Meta released Muse Wonder and Muse Wonder Full, its first frontier models, which its own testing puts roughly even with Anthropic’s Claude Fable 5.1 and OpenAI’s GPT-6 Astra on benchmarks. Wonder costs $4 per million input tokens and $20 per million output through the Meta Model API, close to what GPT-6 Sol and Claude Opus 5.5 charge, but Wonder Full is open only to people paying for Muse Code. Meta built both models to power its Muse agent and to remember what users tell them over weeks and months, and says Wonder outscores Fable, Opus, and Astra on the MemoryBench memory test. Outside of Meta’s own numbers, Vals ranked Wonder first on cybersecurity tasks and Artificial Analysis called it a big step up from Meta’s earlier models.
