Meta Launches Muse Spark Open, an 816B-Parameter Open-Weight Model
Muse Spark Open pairs a 1 million token context window with local Mac support, but Meta says it needs at least 1 terabyte of RAM.
Meta launched Muse Spark Open on Thursday, describing it as the most capable open-weight model in the United States. The model uses 816 billion parameters and a 1 million token context window, and is available now through several API platforms.
Meta released evaluation results on Thursday claiming Muse Spark Open outperforms other popular open-weight models. According to Meta, the model is natively multimodal and built for complex coding tasks and agentic work.
Muse Spark Open was designed to run locally on Apple computers, and Meta says it requires at least 1 terabyte of RAM, a figure reported by Engadget. Meta said the model runs best on a Mac Studio with the M5 Ultra chip.
Meta paired the model with Muse Code, its AI coding terminal, and deployed it to agentic workflows, including coding fixes for a Mastodon client. Meta also deployed Muse Spark Open to run DaVinci Resolve, where it was capable of reviewing and cropping footage.
Quick answers
How many parameters does Muse Spark Open have?
Muse Spark Open uses 816 billion parameters and a 1 million token context window.
What hardware does Muse Spark Open need to run locally?
Meta says the model requires at least 1 terabyte of RAM and runs best on a Mac Studio with the M5 Ultra chip.
How can I access Muse Spark Open?
Meta is making the model available now through several API platforms.