OpenAI released ChatGPT Images 2.6 on November 4, two months after GPT Image 2.5. According to the ChatGPT release notes, the new model creates images faster and is better at preserving details across edits, for example keeping a character's outfit the same when you change the background.
Two new features in ChatGPT
GIFs. GIF creation is rolling out in ChatGPT. You can ask for a looping animation or upload a sketch and ask for the character to wave, then adjust the playback speed and share the result.
Saved reference images. A photo can be saved once and reused in later chats. OpenAI's example: save a photo of your dog as Pepper, then ask for a watercolor portrait of @Pepper without uploading the photo again.
For developers
In the API, the model comes in the same two variants as before: gpt-image-2.6-flare for fast everyday generation and gpt-image-2.6-sunburst for the most demanding generation and editing. The changelog lists SVG and animated GIF output, adjustable complexity and markup-guided editing. Prices are unchanged from 2.5: $8 per million image input tokens and $30 per million image output tokens.
What this means if you make images with AI
Consistency is the main gain. Series of images with the same character, product or pet were the weak spot of most generators; saved references and better detail retention target exactly that.
GIFs remove a step. Simple looping animations for chats and social posts no longer need a separate video tool. For longer clips, a video generator is still the right choice. Compare options on our ChatGPT Images page, in ChatGPT Images vs Adobe Firefly and in the list of AI image generators.