Meta launches first frontier models, Muse Wonder
Meta's Superintelligence Lab released Muse Wonder and Muse Wonder Full, its first frontier models, claiming benchmark scores comparable to Fable 5.1 and GPT-6 Astra. The models emphasize long-term memory through a property Meta calls active memory, outscoring rivals on the MemoryBench benchmark. Wonder is priced at $4 per million input tokens and $20 per million output tokens via the Meta Model API, while Wonder Full is exclusive to Muse Code subscribers. Both models will power the Muse agent and were trained at Meta's Prometheus data center in Ohio.
Meta says the big draw is memory: the model should remember what matters to you over months.
