Meta ships Muse Wonder models to its API and Muse Code
- Source
- x.com
- Date
Meta introduced Muse Wonder and Muse Wonder Full as the successors to April's Muse Spark, aimed at personal and professional agent work. Meta says the models were trained at its Prometheus data center and add features it calls Active Memory and Active Focus. Muse Wonder is available today through the Meta Model API and Muse Code. Muse Wonder Full is limited to Muse Code subscribers, with the Muse app to follow. Artificial Analysis scored Muse Wonder at 52 and Muse Wonder Full at 56 on its Intelligence Index, up from 48 for Muse Spark 1.3, with the largest gains on agentic evals such as Terminal-Bench 4.0, where Full gained 24 points. Theo of t3.gg pointed out the practical catch: the headline results come from Full, which you cannot reach over the API. If you are evaluating Meta for an API workload, run your own tests against standard Wonder rather than relying on the launch numbers.
We’re excited to introduce the Muse Wonder series with Muse Wonder and Muse Wonder Full, our most capable models yet. Both are frontier models, built to deliver state-of-the-art performance for personal and professional agentic tasks. Muse Wonder is available today via the Meta Model API and Muse Code. Muse Wonder Full is available today for Muse Code subscribers. Muse Wonder will also soon power experiences in our personal agent product, Muse, rolling out first to subscribers.
Context
Meta's October 12 announcement of the Muse Wonder series names two models with different access. Muse Wonder is available through the Meta Model API, which developers call from their own software, and through Muse Code, Meta's coding . Muse Wonder Full is available only to Muse Code subscribers. Meta's launch page adds that Muse Wonder will later power Muse, its personal agent product, rolling out first to subscribers.
Meta describes both as frontier models with state-of-the-art performance on agentic tasks, which are its own claims, and the post does not say how the two models differ. An independent measurement makes the access split more consequential: Artificial Analysis scored Muse Wonder Full at 56 and Muse Wonder at 52 on its Intelligence Index at maximum reasoning effort. On that measure, the higher-scoring model is the one Meta does not offer through its Model API.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
Checking sign-in…
Loading comments…