Meta launches Muse Wonder AI model to challenge OpenAI and Anthropic
The model will power the subscription version of Meta's Muse AI assistant and is available to developers through Meta Model API and Muse Code
Published
Chart: META, one-minute prices, three sessions
Meta Platforms has launched Muse Wonder, the latest model from its AI unit, which the company says matches OpenAI and Anthropic on coding, financial and legal tasks. The model will power the subscription version of Meta's Muse assistant and is available to developers through Meta Model API and Muse Code.
Muse Wonder follows Muse Spark, which debuted in April 2026, and Muse Image and Muse Video, added in July 2026, according to Crypto Briefing. The site described the launch as the first major release from Meta's Superintelligence Labs since the unit was overhauled in 2025. Both Muse Wonder models carry a 1-million-token context window and are priced at $4.00 per million input tokens and $20.00 per million output tokens, positioning them against Anthropic's Claude Opus 5.5, according to OrcaRouter.
Meta shares are up 0.15% in afternoon trading. Zuckerberg told the Financial Times, in comments summarized by Axios, that Meta has "reached the AI frontier," and rejected calls for a coordinated industry slowdown on AI safety.
Muse Wonder puts Meta directly in the frontier-model race that OpenAI and Anthropic have so far dominated, and its claimed parity on coding, financial and legal tasks lands just as Anthropic pushes toward a pre-Thanksgiving IPO built partly on enterprise demand for those same capabilities. The release also builds on the reaction already seen around Meta's Muse platform, which lifted Taiwanese hardware names and pulled Shopify higher on its partnership announcement. A credible frontier entrant from Meta raises the bar Anthropic's IPO pitch has to clear on product differentiation, even if it leaves untouched the legal and supply-chain questions still sitting in Anthropic's prospectus.