MENLO PARK, CALIFORNIA / RankWire.AI / – Meta has launched its latest advanced AI models, Muse Wonder and Muse Wonder Full. These highly capable systems are tailored for programming, extended AI agents, and handling intricate personal or professional responsibilities. According to Meta, these models assist users in managing insurance claims, deciphering lease agreements, analyzing financial data, and creating functional applications. Muse Wonder was made available on October 12 via the Meta Model API and Muse Code, with Muse Wonder Full debuting exclusively for Muse Code subscribers.
This launch marks Meta’s strategic move into AI systems capable of executing multi-step tasks, moving beyond simple prompt responses. Muse Wonder can strategize extended projects, assign subtasks to sub-agents, utilize software tools, and recover from errors during operation. The models are optimized for use across browsers, desktop applications, and mobile platforms. Both models feature a context window of approximately one million tokens, enabling them to process extensive data within a single session.
Meta introduced two new features called Active Memory and Active Focus to enhance their ability to handle longer, more tailored tasks. Active Memory maintains relevant information throughout conversations and retrieves details when needed for new tasks. Active Focus allows the models to persist through challenging parts of a task, test alternative strategies, and recover from mistakes. Additionally, Meta states that the models can solicit clarification from users and verify actions before executing significant operations.
Models designed for coding and persistent AI agents
Through the Meta Model API, Muse Wonder is priced at $4 per million input tokens and $20 per million output tokens. Cached input data costs $0.20 per million tokens, as per Meta’s developer resources. The full version, Muse Wonder Full, is accessible only through paid Muse Code subscriptions, not separate API fees. Developers with intensive coding needs can subscribe to Power Usage at $50 per month and Maximum Usage at $200 per month.
Meta reports that Muse Wonder Full surpasses several benchmarks related to agentic and coding performance, with results from tests like DeepSWE and SWE-Marathon used to evaluate software engineering and long-term coding capabilities. An independent evaluator, Artificial Analysis, assigned Muse Wonder an Intelligence Index score of 52, while Muse Wonder Full achieved a score of 56 at maximum reasoning effort. These scores place the models ahead of Meta’s earlier Muse Spark generation on the evaluator’s index.
Enhanced safety features accompany expanded model functions
Meta has trained the Muse Wonder series to prevent sharing private information without consent and to decline harmful requests. The models also retain cybersecurity functions to help developers identify vulnerabilities in their code and bolster security systems. Meta states that its most advanced models are evaluated using the Meta Superintelligence Scaling Framework, with plans to publish a comprehensive report on additional safety assessments following the initial launch.
The company initiated a broad overhaul of its AI efforts approximately 14 months prior to the Muse Wonder release. Meta explains that these new models build upon the foundations established with Muse Spark, with training conducted at its Prometheus data center. The company intends to integrate Muse Wonder into more features within its Muse ecosystem, beginning with subscribers. The October launch provides immediate access for developers to Muse Wonder, while reserving the larger Muse Wonder Full model for Muse Code members.