OpenAI’s Images 2.6 announcement introduces GIF creation and new editing controls alongside faster image generation. The company says the release is available in ChatGPT and Codex, with two named variants for API developers.
Motion becomes part of the workflow
The announcement presents GIFs as a way to animate characters, product imagery and storyboards. These examples expand the output a creator can work toward, but they do not remove the need to inspect motion, frame consistency and the intended loop.
For a small creative project, start with a clear subject and a simple motion, then review the result at its actual playback speed. That is an editorial evaluation suggestion, rather than a claim of a new Franklin performance test.
More control over an edit
OpenAI says image markup is coming to the API so developers can identify where a change should be made. The release also includes SVG exports and a complexity parameter.
The announcement distinguishes complexity from image quality: complexity controls how much planning goes into the scene, while quality controls detail and fidelity. A routine edit and a scene with many constraints may therefore call for different settings.
Availability and speed claims
The company names GPT-Image-2.6 Flare and Sunburst as API options. It describes Flare as a faster option and Sunburst as offering higher quality and consistency at medium or high quality settings.
OpenAI’s stated generation-speed improvement is a vendor-reported comparison. Actual turnaround and the number of revisions needed will depend on the request and chosen settings. Check the current model documentation and pricing before building a production workflow around either variant.
The practical change is a broader set of creation and revision controls in one release. A useful trial should evaluate a complete finished asset, including edits and animation review, instead of judging the update only by a first-generation speed figure.