Muse Wonder 1.0
Muse Wonder is Meta's frontier model for coding, long-running agents, and computer use, launched October 12, 2026.
Provider
Meta
Model family
Meta Muse
Frontier agentic and coding LLM
Cost tier
Wonder
Status
Current
Release Oct 12, 2026
Technical specs
- Inputs
- text, image
- Outputs
- text
- License
- Proprietary API
- API ID
muse-wonder-1.0- Context tokens
- 1,000,000
- Input Usd Per Million
- 4
- Output Usd Per Million
- 20
- Cache Read Usd Per Million
- 0.2
- Capabilities
- Frontier coding across large repositories
- Long-running agents with sub-agent delegation and parallel tool calls
- Computer use across browsers, desktop, and mobile apps
- Active memory across conversations
Why teams choose it
Complex reasoning
Useful for workflows that require structured thinking, multi-step logic, and deeper analysis than lightweight models provide.
Long-context analysis
Helps teams summarize, compare, and extract insights from long documents without losing important nuance.
Meta roadmap vigilance
Use published model pages—not stale marketing blurbs—for modalities, quotas, pricing, and policy; schedule revalidation tied to vendor release notes.
Cost-efficient routing
Useful as part of a routing stack where cheap models handle drafts and confirmations and this tier handles genuinely hard passages.
Tradeoffs to know
- Benchmark figures come from Meta's own comparison table; only some runs are third-party.
- Meta's model page does not list a max output token limit.
- Muse Wonder Full is a separate, subscription-only model and is not this API ID.
Benchmarks
- Deepswe 1 1 (vendor-reported)
- 78.8
- Osworld V2 1 (vendor-reported)
- 70.5
- Programbench (vendor-reported)
- 50
- Swe Marathon (vendor-reported)
- 53.1
- Terminal Bench 4 0 Aa (vendor-reported)
- 52
- Artificial Analysis Index (vendor-reported)
- 52
Meta Muse family lineup
Related models
Muse Wonder 1.0 FAQ
What is Muse Wonder 1.0?
Muse Wonder is Meta's frontier model for coding, long-running agents, and computer use, launched October 12, 2026. 00 output per million tokens. It is available in the Meta Model API and Muse Code. Meta says it plans long tasks, delegates parts to sub-agents, calls tools in parallel, recovers from errors, and can operate browsers, desktop, and mobile apps. 1, and 52 on the Artificial Analysis Index. A larger sibling, Muse Wonder Full, is available only with a Muse Code subscription.
When does Muse Wonder 1.0 fit best?
Multi-hour coding and agent tasks through the Meta Model API
What should teams watch out for with Muse Wonder 1.0?
Benchmark figures come from Meta's own comparison table; only some runs are third-party.
Explore next
Models, tools, and comparisons that connect to this reference.