We have heard a lot about open source Chinese models. Muse Glimmer gives you an American version. It is a 30B parameter model that can run locally. It is already available through Ollama’s MLX engine, so as long as you have a decent MacBook, you can use it locally. Muse Glimmer can power Claude Code, Codex, and other local agents. Optimizations for Apple Silicon, NVIDIA, AMD, and other platforms will be available soon.
Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally. Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases. Congrats…
— Mark Zuckerberg (@finkd) August 10, 2026
People are already doing amazing things with this model. For example, they can rely on it to create images and create interactive websites. Muse Spark 1.2 will have its weights released soon. It is already one of the smartest models that sits just behind Kimi K3.
thanks zuck new chapter added now in the journey created by muse 1.2pic.twitter.com/QSlzVGNRGg
— Chetaslua (@chetaslua) August 10, 2026
