What the evaluator found
Artificial Analysis published results for Muse Wonder and Muse Wonder Full on October 12. In its Intelligence Index v4.3, both improved on the tested Muse Spark 1.3 configuration. The report separates the two models: it says Wonder is offered through Meta’s API and Muse Code, while Wonder Full is limited to Muse Code at launch.
Why it matters
A new family name can hide meaningful differences in access, settings, and cost. Treat the evaluator’s results as evidence about its tasks and configurations. They do not establish which assistant will be best at your everyday work.
Our coverage boundary
This is a report on an independent evaluator’s published work, not our hands-on test. We could not retrieve a readable Meta launch article during this check. Verify current access with Meta before subscribing or building around either version.