OpenAI launched Images 2.6 on November 4, making its new image model available in ChatGPT, the Codex coding tool and through its API. According to OpenAI, the model works worldwide on all subscription plans across desktop, mobile and web.
The update also adds two features that go beyond still pictures: animated GIF creation and the ability to save personal photos for reuse in later chats.
Faster generation and more consistent edits
Images now generate 50% faster than before, OpenAI said. Edits are also more consistent, so clothing, faces and product details stay unchanged when users adjust the background or another part of an image.
Changes are requested through the chat window, and the company says the same settings apply across its supported platforms.
Animated GIFs arrive in ChatGPT
ChatGPT now lets users create animated GIFs by describing what should move or by uploading a still image and asking for an animation. OpenAI’s help article on ChatGPT Images says users can adjust the playback speed afterwards and then save or share the GIF.
There are limits. ChatGPT cannot edit an animated GIF that a user uploads. As of November 7, the help article still described GIF creation as rolling out, so it may not be available in every account yet. On November 8, the ChatGPT Images page in Germany showed an anime reaction GIF template without sign-in. OpenAI has not shared any GIF limits for any plan.
Saved photo references raise privacy questions
Users can upload a picture, give it a name and mention it later with @ to reuse it. OpenAI’s example is a dog photo saved as Pepper, which can then be used to request a watercolor portrait of @Pepper without uploading the photo again. ChatGPT may sometimes suggest saving a photo when a user describes it.
Once saved, a photo stays in the account, but OpenAI’s help page does not explain where users can manage or delete saved references. Generated images are handled differently: they are saved under Images and are deleted only when the conversation that created them is removed.
One report said that combining a saved reference photo with animation could make it easier to place people in scenes that never happened. OpenAI has not explained how it prevents misuse of GIFs showing real people. Its provenance help article says images made with ChatGPT include Content Credentials and an invisible SynthID watermark, but GIFs are not mentioned separately.
Users should save photos of children or other people only if those people agree. Uploaded photos form part of chats, and users can decide whether OpenAI uses them for training through Data Controls.
