Meta Launches Muse Wonder and Muse Wonder Full as Its Most Capable AI Models Yet

Meta introduced Muse Wonder and Muse Wonder Full on 12 October 2026, calling them its most capable models to date. They build on the groundwork of Muse Spark and are aimed at personal and professional agentic tasks, from working through an insurance claim to understanding a lease to building a working app.

Access is split. Muse Wonder is available now to developers through the Meta Model API and Muse Code. Muse Wonder Full is available to Muse Code subscribers. Meta says Muse Wonder will soon power experiences in the Muse app, rolling out first to subscribers.

Two features lead the announcement. Active Memory lets a model carry what matters across conversations and bring back relevant details for the task in hand. Active Focus is Meta's name for staying on a goal when the work gets hard: trying other approaches, recovering from errors, asking for clarification when needed, and checking in before doing anything important.

Meta also pitches the models at expert domains including law, finance and health, alongside frontier coding performance. On safety, it says the models were trained to avoid sharing private information without permission, to refuse requests to help with harmful acts, and to strengthen defences against cyber attacks while still being able to help developers find flaws in their own code. A full report with additional safety evaluations is promised soon, so the claims are not yet independently checkable.

Why it matters for studios like ours: the interesting line is the one about checking in before consequential actions. As agents move from writing text to touching real files, accounts and websites, the behaviour that matters is when a model stops and asks. We would rather see that tested in public than described in a launch post, so we will be watching for the safety report and for independent benchmarks.

For now, treat the performance language as Meta's own. The practical step is to try Muse Wonder on a real task from your own work and see where it asks questions and where it charges ahead.