OpenAI launched Images 2.6 on November 4. The new image model is now available in ChatGPT, the Codex coding tool, and through the API. According to OpenAI, it works worldwide on all subscription plans for desktop, mobile, and web. Images now generate 50% faster than before. Edits are also more consistent, so clothing, faces, and product details remain unchanged when you adjust the background or another part of the image. There are also two new features that go beyond still images.
ChatGPT now lets you create animated GIFs. You can describe what should move or upload a still image and ask for an animation. The OpenAI help article on ChatGPT Images says you can adjust the playback speed afterwards, and save or share the GIF.
You request changes in the chat. ChatGPT cannot edit an animated GIF that you upload. As of November 7, the help article still said GIF creation was rolling out, so it might not be available in every account yet. On November 8, the ChatGPT Images page in Germany already showed an anime reaction GIF template without signing in. OpenAI has not shared any GIF limits for any plan.
The second new feature is about personal photos. You can upload a picture, give it a name, and use it in later chats by mentioning it with @. For example, OpenAI shows a dog photo saved as Pepper.
Later, you can ask for a watercolor portrait of @Pepper without uploading the photo again. Sometimes, ChatGPT may suggest saving a photo when you describe it. Once saved, the photo stays in your account. OpenAI’s help page does not explain where to manage or delete saved references. Generated images are handled differently: they are saved under Images and are only deleted when you remove the conversation that created them.
Saving photos of real people brings up another concern. Trending Topics notes that combining a saved reference photo with animation can make it easier to put people in scenes that never happened. OpenAI has not explained how it prevents misuse of GIFs showing real people.
Its help article on provenance signals says images made with ChatGPT include both Content Credentials and an invisible SynthID watermark. GIFs are not mentioned separately. Anthropic now marks AI output in a similar way, as our report on watermarks in Claude text explains.
Photos of children or other people should only be saved if they agree. Uploaded photos are part of your chats, and you can control whether OpenAI uses them for training in Data Controls, as explained in our guide to opting out of AI training. Our AI subscription comparison shows the prices for ChatGPT, Gemini, and Claude plans.
