Meta releases Muse Wonder AI models for complex tasks and agents

Meta Platforms released two new artificial intelligence models on October 12, named Muse Wonder and Muse Wonder Full. Bloomberg, SiliconANGLE, Digital Trends, and TechCrunch all reported on the release, which marks the debut of the new systems. According to SiliconANGLE, Meta built the models to run AI agents that handle longer, multistep tasks in personal and professional settings. [1][2][3][4][8][11]

The company stated that the new models are built to perform well on longer jobs, such as working through an insurance claim or writing computer code, SiliconANGLE reported. TechCrunch noted that the Superintelligence Lab at Meta released the pair as larger, more capable successor models to Muse Spark. Engadget reported that the models include new active memory and active focus abilities, while Wired reported that the software includes features designed to keep agents from going off the rails. [2][4][6][7]

According to CNBC Tech, users must pay for subscriptions to the Muse application to access the new Muse Wonder model. Digital Trends reported that the release puts Meta into direct competition with other frontier models, including Claude Opus 5.5 from Anthropic, Gemini 4 Argon from Google, and GPT-6 Astra from OpenAI. [3][5][11]