Meta released its Muse Wonder and Muse Wonder Full AI models on Monday, marking the company's first foray into frontier models, according to its announcement. The models are designed for personal and professional agentic tasks.

The release signals Meta's renewed push to compete with leading AI models from OpenAI and Anthropic. Muse Wonder is available through the Meta Model API and Muse Code, while Muse Wonder Full is restricted to Muse Code subscribers, the company said.

Meta claims that Muse Wonder and Muse Wonder Full achieve state-of-the-art results across key benchmarks for personal and professional agents, memory, and coding. The models were tested internally, with results showing comparable performance to Fable 5.1 and GPT-6 Astra, according to Meta's testing presented in the model card. The models also scored higher on the MemoryBench benchmark, which evaluates continual learning and memory construction frameworks; Meta highlighted that Muse Wonder outperformed Fable, Opus, and Astra in this area, according to the company's testing presented in the model card.

The announcement includes no independent test results or peer review.

Muse Wonder is priced at $4 per million input tokens and $20 per million output tokens, making it roughly equivalent in cost to GPT-6 Sol or Opus 5.5, according to TechCrunch's reporting. Muse Wonder Full is only available to Muse Code subscribers and is not served through the API.

Meta emphasized the models' safety and alignment features, stating that they were trained to lead on both dimensions with robustness to adversarial instructions and careful handling of private information. The company also mentioned that Muse Wonder offers its strongest safeguards to date against chemical, biological, radiological, and nuclear (CBRN) and cyber misuse. The models were trained at Meta's Prometheus data center in central Ohio, which came online earlier this year, according to TechCrunch's reporting.

Meta's release of the Muse Wonder models follows a period of restructuring and investment in AI efforts, including the launch of the lightweight Muse Spark model and the Muse agent. The company has been steadily expanding its AI lineup, with a focus on bringing useful AI agents to the masses through Muse, according to The Verge's reporting.

A prior Automatica article in July noted Meta's plans to expand its AI agent lineup, which this release builds upon.