Meta released frontier AI models Muse Wonder and Muse Wonder Full through its Superintelligence Labs and Muse services
Model release ● Confirmed 78% confidence first seen
Meta launched Muse Wonder and Muse Wonder Full as part of its Superintelligence Labs effort. The models were made available to Muse Code subscribers, with Muse Wonder also accessible via the Meta Model API, and Meta positioned them as high-performing models on AI benchmarks for tasks including coding and its Muse agent.
Decision brief
- What changed
- Meta’s Superintelligence Labs released two new frontier AI models, Muse Wonder and Muse Wonder Full. Muse Wonder is available through the Meta Model API and to Muse Code subscribers, while Muse Wonder Full is limited to Muse Code subscribers; reported API pricing for Muse Wonder is $4.00 per million input tokens and $20.00 per million output tokens.
- Why it matters
- This gives businesses a new Meta-supplied frontier model option to evaluate for coding, agent, and API-based workloads, with a clear split between broadly accessible access (Wonder) and subscriber-only access (Wonder Full). Decision-makers should care because Meta is explicitly positioning these models as the foundation for its Muse AI agent, which could affect vendor selection, developer tooling choices, and the economics of production AI deployments if the pricing and access model prove competitive for targeted use cases.
- Evidence
- The Verge and TechCrunch AI both report that Meta launched Muse Wonder and Muse Wonder Full through its Superintelligence Labs, with Wonder available via API and both available to Muse Code subscribers. TechCrunch AI additionally reports the API pricing and says the models are being positioned for coding and to power the Muse agent, broadly consistent with The Verge’s description of Meta’s rollout.
- What remains uncertain
- The coverage does not establish independent performance validation against competing frontier models for real enterprise workloads, and Wonder Full’s lack of API availability limits clarity on broader deployment options. It is also unverified from the provided coverage how durable the reported pricing will be, what enterprise terms or SLAs apply, and when or whether broader access to agent-related specialized models will arrive.
- Monitor next
- Watch for independent benchmark results and customer adoption signals for Muse Wonder, especially any Meta disclosures on enterprise availability, API expansion for Wonder Full, and Muse agent rollout beyond subscribers.
Analytical support, not advice — assumptions and open questions stated above.