Create with GPT Image 2.6
Choose Flare for a quick creative direction or Sunburst for detailed work. Describe your image, add an optional reference, and create in the workspace below.
Raster images offer 1K, 2K and 4K tiers. SVG and GIF use the 1K selection; GIF frames are 512 × 512 pixels. Flare uses medium quality and Sunburst uses high quality in this first version.
3 successful results for $2.99: use new or remaining $2.99 pack passes for Flare, Sunburst, SVG or GIF. One successful result uses one pass, with no subscription required.
Generation is available only when the selected model is available in the editor. If it is unavailable, you can prepare a prompt and return later.
Start with Flare, or select Sunburst for detailed work.
Describe the subject and layout. Add a photo when an existing reference matters.
Check the settings and your selected credit or pass cost, generate, then preview and download your result.
Flare is the speed-focused version of GPT Image 2.6. Use it to explore a composition, draft a poster, or try several product-image directions. Describe the layout, lighting and exact wording you want, then review the result before choosing your next edit.
Sunburst is designed for detailed output and consistency across edits. Use it when material, composition or reference details need closer attention. Name what should stay the same and make one deliberate change per edit; check faces, lettering and small objects in every result.
Use PNG or JPEG for conventional image workflows, static SVG for scalable illustrations, or GIF for a short animated loop. The format controls show the available choices for your selected model and size. SVG and GIF are generated through the model; renaming a PNG file is not part of this workflow.
Adapt these prompt ideas to your own project. They are starting points, not verified output examples.
A studio product image of a matte terracotta ceramic mug on a warm cream surface. Keep the mug in the left third, a soft shadow below, and clean negative space on the right. Soft window light, visible ceramic texture, no text.
Create a minimal botanical exhibition poster. Put the exact headline "GROW SLOW" at the top, a single large fern in the center and the small text "SATURDAY 10 AM" at the bottom. Deep green and warm ivory, clear spacing, restrained typography.
GPT Image 2.6 is OpenAI's image model release with Flare and Sunburst variants. EzImageAI brings these choices into its existing image workspace for prompt-based creation and reference-guided editing, with PNG, JPEG, static SVG and GIF output options.
Choose Flare for speed-focused exploration and Sunburst for detailed output and consistency across revisions. This version uses medium quality for Flare and high quality for Sunburst. Actual completion time and quality depend on your request and service availability.
Yes. Add a reference photo to the composer and describe what to change and what to keep. Leave the reference empty for text-to-image generation. The editor currently accepts one reference image per request.
Yes. Choose SVG for a static illustration or GIF for an animation with 2–16 frames. This version uses 512 × 512 pixels per GIF frame. Select the 1K output tier to access these formats; the separate SVG and GIF pages open the same workspace with the relevant format selected.
Choose account credits or an eligible image pass, then check the cost before generating. The $2.99 pack includes 3 successful results, with no subscription required when paying with these passes. Credit-funded requests retain their account and subscription requirements. Each new generation or edit has its own quote.
Yes. Both new purchases and remaining uses from existing $2.99 packs cover GPT Image 2.6 Flare and Sunburst: PNG or JPEG at 1K, 2K or 4K, static SVG, and GIF with 2–16 frames. One successful result uses one pass, including a whole GIF regardless of its frame count. Choose Image pass as the payment source; buying a pack does not automatically switch your selection. Earlier packs sold for a different price keep their original terms.
Your reference images and results use the private EzImageAI workspace. They are not automatically published in a public gallery. Upload only material you have permission to use, and follow the content safety and usage rules shown in the editor.