Muse Wonder and Muse Wonder Full
Meta's newest and most capable AI models, built for agents that work through long, multi-step tasks, available now to developers and Muse Code subscribers.
Details
- Muse Wonder is the general model, available through the Meta Model API and Muse Code. Muse Wonder Full is the top-end version and is only available inside Muse Code subscriptions
- Both have Active Memory, which keeps relevant details across conversations, and Active Focus, which keeps an agent working toward a goal, tries other approaches after errors, and checks in before important actions
- Both handle up to 1 million tokens of context and can reason over screenshots, charts, documents, images, and video
- Muse Wonder costs $4 per million input tokens and $20 per million output tokens on the API. Developers already using OpenAI's code libraries can switch by changing the address and model name (muse-wonder-1.0)
- Muse Wonder Full comes with the Power Usage plan and a new Maximum Usage plan, which gives 80 times the Everyday Usage limit
- Muse Wonder will soon power Meta's personal agent app, Muse, rolling out to subscribers first
Best for
Long-running agent work like coding projects, multi-step research, and tasks that involve operating a computer on screen.
When to use it
You build on Meta's models or already pay for Muse Code and want the strongest option Meta offers.
Who benefits
Developers and technical builders today, and Muse app users once it reaches subscribers there.
Join the newsletter to get ahead with AI.
I share strategies, new modules, free guides, and tools to get you better at using AI in your work.
Leading a team? See team programs →