GPT Image 2.6 AI Image & GIF Generator

GPT Image 2.6 is the image generation and editing model behind Images 2.6, which OpenAI released on November 4, 2026 in ChatGPT, Codex and the API. In the API it comes in two versions, GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst. OpenAI says it generates images 50% faster than before, can output animated GIFs, and keeps characters, products and brand details consistent as you edit.

Three things: it is faster, it can make animated GIFs, and edits are steadier, so the parts you did not ask to change stay the same. Both models can return a transparent background. GPT Image 2.5 is still on Veevid and has a few aspect ratios 2.6 does not, such as 27:16 and 9:8, plus an auto ratio.

OpenAI says Images 2.6 generates images 50% faster than before. In our tests on Veevid, a 1K image took about 9 seconds with Flare and about 55 seconds with Sunburst. Times vary with resolution, prompt and load.

Yes. Choose GIF as the output and you get a looping GIF with 9 frames. It works from a text prompt or from an uploaded image, for example "Animate this: the dog wags its tail and blinks". GIFs come in 1:1, 16:9, 9:16, 4:3 or 3:4 and cost 16 credits each. OpenAI's own examples include animating a brand mascot or character, turning a product shot into a loop, reactions and memes for social media, and animated storyboards before video production; even a simple sketch can become an animated character. Prompt tip: describe the motion step by step, keep the same character and a fixed camera, and ask for the last frame to flow back into the first so the loop is seamless.

Yes. Set Background to Transparent and you get a PNG (or WebP, if you pick that format) with a real alpha channel, ready for logos, stickers and product cutouts. It works for GIFs too: the animation comes back with a transparent backdrop. GIF pixels are either fully transparent or fully solid, so the edges have no soft fade and can show a thin hard outline on dark backgrounds. If that matters, pick Opaque and ask for a background color that matches where the GIF will go. Opaque paints the background your prompt describes, and Auto lets the model choose. GIFs start on Opaque, because on Auto the model often ignores the background color you asked for. The price is the same whichever you pick.

Start with Flare: it is the default and the fastest. Switch to Sunburst for final assets where detail matters, or for edits where consistency is critical; OpenAI says Sunburst brings higher-quality images and even greater consistency across edits. Both cost the same credits on Veevid, so the choice is only speed against detail.

Switch to Image to Image, upload up to 16 reference images and describe the change, for example "Change the sneaker color to forest green suede and keep everything else the same". Saying what should stay unchanged helps. The first reference image is included in the price; each extra image adds 2 credits (3 images at 1K = 16 credits).

8 aspect ratios: 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, 21:9. Output comes in 1K, 2K or 4K, and 4K goes up to 3840×2160. GIFs come in five sizes: 1:1 512×512 / 16:9 768×432 / 9:16 432×768 / 4:3 576×432 / 3:4 432×576. Still images download as PNG (default), JPEG or WebP, all at the same price. Making a logo, icon or illustration? Pick SVG to get a scalable vector file (12 credits, text to image only).

12 credits for a 1K image, 24 for 2K, 24 for 4K, 16 for a GIF, and 12 for an SVG. When editing, the first reference image is included and each extra one adds 2 credits, GIF edits included. Flare and Sunburst cost the same.