T
Text-rich visuals
Create layouts where readable text is part of the image.
Models
Dolly·May 13, 2026
GPT-2-Image
Generate images with OpenAI’s GPT-2-Image model.
Dolly·May 13, 20261pinned model
3core strengths
2task types
1+reference images
What it does
Use GPT Image 2 without switching models mid-request
This model-specific skill keeps the request pinned to GPT Image 2 for generation and editing. It is suited to instruction-led changes, readable text inside images, and compositions that combine multiple references.
- 1Pin every request to one image model
- 2Handle text rendering, instruction edits, and multi-image composition
- 3Support both generation and editing tasks
GPT Image 2 example
A first-party Drama.Land model output
How it works
Generate with GPT Image 2
1
Provide the brief
Write the desired composition and attach any source or reference images.
You bring
Prompt and optional references
You get
GPT Image 2 request
2
Generate or edit
Run the request with the model pinned to GPT Image 2.
You bring
Configured request
You get
Generated image candidate
3
Review and refine
Inspect text, composition, and instruction fidelity, then request a targeted adjustment.
You bring
Image candidate and feedback
You get
Saved final image
Use cases
Use GPT Image 2 for its specific strengths
I
Instruction edits
Make precise changes while preserving specified content.
M
Multi-image composition
Combine roles from several references into one coherent result.
Good to know
GPT Image 2 questions
Is this page model-specific?+
Yes. Every request from this tool is pinned to GPT Image 2.
Can it edit an image?+
Yes. Provide the source and describe both the change and the elements that must stay fixed.
How is this different from Image Studio?+
Image Studio can route across multiple services; this tool always uses GPT Image 2.