Style Spectrum: Illustration Translation from Realism to Flat

Using a single still-life photograph as the only source image, use the gpt-image-2.6-sunburst image-to-image pipeline to translate it in sequence into four illustration styles: realism, semi-realistic flat color, hard-edge flat, and minimalist geometric. The aim is to make "style strength" into a comparable spectrum—so the trade-off differences of the same composition, same subject, and same light-source relationship across the four styles are clear at a glance. Each skin is first generated as independent source material, then submitted as image_urls to the editing pipeline, to observe how the model maintains subject recognizability and compositional stability as it reduces texture layers. The schema side also sets non-media parameters such as resolution, aspect_ratio, output_format, and quality, making it easy for developers to reproduce against a real submit input. It is suitable for scenarios that need to establish a style baseline, conduct visual exploration, or showcase model editing consistency externally.

Style style-transfer · Illustration · image-to-image

Model GPT Image 2.6

Reference inputs

Source images

Try it yourself

Prompt

Try this model

Edit the illustration layout to show the same still-life composition: a ceramic vase holding dried branches on a linen-covered table, presented as a four-panel style lineage. The leftmost panel is photo-level realism, showing ceramic glaze, linen fiber weave, and soft window light; the second panel is semi-realistic flat color with simplified shadows that retain hints of texture; the third panel is hard-edge flat vector illustration with two-tone shading and no gradients; the rightmost panel is minimalist geometric abstraction that reduces the objects to a few shapes. All panels keep a consistent camera angle, the same object placement, and the same warm daylight direction, with ample negative space, using low-saturation terracotta and sand tones, in a 16:9 landscape layout.

Copy the full prompt and adapt it for your own generation.