Meta Launches Muse Wonder AI Models, Claims Rivalry with OpenAI and Anthropic

Two new models join Meta's AI lineup

Meta released two artificial intelligence models called Muse Wonder and Muse Wonder Full on Monday, October 13, 2026. The company described them as its most capable models to date and a step toward what it called personal superintelligence. Meta co-founder Mark Zuckerberg was associated with that framing.

The models compete directly with offerings from OpenAI and Anthropic, the other two major players in the AI race.

Benchmark scores and price tags

• Independent evaluation platform Artificial Analysis scored Muse Wonder Full at 56 on its Intelligence Index and the standard Muse Wonder at 52, both measured at maximum reasoning effort.
• That places Muse Wonder Full above Claude Fable 5.1 and GPT-6 Astra, which each scored 53, but below Claude Opus 5.5 at 58.
• The standard Muse Wonder model costs $4 per million input tokens, matching Claude Opus 5.5 but undercutting GPT-6 Astra at $10 and Claude Fable 5.1 at $50.
• Muse Wonder Full is available only to subscribers of Muse Code's $50 per month power usage tier or the $200 per month maximum usage tier.

What Meta says the models can do differently

Meta highlighted two new features in these models: Active Memory, which carries relevant information across conversations, and Active Focus, which helps the AI agent keep working on difficult tasks without losing context.

Doubts from another evaluator

Bindu Reddy, CEO of Abacus.AI, which runs its own model evaluations, publicly challenged Meta's performance claims. Writing on X, she said Muse Wonder was nowhere near Fable or Astra level and indicated that detailed benchmark results from Abacus.AI would follow. Those follow-up results were not included in the source material available at time of writing.

Sources

• Cointelegraph