Meta unveils Muse Wonder models in challenge to AI leaders

what happened

Meta has introduced two new artificial intelligence models, Muse Wonder and Muse Wonder Full. Cointelegraph reports that the company described them as its most capable models so far and as a step toward “personal superintelligence.” On Monday, Meta said the releases add two features called Active Memory and Active Focus. The first retains relevant information between conversations, while the second helps an agent continue working on demanding tasks.

Independent evaluation platform Artificial Analysis gave Wonder Full a score of 56 and the standard Wonder a score of 52 on its Intelligence Index, with both tested at maximum reasoning effort. Those results placed Full ahead of Claude Fable 5.1 and GPT-6 Astra, which each scored 53, but behind Claude Opus 5.5 at 58.

Bindu Reddy, CEO of Abacus.AI, which runs its own model evaluations, disputed Meta’s performance claims. In a post on X, she said Muse Wonder was “nowhere near Fable or Astra level” and added that detailed benchmark results would be released later.

Meta’s latest models are also its most expensive so far. The standard Wonder costs $4 per million input tokens, matching Claude Opus 5.5 but pricing below GPT-6 Astra and Claude Fable 5.1 at $10 and $50 respectively. Wonder Full is restricted to subscribers of Muse Code’s $50-per-month “power usage” and $200-per-month “maximum usage” plans.

why it matters

The release underscores intensifying competition among AI developers while showing that performance rankings remain contested. Pricing and limited access to Wonder Full are relevant to comparisons of the models’ commercial positioning.