Use Meta AI assistant to get things done, create AI-generated images for free, and get answers to any of your questions. Meta AI is built on Meta's latest Llama large language model and uses Emu.
This is the 4th launch from Meta AI. View more
Muse Wonder
Launching today
Muse Wonder is Meta’s new model series for personal and professional agents, with two models: Muse Wonder and Muse Wonder Full. Active Memory carries your context across conversations, Active Focus keeps working through blockers on hard goals, and it adds domain expertise in law, finance, and health plus stronger coding. Muse Wonder is live on the Meta Model API and Muse Code. Muse Wonder Full is for Muse Code subscribers.
Free Options
Launch tags:Artificial Intelligence
Launch Team
Viktor.comAn AI coworker that actually does the work
Promoted
Muse Wonder is Meta’s new model series for personal and professional agents, from Meta AI Research. It has two models, Muse Wonder and Muse Wonder Full. Meta calls them its most capable models, and the post says they were trained at its Prometheus data center.
Problem: Most AI assistants forget you between chats and give up or drift on long, hard tasks. The post frames personal agents as needing to remember context, stay on a goal and handle private information carefully.
Solution: Meta built the series around four things:
Active Memory: carries important context about you across conversations and pulls up the relevant details for the current task
Active Focus: keeps working toward a hard goal by trying other approaches and working through blockers, and asks for clarification or checks in before important actions
Domain expertise: specialized knowledge in law, finance and health, such as understanding a lease
Frontier coding: tracks requirements, edits codebases and iterates through testing cycles, trained on long-running tasks
What makes it different: the focus on long-running, personal work. Meta shows demos like a Saturday run plan built across a browser and other apps, custom swag ordered from Instagram photos, and a playable 3D kitchen game built from image inputs. You can try the Kitchen Game demo and the step-and-repeat demo.
How it performs: Meta says it reaches state-of-the-art results on benchmarks for personal and professional agents, memory and coding. The named evaluations are Harvey’s Legal Agent Benchmark, FocusBench and SWE-Marathon. These are Meta’s own claims, and the post’s text gives no scores, since the comparisons are in charts. The model report has the full evaluation details.
Availability:
Muse Wonder is available now through the Meta Model API and Muse Code
Muse Wonder Full is available now for Muse Code subscribers
The Muse app will soon be powered by Muse Wonder, rolling out first to subscribers
Safety: Meta says it has its strongest safeguards to date against chemical, biological, radiological, nuclear and cyber misuse, under its Superintelligence Scaling Framework.
Who it’s for: developers building agents through the API or Muse Code, and people who want a personal assistant that keeps context over time.
Read the full announcement.
P.S. I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends
Hey, congrats on the Muse Wonder launch! We're sharing a launch day with Benson: https://www.producthunt.com/products/benson-2. Would appreciate a comment or a little support if you have time today. Wishing you a great launch!