Meta’s New Muse Wonder Models Challenge OpenAI, but Access Comes With Limits
Meta's Muse Wonder AI models challenge OpenAI and Anthropic with new agent capabilities. Explore benchmark results, pricing, access restrictions, and privacy risks. The post Meta’s New Muse Wonder Models Challenge OpenAI, but Access Comes With Limits appeared first on TechRepublic .
Meta has unveiled its latest frontier AI models, Muse Wonder and Muse Wonder Full, which challenge the dominance of OpenAI and Anthropic. These models are designed to power autonomous agents for tasks like coding, complex reasoning, and document analysis. Unlike earlier open-weight Llama models, Muse Wonder models are not publicly downloadable, limiting access to developers through Meta's Model API and Muse Code programming assistant.
Pricing starts at $4 per million input tokens and $20 per million output tokens for the Muse Wonder model, with Muse Wonder Full available only to Muse Code subscribers for $50 or $200 per month. Both models feature Active Memory and Active Focus, which allow them to retain information and adapt to errors during multi-step tasks.
Independent evaluations rank Muse Wonder Full at 56 on the Intelligence Index, tying it with Anthropic's Claude Sonnet 5.5, while Muse Wonder scores 52 and OpenAI's GPT-6 Astra 53. This puts Meta on par with leading providers, but benchmark results don't guarantee reliable real-world performance. Privacy is a concern, as Meta trains the models to protect sensitive data, but defaults to a Contributor mode for Muse Code subscriptions, which may use submitted content for model improvement.
Businesses must weigh factors like cost, reliability, data handling, and real-world task performance when evaluating these models for automation tasks.
Written by urgent.news from TechRepublic's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
