Image prompt
Optional reference images for subject, product, or style
Aspect ratio, output size, and background
Image tools
GPT Image 2.5 tool
Create images with GPT Image 2.5.
Describe the image, pick its shape and output size, and ask for a transparent background when the subject has to sit on another design. Flare runs by default; Sunburst is one word away.
1K, 2K, or 4K output on one page Transparent background on request Up to 8 references here, 16 per model request
Start with the messy versionOptional choices help Vizify shape the handoff.
Try a starting point
Your prompt, settings, and references open in Vizify, which keeps one GPT Image 2.5 job and one artifact from request to completion.
One GPT Image 2.5 image artifact
An image with alpha when a transparent background succeeds
Reusable prompt and settings for the next variation
What it does
Built to hand back finished work.
Output sized for the job
Pick 1K for drafts and quick exploration, 2K for web and social layouts, or 4K when the image has to hold up on large screens and detailed crops.
Transparent backgrounds without a cutout step
Choose Transparent and GPT Image 2.5 generates the subject with an alpha channel. Vizify inspects the result, because hair, glass, and soft shadows can still leave halos.
Two variants behind one page
Flare is the default. Ask for Sunburst in the prompt to use it instead; Vizify's catalog lists the same parameters for both and does not rank one above the other.
References that stay in the request
Add product shots, a character sheet, or a style board. They travel with the GPT Image 2.5 job instead of being summarized into text first.
Model referenceModel details
Review GPT Image 2.5 before choosing the workflow.
See current Vizify variants, supported inputs, resolutions, limitations, and other compatible tools.
FAQ
Questions, answered.
What is GPT Image 2.5?
An OpenAI image model family that Vizify offers as two variants, GPT Image 2.5 Flare and GPT Image 2.5 Sunburst. This page runs Flare unless you ask for Sunburst.
Which aspect ratios can I pick?
Auto, square 1:1, landscape 3:2, portrait 2:3, widescreen 16:9, and vertical 9:16. The model also lists 4:3, 3:4, and 21:9, which you can ask for in the prompt.
Can GPT Image 2.5 generate 4K images?
Yes. Choose 1K, 2K, or 4K. 1K is the default; larger outputs take longer and cost more.
Can it make a transparent background?
Yes. Set Background to Transparent and the request asks for alpha output. Vizify checks that the result really is transparent rather than white or checkered, and fine edges can still need cleanup. Opaque forces a solid background; Auto lets the model decide.
How many reference images can I use?
GPT Image 2.5 accepts up to 16 per request. This page stages up to eight PNG, JPEG, or WebP files of up to 10 MB each.
What is the difference between Flare and Sunburst?
In Vizify's catalog both share the same resolutions, ratios, reference limit, and background options, and Vizify does not claim a quality difference. Flare is the default here; Vizify may use Sunburst when you ask for it or for a detail-heavy edit of your references.
How is GPT Image 2.5 different from GPT Image 2?
GPT Image 2.5 adds the background option and accepts up to 16 references instead of eight. GPT Image 2 remains Vizify's default image model and keeps its own generator page.
Can I edit an existing image with GPT Image 2.5?
Yes, as a reference-led edit: upload the image and describe the change, naming what must stay the same. It is a semantic edit, so nearby details can be redrawn; review the result before reuse.
How many images does one request create?
One. Each request starts one generation job and one artifact. Send separate requests when you want several directions, so each keeps its own status and usage.
What happens while the image is generating?
Vizify creates one pending artifact and updates that same artifact when the asynchronous GPT Image 2.5 job completes or fails.
Choosing size and background
Output size and background are the two settings that change what you can do with the file afterwards. Start at 1K while the composition is still moving, then regenerate at 2K or 4K once it is right. Choose Transparent only when the subject will be placed on another layout; for a finished scene, Opaque or Auto keeps the background the model painted.
Stable from request to artifact
The page exposes the GPT Image 2.5 parameters Vizify actually passes: aspect ratio, output size, background, and references. Quality, file format, compression, moderation, and seed are not public controls. Each request keeps one job and one artifact identity from pending to completed, and a failed job is reported rather than silently retried under another model.
Prompt to finished image
Open Vizify