Intel / article

Muse Wonder

Source: artificialanalysis.ai
Author: Artificial Analysis articles
Date: Published Oct 12, 2026

Context

Muse Wonder and Muse Wonder Full are different subjects and different access options. Artificial Analysis' launch comparison reports stronger aggregate results for Full, but Meta made that variant available through Muse Code rather than its first-party API at launch. A favorable result for Full should not be silently transferred to an API deployment of Wonder.

The same distinction limits cost comparisons. The evaluator calculates per-task API spending for Wonder, including substantial input usage, but gives no equivalent figure for Full because it is not priced per . The article therefore provides evidence about both capability and availability without establishing a directly comparable dollar cost for the two variants.

Terms in this piece · Glossary

eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.

Read the source artificialanalysis.ai