OpenAI Images 2.6 Turns Pictures Into Looping GIFs and Generates 50% Faster

OpenAI has released Images 2.6, and the headline feature is one artists and fan creators will notice right away: still images can now become looping GIFs. The update went live on November 4 in ChatGPT and Codex for every subscription tier, and as two new models in the API.

What changes

• Motion. You can animate a mascot or character, or turn a product shot into a short loop. OpenAI also suggests using it for animated storyboards to try out character movement before real video production.
• Speed. OpenAI says generation is 50% faster than before. In the API, the faster model, GPT-Image-2.6 Flare, can return a low-quality image in about six seconds.
• Consistency. The second API model, GPT-Image-2.6 Sunburst, is aimed at detail and at keeping a character, product or brand looking the same across rounds of edits.
• Editing tools for developers. Image markup, already in ChatGPT, comes to the API so you can point at the exact area to change. There are SVG exports, plus a new complexity setting that controls how much planning the model does on a request.

OpenAI also says people now create about 4 billion images a week across ChatGPT, Codex and the API.

Why I'm watching this one

Character consistency has always been the weak spot of image generators, and it is exactly what anyone making comics, sprites or animated reactions cares about. A tool that keeps a character on-model across edits and then animates it is aimed straight at the kind of work fan artists and indie game makers do by hand. That makes the questions around training data and credit for human artists feel more pressing, not less, and OpenAI's announcement doesn't address them beyond pointing to its system card for safety details.

The full announcement, with sample GIFs, is on the OpenAI site.