Same Berserk spread, 4 AI colorizations. Looks like GPT wins again?

An unofficial AI colorization test compared four models on a dense Berserk spread, focusing on their ability to add color without redrawing the original artwork. The user wanted to see how well the models handled the intricate ink and crosshatching while keeping the drawing intact. The test aimed to compare editing behavior, with one model, Seedream 5.0 Pro, costing $0.045. The title suggests that "GPT wins again?" in this comparison.