Meta Launches Muse Wonder Frontier Model; Wang: Best Model for Personal and Professional Agents
On October 12, Meta released Muse Wonder and Muse Wonder Full, the first closed frontier models from Meta Superintelligence Labs, led by chief SI officer Alexandr Wang. In a post on X, CEO Mark Zuckerberg called it a major step toward building personal superintelligence for everyone.
Meta says Wonder supports a context window of about one million tokens and multimodal input, headlined by two features: Active Memory, which remembers across conversations, and Active Focus, which keeps long tasks moving. Wonder is available through the Meta Model API and Muse Code at $4 per million input tokens and $20 per million output tokens; the more capable Wonder Full is limited to paid Muse Code subscription tiers. Independent evaluator Artificial Analysis scores Wonder Full at 56 and Wonder at 52 on its intelligence index, just below Claude Opus 5.5 at 58.
"We think it is competitive as a frontier model — specifically the very best model for personal and professional agents," Wang said at launch. The models will power the Muse app, with subscribers first. Meta shares rose about 3% on launch day. (GSM, based on The Wall Street Journal and other reports)
Related:
Alexandr Wang on Building Muse: From a Two-Person Experiment to the Top of the U.S. App Store
Published by GSM · USA (original Chinese edition)
Photo: 图：Meta首席SI官汪滔。图片来源：Wikimedia Commons（CC BY-SA 4.0）。Photo: Alexandr Wang, Meta's chief SI officer. Source: Wikimedia Commons (CC BY-SA 4.0).
