GPT Image 2.5 Prompts

【S.012】 Clone

Transform one photo into a borderless vertical triptych. The center preserves an unchanged crop of the original image as clear evidence of the scene, while the upper and lower sections each select a different subject from the photo for cloning and inversion. The result is a surreal image grounded in the original environment. The upper panel can focus on a primary subject such as a person or animal. The lower panel works well with objects that have clear structures, such as boats, buildings, railings, or landmarks. Each generated area keeps one recognizable subject at a visible scale, while allowing smaller inverted figures to blend into the foreground, background, water, rocks, roads, or architectural boundaries. The original scene’s lighting, perspective, occlusion, materials, and depth of field are preserved. Ideal for photographic collages with separately duplicated and inverted people and objects, source-photo-driven surreal triptychs, and visual experiments that retain features such as landmarks, mountains, and skylines. The final output is a single vertical 2:3 image with no text, borders, watermarks, or decorative titles.