MENLO PARK, CALIFORNIA / RankWire.AI / – Meta has unveiled Muse Wonder and Muse Wonder Full, a new lineup of sophisticated artificial intelligence models. These systems are designed for software development, AI agents, and tackling complex everyday challenges. They can assist with areas such as leases, insurance claims, financial data, and application building. Meta launched Muse Wonder via its Model API and Muse Code on October 12. Meanwhile, Muse Wonder Full became accessible to users with eligible Muse Code subscriptions.
The focus of these models is on multi-step tasks rather than simple responses. Muse Wonder can plan projects, utilize digital tools, and break larger tasks into smaller sub-agents. It is capable of working across browsers, desktop applications, and mobile platforms. Meta designed the system to recover when a chosen method fails. The models support context windows of approximately one million tokens, enabling them to process extensive documents, codebases, and other large materials within a single session.
Meta also introduced Active Memory and Active Focus alongside the Wonder models. Active Memory retains useful details from earlier conversations and retrieves them when needed. Active Focus allows an AI agent to persist through difficult problems and modify its approach after errors. The models can ask for clarification if instructions lack key details. They can also seek confirmation before executing sensitive or critical actions. These features aim to facilitate longer work sessions without users needing to restart tasks repeatedly.
New models boost Meta’s agent and coding functionalities
Muse Wonder is priced at $4 per million input tokens via the Meta Model API. Output costs are $20 per million tokens, and cached input is $0.20 per million tokens. Instead of separate API pricing, Meta offers Muse Wonder Full through paid Muse Code plans. The company provides a Power Usage plan at $50 per month and a Maximum Usage plan at $200 per month, aimed at developers requiring more extensive coding resources.
Meta shared benchmark results evaluating software engineering, coding agents, and long-term technical tasks. Tests included DeepSWE and SWE-Marathon. The company also assessed both models at maximum reasoning effort. Muse Wonder received an Intelligence Index score of 52, while Muse Wonder Full scored 56. These results surpass Meta’s earlier Muse Spark models on the evaluator’s scale and demonstrate a clear improvement in reasoning capabilities.
Safety and privacy measures accompany the rollout
Meta states that the Wonder models incorporate safeguards for privacy, harmful requests, and cybersecurity use. They are trained to avoid revealing private information without user consent. The models also decline requests involving harmful activities, while maintaining legitimate security functions. Developers can still utilize the models to identify vulnerabilities in their own software and enhance defenses. Meta assesses its most advanced AI systems under its Superintelligence Scaling Framework and plans to publish additional information on safety testing for the new models.
This launch follows a broad restructuring of Meta’s AI initiatives that began approximately 14 months earlier. The company explains that Muse Wonder builds upon the technical base established with Muse Spark. Development involved significant computing resources at Meta’s Prometheus data center. The company also intends to integrate Muse Wonder into more features within the Muse platform, starting with subscribers. Developers can already access Muse Wonder, while Muse Wonder Full remains available via qualifying Muse Code subscription plans.