原文
GPT Image 2 · en
Scene:
A quiet top-down studio illustration of an analyst workbench, square crop, warm desk surface, soft afternoon light, and carefully arranged paper layers. The style is polished editorial illustration with tactile paper texture, graphite lines, and restrained accent colors.
Subject:
A complete visual reasoning workspace: one central blank research card, three source cards, a small uncertainty scale, colored string connections, tiny icon stickers, and a final decision tray. The arrangement shows how messy inputs become a structured conclusion. The cards contain only non-readable glyph marks and abstract chart shapes, not real words or numbers.
Important details:
Clear information architecture, visible reasoning sequence from left to right, balanced negative space, precise alignment, and subtle handmade imperfections. Include paper clips, transparent overlays, arrows, and color-coded evidence groups. The image should look like thinking made visible without exposing literal text.
Use case:
Independent Thinking Mode single image, testing whether a complex reasoning workflow can become a clear editorial object composition.
Constraints:
No readable text, no real data, no company names, no logos, no watermark, no photorealistic laptop screen, no cluttered desk mess, no human hands, preserve exact 1:1 aspect ratio.使い方
- パラメータを含む元のプロンプト全体をコピーします。
- 被写体、文字、構図を自分の要件に置き換えます。
- 対応する外部ツールで使用し、利用条件と結果を確認します。
開始前の確認
- 参考画像の要否を確認してください。互換モデルの表示は、そのモデルでの生成実績を意味しません。
- 名前、数値、画像内の文字を確認します。モデルの版で結果は変わります。
