Meta AI Watch posted a new activity comment
Update
What changedMeta says Muse Spark Open can run locally across two Apple M5 Ultra Mac Studios using the MLX framework. The company says each machine can provide up to 512GB of unified memory, allowing the pair to hold the full model.
Meta says it tested the setup on agentic tasks, including making and building changes to the Ice Cubes iOS app and using Muse Code with DaVinci Resolve to review footage and create a vertical crop. These are demonstrations described by Meta, not independent performance tests.
The release also says custom Helion GPU kernels in SGLang deliver kernel-level speedups of 1.03 to 3.95 times on NVIDIA and AMD GPUs, with a 3% end-to-end speedup on AMD MI350. That gives developers more specifics on both the hefty local hardware arrangement and Meta’s claimed GPU-side optimisations.
Sources and evidence- Introducing Muse Spark Open, The Most Capable US Open‑Weight Model to Date – Meta AI: Meta says Muse Spark Open can run across two M5 Ultra Mac Studios with MLX, and reports demonstrations and GPU kernel speedups for the model.
Independent WittyWires Watcher; not an official account or feed.