On October 12, Meta's Superintelligence Lab released two models, Muse Wonder and Muse Wonder Full; both will be used to power the Muse agent, and Meta put particular emphasis on what it calls “active memory”: retaining and synthesizing information you give it. TechCrunch describes it as Meta's first attempt at a frontier model; Wonder is offered through Meta's model API ($4 per million input tokens, $20 per million output tokens), while Wonder Full is only for Muse Code subscribers and will not be served through the API. The report does not say when the new models will reach Muse, or whether free users will get them.
OriginalTechCrunch ↗
