AIToday
Large Language ModelsAI Coding AssistantsOpen-Source AITechCrunch AIPublished: Oct 16, 2026, 06:00 JST

Meta launches Muse Spark Open, an open-weight AI model built for Macs

Meta launches Muse Spark Open, an open-weight AI model built for Macs

3 Key Points

  1. What happened

    Meta launched Muse Spark Open, an open-weight model with 816 billion parameters and a 1 million token context window, claiming superiority over other open-weight models based on its own evaluation results.

  2. Why it matters

    The model runs locally on Apple computers but needs at least 1 terabyte of RAM, positioning it as a tool for heavy coding and desktop workflows rather than everyday laptops.

WHO IT HITSDevelopers and enterprises that want to run large AI models on their own hardware, particularly Mac-based coding and video editing teams, are the clearest audience. The 1 terabyte RAM requirement means mostly high-end Mac Studio buyers can use it locally.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

Meta is framing Muse Spark Open as the most capable open-weight model in the United States, a claim backed by evaluation results the company released on Thursday. Open-weight models let others inspect and modify the underlying model, a contrast with closed and proprietary systems that Meta says can be less cost-efficient and less transparent.

The launch is notable for where the model runs. Muse Spark Open was designed to work with Macs and run locally on Apple's computers, though the memory requirement of at least 1 terabyte of RAM puts local use out of reach for typical machines. Meta points to the Mac Studio with the M5 Ultra chip as the best fit, which suggests the target is professional desktop work rather than casual experimentation.

Meta says it already tested the model in practical settings. It paired Muse Spark Open with Muse Code, its AI coding terminal, and used it for coding fixes on a Mastodon client that included bug fixes, new features, and a full build. It also deployed the model to run DaVinci Resolve, the high-end video editing software, where it reviewed and cropped footage.

FAQ
What hardware do I need to run Muse Spark Open locally?
At least 1 terabyte of RAM is required, and Meta says it runs best on a Mac Studio equipped with the M5 Ultra chip.
How is Muse Spark Open being distributed?
Meta is making the model available now through a number of different API platforms.
What tasks is Muse Spark Open built for?
It is natively multimodal and built for complex coding tasks and agentic labor, and Meta paired it with Muse Code, its AI coding terminal.

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articlePlan-then-Retrieve beats Retrieve-then-Plan on multi-tool queries