Meta releases its first frontier model, Muse Wonder

Meta’s Superintelligence Lab released two models on Monday: Muse Wonder and Muse Wonder Full. The company describes them as its first attempt at a frontier model. Unlike the lightweight Muse Spark, designed for low latency, the Wonder models are larger and are said to be more capable, with higher benchmark scores and better results on complex tasks such as coding.

In Meta’s testing, the models achieved benchmark scores comparable to Fable 5.1 and GPT-6 Astra. Muse Wonder costs $4.00 per million input tokens and $20.00 per million output tokens through the Meta Model API. Muse Wonder Full is available only to Muse Code subscribers and is not offered through the API. Both models will also power the Muse agent.

Meta says the models can retain and synthesize user-provided information, a capability it calls “active memory.” According to the company’s tests, Muse Wonder scores higher on the MemoryBench benchmark than Fable, Opus, or Astra. Research scientist Kavya Kopparapu said in a launch video that the team worked to help the model remember what matters to users over days, weeks, months, and beyond.

The models were trained at Meta’s Prometheus data center in central Ohio, which came online earlier this year. Their release is Meta’s first frontier-model launch since Alexandr Wang restructured its AI efforts in August of last year.