MENLO PARK, CALIFORNIA / RankWire.AI / – Meta has unveiled the Muse Wonder and Muse Wonder Full models, a new set of sophisticated artificial intelligence tools. These systems are designed for tasks involving software creation, AI agents, and complex daily activities. They can support functions such as managing leases, processing insurance claims, handling financial data, and developing applications. Meta announced the release of Muse Wonder via its Model API and Muse Code on October 12, with Muse Wonder Full becoming accessible to users holding eligible Muse Code subscriptions.
These models are tailored for multi-step processes rather than simple, single-step answers. Muse Wonder is capable of planning tasks, utilizing digital tools, and breaking down larger projects into smaller, specialized sub-agents. It can operate seamlessly across web browsers, desktop applications, and mobile platforms. Designed with resilience in mind, the system can recover if a chosen approach does not succeed. It supports context windows of approximately one million tokens, enabling it to analyze large documents, extensive codebases, and other substantial materials within a single session.
Meta has also introduced features called Active Memory and Active Focus alongside the Wonder models. Active Memory retains relevant information from previous interactions and retrieves it when necessary. Active Focus assists an AI agent in persisting through challenging problems and adjusting its strategy after encountering errors. The models can ask for clarification when instructions lack critical details and seek confirmation before executing sensitive or significant actions. These features are aimed at facilitating longer work periods without requiring users to restart their tasks repeatedly.
Enhanced agent and coding skills mark the new models’ capabilities
Muse Wonder is priced at $4 per million input tokens through the Meta Model API. Output generation costs $20 per million tokens, while cached input is billed at $0.20 per million tokens. Instead of separate API charges, Meta offers Muse Wonder Full through paid Muse Code subscriptions. The company also provides a Power Usage plan at $50 per month and a Maximum Usage plan at $200 per month, targeted at developers requiring more extensive access for intensive coding tasks.
Meta has shared benchmark results in areas such as software engineering, coding agents, and long-term technical tasks. The tests included evaluations like DeepSWE and SWE-Marathon. The AI models were also tested at their highest reasoning capacity, receiving an Intelligence Index score of 52 for Muse Wonder and 56 for Muse Wonder Full. These scores surpass those of Meta’s earlier Muse Spark models, indicating a clear improvement in reasoning abilities as measured by the evaluation metrics.
Safety and privacy measures accompany the new model launch
Meta states that the Wonder models incorporate protections for user privacy, cybersecurity, and the prevention of harmful requests. They are trained to avoid revealing private information without explicit permission and are programmed to reject requests that involve malicious activities, while still enabling legitimate security testing by developers. Meta also continues to assess its most advanced AI systems under its Superintelligence Scaling Framework. Further details on safety testing for these models are expected to be published soon.
This release is part of a broader restructuring of Meta’s AI efforts that began approximately 14 months ago. The company emphasizes that Muse Wonder builds upon the foundational technology developed through Muse Spark. Development was supported by extensive computing resources at Meta’s Prometheus data center. Moving forward, Meta plans to integrate Muse Wonder’s technology into additional features within the Muse product suite, initially accessible to subscribers. Currently, developers can access Muse Wonder, while Muse Wonder Full remains available through qualifying Muse Code subscription plans.