Meta releases Muse Wonder agent models for long-running tasks
Meta has opened Muse Wonder and Muse Wonder Full to developers, adding cross-conversation memory and sustained focus for difficult agentic work.
AT A GLANCE
Key points
- Muse Wonder is available through Meta Model API and Muse Code
- Muse Wonder Full requires a Muse Code subscription
- Active Memory carries context across conversations
- Meta says an additional safety report will follow
What is changing?
Meta announced Muse Wonder and Muse Wonder Full on October 12 for personal and professional agentic tasks. Muse Wonder is available through the Meta Model API and Muse Code, while Muse Wonder Full is available to Muse Code subscribers. The models are due to reach Muse experiences later, starting with subscribers.
Meta’s Active Memory is designed to carry important context across conversations and retrieve it when relevant. Active Focus aims to help the model try alternative approaches, recover from errors, and check with the user before important actions. Meta positions the models for legal, financial, health, and long-running coding work.
Meta published benchmark results from its own evaluation suite and said an additional safety report will follow. Performance and safety claims should therefore be verified independently in production conditions.
Explore the original sources
Source date: 12 October 2026.
Published by AI Radar: 13 Oct 2026, 11:10 (Istanbul, UTC+3).
FOLLOW THE TOPIC
Topics and tags
Share your discovery.
Send this story to a friend.
Reader comments
What do you think about this development? Share your experience or question. Comments are published after editorial approval.
Your name/nickname and comment become public if approved. Read: Privacy notice.
Loading comments…