Meta has introduced two new artificial intelligence models, Muse Wonder and Muse Wonder Full. The company describes these as its most capable models to date and a significant step toward developing personal superintelligence.
The models feature two new capabilities: Active Memory, which retains relevant information across different conversations, and Active Focus, designed to assist agents in maintaining progress on complex tasks.
According to the evaluation platform Artificial Analysis, Muse Wonder Full achieved a score of 56 on its Intelligence Index, while the standard Muse Wonder scored 52, both tested at maximum reasoning effort.
These scores place Muse Wonder Full above Claude Fable 5.1 and GPT-6 Astra, which both scored 53. However, the model remains below Claude Opus 5.5, which holds a score of 58.
Bindu Reddy, CEO of Abacus.AI, has publicly questioned Meta's performance claims. Reddy stated on X that the models do not reach the level of Fable or Astra and promised to release further benchmark results.
