Metaclosed weightsCurrentReasoning
Muse Wonder Full
Larger Wonder model, Muse Code subscriptions only ($50 and $200 a month), no API. Meta reports DeepSWE 1.1 79.0%, SWE-Marathon 57.5%, JobBench 75.3%.
Oct 12, 2026
—
—
1M
Benchmark scores
| Benchmark | Score | Δ vs prior | Rank (current) | Configuration | Source |
|---|---|---|---|---|---|
Artificial Analysis Intelligence Index | 56 | #4 of 16 | max reasoning · independent run | venturebeat.com ↗ |