Overall rating
78
#3 of 19 models
Per-task results
88 tasks across 8 types, every one recorded. Pick one to watch exactly what the model did.
By task type
Aiming
Disqualified. Broke the task's rule “Do the editing in Photopea, using only rotation, cropping, and proportional resizing — I want the original photograph, with no retouching, filling, or stretc...”
Task IDaiming/banner_crop
Cost is the model / provider cost. Every task counts toward the pass rate, including any whose gold answer is not yet verified (flagged "Gold unverified"). Each trace was reviewed against its task's rules: a pass that broke one is disqualified and counts as a fail.