Meta rolls out Muse Wonder and Muse Wonder Full, introducing memory across chats and a looping problem-solving approach as it climbs higher in the AI race.
Meta has added two new AI models to its Muse lineup: Muse Wonder and Muse Wonder Full. They introduce cross-chat memory (Active Memory) and a persistence-driven approach (Active Focus) that keeps trying different methods until a task is done.
Safety is a key theme, with promises to protect private data and resist harmful requests. Muse Wonder Full shows strong performance in benchmarks, particularly in coding tasks, as Meta pushes Muse from consumer features toward enterprise use.
For everyday users, this could mean more personalized interactions and fewer repeated mistakes as the AI explores solutions. The rollout will take place over the coming weeks.
Read in: English Français Deutsch Español
Meta is expanding its Muse AI family with two new models, Muse Wonder and Muse Wonder Full, signaling a broader push into both everyday and enterprise AI use. The company describes two user-facing upgrades: Active Memory, which lets the model recall preferences across conversations, and Active Focus, a looping system that tests different strategies until the objective is achieved.
In a field crowded with rivals, Meta positions these models as a stronger step forward for its Muse ecosystem, spanning consumer features and business integrations. Early benchmarks show Muse Wonder Full performing well on several AI indices and performing especially strongly in coding tasks. Meta also stresses safety, saying Muse Wonder family won’t share private information without permission and will dial back requests that could enable harm, while tightening defenses against cyber threats.
For readers, the headline takeaway is that Meta is trying to move from a rear-guard position to a front-line participant in the AI race. Muse Wonder and Wonder Full are meant to feel more personal and capable, while staying mindful of privacy and safety in everyday use.
This is an incredible form by @AIatMeta pic.twitter.com/fTcUKkvnqu
— Digital Trends (@DigitalTrends) October 12, 2026
Meta unveiled two new members of the Muse family: Muse Wonder and Muse Wonder Full. The company says these models are already being rolled out to Muse users and will broaden Muse’s reach across consumer apps and enterprise tools, including hardware integrations such as smart glasses.
The company frames Wonder and Wonder Full as two tiers within the Muse lineup. Wonder Full is positioned as the higher-performing option in benchmarks and coding workflows, while both models share memory and persistence features designed to improve how they handle long-running tasks.
Active Memory lets the AI retain user preferences across sessions, which can make later interactions feel more tailored. Active Focus acts like a built-in retry loop, trying alternative methods when the initial attempt falls short, until the job is completed.
Meta says the Muse Wonder family was trained to avoid disclosing private information without permission, decline requests that could enable harm, and bolster defenses against cyber threats. That framework matters for users who rely on AI for planning, drafting, and research while wanting stronger privacy controls.
If you use Muse-powered features now or plan to, you could see more personalized responses across chats and a more proactive problem-solving approach. For example, if you frequently request emails in a certain tone or pull data from specific sources, Active Memory could streamline those tasks over time. Active Focus should cut down on repeated back-and-forth by exploring alternative paths to reach a solution.
Meta says the rollout is ongoing, with broader access likely in the coming weeks as services and integrations update. If you depend on Muse features for work or daily planning, review your privacy settings and data-handling preferences to understand how memory data is stored and used.
They are two new AI models in Meta’s Muse line, designed to be more capable with cross-chat memory and persistent problem solving. Wonder Full is the stronger performer in benchmarks.
Active Memory lets the AI remember user preferences across chats, while Active Focus keeps trying different approaches until a solution is found.
Meta says the models are already rolling out to Muse users, with broader access expected in the coming weeks.
You're reading the quick version.
Intel renames Crescent Island as the GPU D-Series and shares targets for FP8 compute and memory bandwidth geared toward AI inference in data centers.
OpenAI grows in Dublin with 200 new roles and the launch of a second applied AI deployment lab ahead of a major European AI summit.
Google describes how AI aids gene mapping, rapid diagnoses, engineered enzymes, and smarter disease forecasting.
The Daily Brief
Every morning, the news that matters from AI, phones, apps and the people shaping tech. Explained in plain English. Free.
One email a day. No spam, unsubscribe anytime. Privacy policy
