ChatGPT can now produce a finished, looping GIF from a written request or a photo you upload, and it lets you change the playback speed and edit individual frames. The change comes with OpenAI's new Images 2.6 model and is rolling out to all ChatGPT users on web, desktop and mobile.
It is not the first chatbot to do this. Google's Gemini already offered GIF creation a year ago. What is new is that ChatGPT, which has far more people using its image tools, now does it in one step and gives you controls to fix the result.
A GIF is a flipbook, so good image editing carries over
The old ChatGPT route shows what a GIF really is. You asked for a series of pictures that change a little each time, saved them in order, uploaded them together, and asked for them to be joined into one file. That is a flipbook. Ten pictures drawn one at a time have no reason to agree with each other, so you did the stitching and the checking yourself.
When the same tool makes the whole flipbook, it can also open it. Slowing the loop down means changing how long each picture stays on screen. Editing one frame means asking for a change to one picture, the way you would ask to remove a stranger from a photo. The image skills people already use every day are now applied to a short run of pictures.
Why a GIF and not a video
A GIF plays by itself, repeats forever and needs no player. It works in most chat apps, on web pages and in help articles, and in most email programs. A video needs hosting, a play button and often a click. For the small jobs where someone only needs to see one thing move, the GIF is the lighter tool, and that is why it is still everywhere.
Those small jobs are most of what a business actually needs. A support team answers the same question about resetting a router every week, and a three-second loop of the right button pulsing does the work of a paragraph. A café puts steam over the cup in its menu photo. A shop adds a sparkle to the sale badge on a product page. A school office animates the arrow on a map to the right entrance. Each of these used to be a task for a designer or someone patient with an animation app. My estimate is that an hour in an animation app, or a small freelance job, becomes a few minutes of asking and checking.
What to watch as it spreads
People are making images with ChatGPT at a rate that keeps climbing: OpenAI now counts more than 4 billion a week, up from 3 billion in September. That is a third more in about two months. Adding motion extends a habit people already have instead of asking them to learn a new one, so expect loops to show up in the places that used to hold a still picture: email signatures, product listings, menus, internal guides.
When everyone can add motion, motion stops being a way to stand out. What will separate one shop's loop from another's is whether it shows the right thing: the correct button, the real product, the price that is actually current. The frame view exists for that check. Step through every frame before you send a loop to customers, and look hardest at anything with lettering in it.