Muse Wonder 1.0
Language modelCard summarizedPlans long tasks, delegates to sub-agents, and recovers from errors.
Source: Muse Wonder Model Card (13 pp)
Use it via API
Where to call it
- Meta Model API
muse-wonder-1.0
Supports
- Image input
- PDF input
- Structured outputs
- Reasoning control
- Tool calling
- Web search
- Prompt caching
Good for
- General assistance and reasoning: questions, writing, summarizing, analyzing documents and data
- Coding and software engineering, from single tasks to sustained work across larger projects
- Personal and professional agents: research, planning and multi-step workflows through tools and apps
Overview
Muse Wonder
Plans long tasks, delegates to sub-agents, and recovers from errors.
- Cost
- $4.00 in
- Context
- 1M
- Input / output
- $4.00 / $20.00
- Max output
- —
- Knows up to
- —
- Released
- 12 Oct 2026
Meta's most capable model at launch, built on the work behind Muse Spark with larger base models, and aimed at long-horizon agents, coding and specialist knowledge work in law, finance and health. The model card covers Muse Wonder (on the API) and Muse Wonder Full (Muse Code only). Meta did not rule out the unmitigated model reaching its high-risk threshold for chemical and biological misuse, so it ships with extra safety training and system-level classifiers; with those in place Meta rates every domain "moderate or lower". p.1–2, p.7, p.11; research blog
muse-wonder-1.0- Input / output per 1M
- $4.00 / $20.00
- Context window
- 1M
- Max output
- —
- Knows up to
- —
- Reasoning
- Always on; reasoning_effort minimal to xhigh, plus max
- Weights
- Closed
About this card. A model card is the maker's own account of a model: what it's for, how it was trained and tested, and where it falls short. This page condenses Meta's Muse Wonder Model Card. Every claim carries its source page.