Meta unveils Muse Wonder AI models, pitches step toward personal superintelligence
In brief
- Muse Wonder and Muse Wonder Full are Meta's most capable models to date, the company says.
- Active Memory and Active Focus are two features Meta says the models introduce.
- Artificial Analysis scored Wonder Full at 56 and standard Wonder at 52 on its Intelligence Index.
- Abacus.AI CEO Bindu Reddy disputed the claims on X and said detailed benchmarks would follow.
- Wonder Full is limited to Muse Code's $50 and $200 monthly subscription tiers.
What Meta says the models do
Meta's pitch comes down to two features. According to the company, Active Memory carries relevant information across conversations, and Active Focus helps the agent keep working on difficult tasks. In other words, it's an agent that's meant to remember context and stay on hard problems.
The framing is ambitious, and it comes from the top.
Meta co-founder Mark Zuckerberg said its newest models are a step toward “personal superintelligence.”
How the benchmarks stack up
Outside numbers are what most people will look at. Independent evaluation platform Artificial Analysis scored Wonder Full at 56 on its Intelligence Index and the standard Wonder at 52 (both at maximum reasoning effort). By Artificial Analysis's count, that puts Wonder Full above Claude Fable 5.1 and GPT-6 Astra, which each scored 53, but below Claude Opus 5.5, which scored 58.
Not everyone's convinced.
Bindu Reddy, CEO of Abacus.AI (which runs its own model evaluations), challenged the performance claims in a post on X, writing that Muse Wonder was "nowhere near Fable or Astra level." She said detailed benchmark results would follow. Until they do, her criticism isn't backed by published numbers, and the Artificial Analysis scores are the only third-party figures in the report.
Pricing and access
Cointelegraph described the new models as Meta's most expensive to date. The standard Wonder model charges $4 per million input tokens, according to the report.
Wonder Full is harder to get. It's available only to people who subscribe to Muse Code's $50-a-month "power usage" tier or its $200-a-month "maximum usage" tier, so Meta's top model sits behind a subscription rather than a per-token rate.
Frequently asked questions
What are Active Memory and Active Focus in Meta's Muse Wonder models?
They're the two capabilities Meta says its new Muse Wonder models introduce. According to Meta, Active Memory carries relevant information across conversations, and Active Focus helps the agent keep working on difficult tasks.
How did Muse Wonder score on the Artificial Analysis Intelligence Index?
Artificial Analysis scored Wonder Full at 56 and the standard Wonder at 52, both at maximum reasoning effort. That put Wonder Full above Claude Fable 5.1 and GPT-6 Astra (53 each) and below Claude Opus 5.5 (58).
Who can access Muse Wonder Full?
Wonder Full is available only to subscribers of Muse Code's $50-a-month "power usage" tier or its $200-a-month "maximum usage" tier. The standard Wonder model charges $4 per million input tokens, according to Cointelegraph.
