ChatGPT Images 2.6 adds GIFs and saved reference images
OpenAI released Images 2.6 on November 4. The new image model is in ChatGPT, Codex and the API, on every subscription tier, on desktop, mobile and the web. OpenAI says it makes images 50% faster than before and keeps details such as a character's outfit, a product or brand elements the same across rounds of edits.
Two new features come with it in ChatGPT. You can now ask for a looping GIF, or upload a sketch and have ChatGPT animate it, then change the playback speed and share the result. GIF creation is still rolling out. You can also save a reference image and use it again in later chats by name. OpenAI's example is a photo of a dog saved as Pepper, then called with @Pepper, so you do not upload the photo again.
For developers, the API has two models, GPT-Image-2.6 Flare for speed and GPT-Image-2.6 Sunburst for detail. The API also gets image markup and SVG export.
What it means for your work
- Save your brand assets once: Store your logo, product shots or mascot as reference images and call them by name in new chats.
- Try GIFs for small motion: A short loop of a product or mascot can work in a newsletter, a slide or a social post where video is too heavy.
- Check every edit: Look at text and details after each round before you publish.
Images 2.6 is also in Codex, which will share one allowance with ChatGPT from December; see our earlier report.