Meta releases Muse Wonder AI models built to run personal and professional agents
Meta ships agent-native models. Muse Wonder and Muse Wonder Full are built for multistep autonomous work—the category everyone is racing to own.
Why it matters
Meta enters the agent-capability race with purpose-built models for personal and professional agents. The release signals competitive positioning in agentic AI, but specifics on benchmarks, guardrails, and deployment readiness are sparse in this announcement.
The key facts
6 to knowTwo models: Muse Wonder and Muse Wonder Full
Designed for longer, multistep agentic tasks
Meta positions as 'most capable yet' for agent workloads
Personal and professional agent use cases cited
Announced October 12, 2026
Benchmarks, pricing, availability, and safety/guardrail specifics not disclosed in article excerpt
The story so far
Earlier coverage of this storyline
- Meta Muse gives AMD a boost as AI momentum shifts to personal agentsCNBC Technology
- Muse Wonder Puts Meta Back at the AI FrontierWired AI
- Meta debuts Muse Wonder AI model in latest test of users' willingness to payCNBC Technology
- Meta releases its first frontier model, Muse WonderTechCrunch AI
- Meta's new frontier model Muse Wonder Full breaks into the AI top tierThe Decoder
- This story
Go to the source
SiliconAnglesiliconangle.com
Publisher excerpt: Meta Platforms Inc. today released Muse Wonder and Muse Wonder Full, two artificial intelligence models it says are its most capable yet at running agents that handle personal and professional tasks. Meta aimed the release at the longer, multistep jobs people hand off to an AI agent. One example it…