Detailed stills
Scenes that need materials and small objects held in place.
Online tool
GPT Image 1.5 - AI Image Generation
Create stunning images with GPT Image 1.5. Supports text-to-image and image editing capabilities.
GPT Image 1.5 for prompt-faithful stills
GPT Image 1.5 is the previous OpenAI still-image model in this studio. It keeps lighting and fine detail from a written brief. This page opens text-to-image with GPT Image 1.5 selected.
Prefer it when you already have a 1.5 workflow, or when you want a lower-cost OpenAI option than GPT Image 2. Quality in the rail still changes credit use.
What this model is for
Composition lock
Keep camera and layout from a long prompt.
Product frames
Catalog-style pictures from a description.
Edits from a photo
Add a source image when you need a change, not a new scene.
A clearer path from idea to output
Learn what the tool does here, then create in a focused workspace built around the task.
Preselected model
The dropdown opens on GPT Image 1.5.
Quality tiers
Low, medium, and high, same family as other GPT Image models.
Shared workspace
Same canvas and rail as every other text-to-image model.
History
Download again from History after the job finishes.
Built for everyday visual work
Campaign drafts
A first photoreal pass before you switch to Image 2.
Catalog comps
Simple product on a described background.
Look tests
Try lighting language without a photoshoot.
Iterative edits
One source photo, several prompt changes.
Create in three steps
- 1
Describe the result
Write a clear prompt describing the image you want to create.
- 2
Review the setup
Choose the available task and output options in the control rail.
- 3
Process and export
Review the result in the workspace, then download it or continue with another image.
Related tools
Frequently asked questions
How is this page different from the workspace?
This page explains the tool and its use cases. The /app workspace is where creation happens.
Do I need to upload an image?
No. This tool starts from a text prompt; an image upload is not required.
Which image formats are supported?
Image inputs generally support JPG, PNG, and WEBP. The upload area shows the exact limits.
Can I reuse a result?
Yes. Download the result, revisit it from history, or use it as the next input where supported.
Should I use Image 2 instead?
Use Image 2 when type inside the picture must be sharp. Stay on 1.5 if you already like its look or cost.
Does the page generate here?
No. Generation runs in /app/text-to-image with this model selected.
What quality can I expect?
Quality depends on the model and settings in the workspace. Download the result, then run again with a different prompt, model, or size if you need another take.
Do I need an account?
You can read this page without signing in. Creating in the workspace uses credits and requires an account.
Can I use the result in my work?
Download the file you generate and use it in your projects. Read the current terms if you need commercial-use details.
Ready to create?
Open the VibePicture workspace and turn your idea into a usable visual.
Open Studio