Original document
GPT Image 2 · en
Create a premium conceptual beverage advertisement for {argument name="brand name" default="Coca-Cola"} on a vertical 4:5 canvas. Center a chilled red aluminum soda can, front-facing, covered in realistic condensation droplets, with a silver pull tab and rim visible at the top and the classic white script logo across the can; include the smaller can text “ORIGINAL TASTE.” The background is a seamless saturated red studio backdrop with luxury commercial lighting and a soft product shadow. A large metallic zipper runs horizontally across the image and wraps around the can at mid-height, with detailed interlocking teeth and a central zipper pull in front of the can; the zipper opens the scene into two contrasting worlds: on the left, a warm orange sunset over a calm water horizon with a small distant city skyline silhouette, and on the right, a cool blue snowy alpine landscape with frosted pine trees and mountains. The zipper seam should curve organically like a wave, visually implying the can contains magical worlds inside. Add minimal white ad copy: at the top center write {argument name="top slogan" default="open happiness"}, with “happiness” in bold; near the bottom center write {argument name="bottom slogan" default="real magic is inside"}, with “inside” in bold; place a small white Coca-Cola logo beneath it. Use hyper-realistic FMCG product photography, crisp macro details, glossy reflections, balanced red-orange-blue color contrast, Apple-level minimalist composition, Cannes Lions conceptual advertising style, no extra objects, no people, no watermark.How to use
- Copy the complete original prompt, including any parameters.
- Replace the subject, text and composition with your own requirements.
- Use a compatible external tool; check its current access conditions and review the result.
Before you start
- Check whether a reference image is required. A compatible model label does not mean this example was generated with that model.
- Keep names, figures and visible text accurate; results can vary between model versions.
