Knowledge base · images
ChatGPT
Text-to-image generation with strong instruction following and neat text in the image.
Open in studio →What it can do
- ✓Follows complex instructions: several objects, their placement and conditions in one prompt.
- ✓Photorealism and illustrations, plain natural language.
- ✓Text and lettering in the image are more stable than in most models.
How to write a prompt
Describe the scene in a full sentence: subject → environment → style → lighting → angle. For lettering, put the text in quotes and keep it short. Describe a complex composition part by part in one request.
A prompt that works
A "20% off" poster in flat style, bright contrasting colors, clean typography, a basket of groceries in the center
Avoid this
“make a picture”
No subject, style or context. Describe the scene in a full sentence.
Examples
Poster with text
An advertising poster for a coffee shop with the words "Good morning", warm tones, a photo of a cup of coffee, soft light
Complex scene
A designer’s desk from above: a laptop, coffee, a notebook and pencils, a neat flat lay, natural light
Variants within the group
- GPT Image 2 / GPT Image 1, Version 2 is newer and higher quality, version 1 is faster.
Tips
- →Write a description of the scene, not a set of words.
- →For lettering, use quotes and short text.
FAQ
How much does a generation in ChatGPT cost?
from 2 to 4 tokens per generation, depending on the chosen version and settings. Current plans and token packs are on the /en/pricing page.
Can the results of ChatGPT be used commercially?
Yes, with an active Pixyn subscription. Note that some models have provider restrictions by industry, check the terms of the specific version.