Meta has unveiled two new AI models, Muse Wonder and Muse Wonder Full, which it says deliver state-of-the-art performance on personal and professional agentic tasks. The models are available starting today for Muse Code subscribers, with Wonder also accessible through the Meta Model API, according to The Verge.
The launch marks Meta's latest attempt to catch up in the AI race. The company says these models will eventually power its Muse AI agent, starting with subscribers, though it hasn't shared a timeline or specific details for that rollout.
Meta Muse Wonder Performance and Benchmarks
Meta claims the new models can rival leading offerings from OpenAI and Anthropic, including GPT-6 Astra, Claude Opus 5.5, and Claude Fable 5.1. These claims are based on internal testing on several industry benchmarks, not independent evaluations.
It remains unclear how Muse Wonder and Muse Wonder Full compare to Google's Gemini 4 Argon, whose capabilities are still being assessed. Meta has not provided third-party benchmark results to verify its performance claims.
Meta's AI Strategy and the Muse Ecosystem
The launch is part of Meta's aggressive push to reestablish itself as a major force in AI. The company has invested billions in talent and infrastructure, following the launch of Muse Spark, the first model from its revamped Superintelligence Labs division, and the buzzy AI agent Muse.
Meta has been steadily expanding its AI lineup, including with image model Muse Image, and maintains a heavy focus on bringing useful AI agents to the masses through Muse. The company says it plans to develop specialized models it "can bring to everyone using Muse," though specifics remain scarce.
What it means for you
For now, Muse Wonder and Muse Wonder Full are available only to Muse Code subscribers, with Wonder also accessible via the Meta Model API. There's no word on when or if these models will be available in Tunisia, or what pricing might look like for local users.
