source
machinelearning.apple.com
Author
antipaul
Date
Why it matters
A near-trillion-parameter open weightsA model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.Full definition → model now runs at reading speed on a four-Mac cluster using tensor parallelism over RDMA. That makes private local inferenceRunning a trained model to get answers — the phase where AI is actually used, as opposed to trained.Full definition → of frontier-class models practical.
Terms in this piece · Glossary
mixture-of-experts — A model built from many specialist sub-networks where only a few activate per token, giving big-model capability at small-model running cost.
open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
