Reasoning and Instruction Following Puzzle Prompt

A complex prompt designed to test the model's reasoning and instruction-following capabilities by asking it to generate a visual puzzle: four related realistic objects on a table, a title text asking 'What do these objects have in common?', and very small, upside-down text at the bottom revealing the answer.

Prompt by @Joe β­• Dot β­• AverageView the original post β†—Published 2025-12-18CC BY 4.0

The prompt

Fill in the blanks β€” 1 placeholder

Edit any of these and the prompt below updates. Copy gives you the filled-in version.

"Show me 4 realistic objects on a table. The objects are related in some way, this is a puzzle. Add title text across the top, "What do these objects have in common?" And very small upside down text on the bottom which reveals the way in which they are related."

Generate opens this prompt in the editor, ready to run.

Model
GPT Image 1.5
Media
Image
Language
English
Words
53

How to use this prompt

  1. 1Copy the prompt β€” fill in any placeholders first and the copy takes your version.
  2. 2Paste it into your image model, or press Generate to open it here with your own photo.
  3. 3Run it twice. Generation is not deterministic and the second pass is often the keeper.
  4. 4Finish it with the free tools: upscale, crop to the platform you need, compress before sharing.

Go beyond generating the image

Everything after the generation, on the same site.

Upscale to 4K β†’Remove the background β†’Crop for a platform β†’Compress before sharing β†’All AI apps β†’

More from GPT Image 1.5

Same model, same dataset β€” GPT Image 1.5 prompts people have shared.

View all GPT Image 1.5 prompts β†’

All image categories

Use cases
Styles
Subjects