Meta Releases Muse Spark Open: 816B MoE Model Enables Local Execution on Mac Studio with 42B Speed
Meta released the weights of Muse Spark Open on Hugging Face, an 816B-parameter MoE model optimized for coding and agentic tasks. Featuring a sparse architecture with 256 experts per layer (activating 8), it delivers inference speeds comparable to a 42B model while requiring significant memory, enabling local execution on Mac Studios via collaboration with Apple.
Full Report
Rohan Paul (@rohanpaul_ai) · 2026-10-15 20
Meta released the weights of Muse Spark Open, on Huggingface, an 816B-parameter model built for coding and agentic work.
The open version shares the same base model as Muse Spark 1.3, with additional safety post-training to resist misuse as a standalone model without system guardrails, reversing Meta’s first proprietary, closed-weight frontier model strategy from April.
Each of its 62 layers holds 256 expert networks and routes every token through 8, so deployment needs the memory of an 816B-parameter model but runs at the speed of a 42B model.
That storage bill explains the Apple tie-up, because the M5 Ultra pushes 1.2 TB/s of unified memory and a Mac Studio can carry up to 512GB.
A 3-bit quant tested on a 512GB M3 Ultra took about 344 GB and ran at about 21 tok/s generation on a short prompt.
↳ 引用 @finkd (Mark Zuckerberg): Muse Spark Open -- weights releasing today -- runs great on new Mac Studios thanks to work with @johnternus and Apple.
