ChatGPT Images 2.6 adds GIF editing and reusable reference images

What changed

The reference-image feature lets users reuse images of people, pets or characters without uploading them again for each creation. The GIF workflow adds editing at the frame level. Image markup and SVG export provide further ways to refine or use an output, as described by OpenAI.

OpenAI presents Flare as the faster option, capable of generating an image in as little as six seconds, and Sunburst as the higher-quality option with greater consistency across edits. These are vendor descriptions, not an independent benchmark of a particular advertising-production workflow.

The company says Images 2.6 is available across tiers in ChatGPT and Codex on web, desktop and mobile, with the two named models available through the API. Product availability does not establish that every requested format, output or edit will satisfy an advertiser’s brand and platform requirements.

For creative teams, a useful evaluation would use an approved reference asset and a defined change, then compare the result with the original. Inspect product shape, labels, repeated characters and the consistency of frames in an animation. Faster generation can reduce waiting time without removing the need for an asset review before a campaign uses the result.

The source is OpenAI’s unsigned official announcement, supported by the dated ChatGPT release notes. The release date is 4 November; later social reporting should not shift the feature to a different launch day. This article covers the image update separately from the company’s cybersecurity model, which has a different purpose and access programme. It is a product update rather than a recommendation to replace a creative vendor.