Meta has unveiled Muse Wonder and Muse Wonder Full, two frontier AI models built for personal and professional agentic tasks. Announced on the Meta AI Research blog dated October 12, 2026, the Muse Wonder series is described as Meta's most capable models yet for agents that can work across work and daily life.
Unlike standard chatbots, Muse Wonder is designed to keep track of what matters to you and work toward your goals over time. It combines two new systems, Active Memory and Active Focus, which Meta says are the core building blocks for reliable personal and professional AI agents.
Muse Wonder specs and key features
Meta positions Active Memory as a critical step toward useful agents. The model carries important context about you across conversations and actively retrieves the details relevant to the task at hand, while recognizing when earlier information becomes relevant again.
Active Focus handles sustained effort. According to Meta, Muse Wonder keeps working toward a challenging goal, tries alternative approaches, works through blockers, and adapts to new information. It also asks for clarification when needed and checks in before taking important actions.
The company evaluated both models across personal and professional agents, memory, and coding benchmarks, including Harvey's Legal Agent Benchmark for legal work products and SWE-Marathon for long-running software projects. Meta says the series achieves state-of-the-art results in these areas.
- Active Memory: carries important context about you across conversations and retrieves details relevant to the task at hand
- Active Focus: continues working toward challenging goals, tries alternative approaches, works through blockers, and adapts to new information
