Compare how closely each model reproduced the reference designs. Rank the results using the average of all judges or an individual judge’s scores.
8 aspects · scores from 1 to 10 Higher scores rank first
Rank
Result
Average total/ 80
Average score/ 10
† Qualified assessment: the native client resumed the same session before returning any assessment text. The original audit failure is retained; open the result for details.
Judge totals: ↑ above, ↓ below, or = equal to that judge’s average total across the available ratings for results. These color baselines stay fixed when you filter designs.