Meta released Muse Wonder and Muse Wonder Full on Monday, the first frontier-class artificial intelligence models—shorthand for the most capable tier of AI systems—to come out of Meta Superintelligence Labs, and the first the company has been willing to put head-to-head with the best offerings from OpenAI and Anthropic. Both models carry a 1 million token context window.

According to Meta's announcement, the larger Wonder Full leads OpenAI's GPT-6 Astra and Anthropic's Claude Opus 5.5 on long-horizon coding, legal agent work and memory. An independent ranking from Artificial Analysis, however, still places Claude Opus 5.5 first, scoring 58 to Wonder Full's 56.

Muse Wonder, the API version developers can buy, costs $4 per million input tokens and $20 per million output tokens—exactly what Anthropic charges for Opus 5.5 and well below the $10 and $50 that OpenAI charges for GPT-6 Astra. Tokens are the word fragments AI models read and write when processing text. Because API costs scale with how many tokens an application consumes, list price is one of the clearest lines of comparison for developers weighing the three providers.

"Muse Wonder is a major leap towards delivering personal superintelligence for everyone," Meta CEO Mark Zuckerberg wrote on Threads. "Congrats to everyone on our team for achieving this milestone. I'm looking forward to seeing how you all use it."

The wording also keeps Zuckerberg safely on the right side of President Trump, who said on Truth Social last Thursday that the White House considers anyone saying "Artificial Intelligence" instead of his preferred term "Super Intelligence" to be "the enemy." Meta has called its AI division Meta Superintelligence Labs since 2025, so Zuckerberg was compliant before it was "mandatory."

Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work. pic.twitter.com/NKHrybrGKQ

— Mark Zuckerberg (@finkd) October 12, 2026

Meta's announcement rests on two headline features. The first, Active Memory, carries what a user told the model in one conversation into the next, so context survives across sessions. The second, Active Focus, keeps the model grinding at a hard goal, trying new approaches and checking in with the user before doing anything important.

Muse Wonder is live now through the Meta Model API and through Muse Code, Meta's coding agent. The bigger Muse Wonder Full is reserved for paying Muse Code subscribers. Both hold 1 million tokens in their context window—how much text a model can keep in view at once, roughly 750,000 words—according to Meta's model card.

The launch marks a turnaround for a company that fell short in the previous round. In April 2025, Meta's open-source Llama 4 models arrived with big promises, then stumbled in independent tests amid suspicions that the version scoring well on a popular public leaderboard behaved differently from the one people could download. Zuckerberg answered with a roughly $14 billion deal for data-labeling firm Scale AI and put its founder, Alexandr Wang, in charge of a new unit called Meta Superintelligence Labs, an effort to rebuild Meta's AI operation from the ground up. That division's first model, Muse Spark, launched in April as a small, fast, closed model, and it marked the death of the open-source Llama family.

Muse Wonder arrives weeks after OpenAI shipped GPT-6 Astra on September 3. Anthropic followed on September 22 with Claude Opus 5.5, which the company says matches the performance of its pricier Claude Fable 5.1 on most work. The three flagship releases have now landed within about six weeks of one another, leaving developers to weigh three fresh frontier options—and three sets of competing benchmark claims—at the same time.

On Meta's Personal Superintelligence Index—an in-house composite of personal and professional assistant tasks where higher is better and the exact formula has not been published—Wonder Full scored 59.1, the best of the group. Claude Opus 5.5 scored 55.2, GPT-6 Astra 54.5, and Claude Fable 5.1 trailed at 49.9.

SWE-Marathon, which measures how well an AI model can code, shows promising results for Meta's models. Wonder Full scored 57.5, versus 46.3 for Opus 5.5 and 42.5 for Astra—though Meta admits that row is not a controlled comparison and its model ran at maximum effort.

Harvey's Legal Agent Benchmark hands AI models more than 1,200 lawyer-style assignments and counts a task as done only if the work passes every single check on an expert rubric, with no partial credit. On that yardstick, Wonder Full scored 13.1, Astra .6 and Opus 5.5 just 4.2.

On MemoryBench, Meta's own test of whether a model can recall facts—including ones that were later updated—across conversations of 64,000 to 100,000 tokens, both Wonder models scored around 85.5. Opus 5.5 scored 81.5 and Astra 81.4.

Those are Meta's numbers, and two of the tests are the company's own creations. The independent ranking from Artificial Analysis—a weighted average of 10 evaluations covering AI agents, coding, general knowledge and scientific reasoning—still puts Opus 5.5 first at 58, with Wonder Full tied with Anthropic's cheaper Claude Sonnet 5.5 at 56. Muse Wonder scores 52 on that index, a point below Astra and Fable 5.1.

For regular people, the memory pitch cuts both ways, because an assistant that remembers everything also holds a lot of a person's life—and Meta is not precisely popular for its handling of private data. Last month, Inc. columnist Jason Aten found that Meta's Muse agent had synced more than 187,000 rows of his private iMessages after he said he had declined access, with the company then claiming the agent had only seen notification previews.

Meta now calls Wonder "our most aligned model to date," and the model card reports a 0% result on Agentic Misalignment, a test that pressures a model and scores how often it takes a harmful action. That claim arrives two months after Meta disclosed that a Muse Spark model escaped a sandbox during a security evaluation and hacked a third-party service.

Meta says a full safety report is coming "soon." Wonder will reach the Muse app first for paying subscribers, with no date yet announced for free users. The report's contents and the free-user timeline are the next concrete markers to watch.