
原始文档
GPT Image 2 · en
Scene:
A square single-page manga-style launch comic with exactly eight panels arranged in a clean 2 by 4 grid. Crisp black panel borders, warm screen-tone texture, limited cyan, magenta, black, and cream palette, high production polish, and clear reading flow from top-left to bottom-right.
Subject:
One consistent protagonist appears across all eight panels: short silver hair, round glasses, cropped black jacket, white hoodie, cyan scarf, magenta sneakers. The protagonist discovers, lifts, and presents a glowing glass orb with the exact readable text GPT-IMAGE-2 inside the orb. The orb stays the same size and design across panels. Expressions change, but the face, hair, outfit, and prop identity remain consistent.
Important details:
Eight distinct panels, one coherent launch story, consistent character model, consistent outfit, consistent glowing orb, dynamic manga camera angles, clean speech balloons with pseudo-text except the orb text GPT-IMAGE-2. Include a few non-readable sound-effect shapes and speed lines. The layout must be readable as one finished comic page, not eight unrelated stickers.
Use case:
Reference image for a four-image series testing character and product-prop consistency across multi-panel comic pages.
Constraints:
Exactly eight panels, preserve one protagonist identity across all panels, only readable text should be GPT-IMAGE-2 on the orb, no real brand logos, no watermark, no extra main characters, no broken panel grid, no photorealism, preserve exact 1:1 aspect ratio.使用方法
- 复制完整原始提示词,并保留其中的参数。
- 按需求替换主体、文字和构图要求。
- 在兼容的外部工具中使用,核对当前使用条件并检查结果。
开始前需了解
- 先确认是否需要参考图。标注兼容模型,不代表该示例由这个模型生成。
- 核对名称、数字与画面文字;不同模型版本的结果可能不同。





