Wired described Muse Wonder on 12 October 2026 as Meta’s biggest model and said it rivals the world’s best. The model also includes new features intended to keep AI agents from going off the rails, Wired reported.
Sorting a work inbox would test whether an agent could stay with the task it was given, if it were allowed to handle those messages. Features intended to keep an agent on task would matter in that setting because the agent’s actions, rather than just its answers, would need to remain within the job assigned to it.
Meta introduced a range of AI models it says could rival the work of OpenAI, Anthropic and Google at the frontier of the technology, Axios reported. The launch follows heavy spending by Meta to assemble an AI research team and the computing infrastructure needed to compete, according to Axios.
.jpg){kind=link}