Meta reports two new AI models called Muse Wonder and Muse Wonder Full

In short: Meta is reported to be working on new AI models called Muse Wonder and Muse Wonder Full to power its Muse personal assistant.

What happened

Meta has new AI models in development called Muse Wonder, and a related name, Muse Wonder Full, according to reporting by The New York Times. The models are meant to support Meta’s Muse personal assistant, which is the company’s tool for helping people with everyday and work tasks.

Based on other public listings and reporting, Muse Wonder is described as a “reasoning model,” meaning it is designed to handle multi-step requests, like planning a trip, organizing a project, or breaking a big task into smaller parts (like assigning chores to different helpers). It is also described as being able to take in several types of information, including text, images, video, audio, and PDFs.

One listing says Muse Wonder has a 1-million-token context window. A token is a small chunk of text (roughly a few characters to a word), so this is like giving the assistant a very large notebook to keep track of long documents and long conversations. The model is also described as being able to use tools, browse the web, and operate apps, and to carry helpful context across chats.

Meta has positioned these models as aiming to compete with leading AI systems from companies like Anthropic and OpenAI. However, the available sources do not clearly explain what “Muse Wonder Full” is, or provide reliable head-to-head performance results.

Why it matters

If these models work as described, Meta’s assistant could be better at longer, messier tasks that usually require lots of back and forth. For regular people, that could mean less time spent repeating details and more help turning plans into finished work.

Source: NYTimes