The useful result
One short looping GIF (or four still frames if the feature is missing) of your own picture, checked against the original and labelled as made with AI.
- Choose a picture that is yours and write down what must not change.
- Ask for one simple motion with a fixed camera, with a safe fallback if the GIF feature is missing.
- Check that the subject and any text stayed the same, then post it with a clear made-with-AI label.
You ask ChatGPT to animate your mascot. The cup grows a second handle and the words on it turn into nonsense. This lesson fixes that with one simple motion and a fixed camera.
Careful: the GIF tool is rolling out slowly and may not be on your account yet. If it is missing, step 2 gives you four still frames instead.
What was reported: on 4 November 2026 The Next Web wrote that OpenAI is rolling out animated GIFs in ChatGPT with a new image model, Images 2.6. Users can turn a written prompt or an uploaded photo into a GIF, change its speed and edit single frames. The Next Web said the update is rolling out to all ChatGPT users on web, desktop and mobile; that wording is about the update as a whole, and the report does not say whether every account gets the GIF tool at once. Decrypt, the same day, said the rollout appears to be gradual, that Images 2.6 had no standalone official announcement yet, and that the feature was not available on its accounts. It also said the model generates a series of frames, but not the animation.
We read those reports on 9 November 2026. We have not run this on a live account, and OpenAI's own announcement page would not open for us. Neither report gives GIF length, limits or plan details.
Use your own picture or an invented mascot. Never use another person's face.
Before you start
- A ChatGPT account and one picture you made or own: a mascot, a logo character, a drawing or a product you sell. No photos of other people.
- About 20 minutes.
- If you do not see the GIF feature yet, that is normal for a gradual rollout. Follow the four-frame fallback in step 2 and try again in a few days.
Optional attachment
Review evidence
Documentation-based review · 2026-11-09
Sources opened and checked on 9 November 2026. Figures are dated and describe what the sources reported, not rules. Prompts and examples are illustrative; no model output or outcome is claimed.
- The Next Web: ChatGPT can now make animated GIFs with Images 2.6 (4 November 2026)
- Decrypt: OpenAI rolls out GIF creation in ChatGPT; feature not available on its accounts, reports the model returns frames (4 November 2026)
- AI Weekly, In the Wild (page undated; week taken from its web address; quotes Unite.AI)
Pick your own picture and list what must not change
Choose a picture you made or own, such as a mascot or drawing. Never use a photo of another person, a child, or a character someone else owns. Attach it to a new chat and ask for a plain description. That is your "must not change" list for step 3.
Look at the attached picture. Describe the main character or object in 5 short lines: shape, colours, face, any accessories, and any words in the picture, spelled exactly. Do not change the picture. Do not guess names or add details you cannot see.Example / expected result
Illustrative example; not a recorded model result.
Imagined example: an invented mascot, a round teal tea-cup called Steamy.
- Round teal cup with a small white handle on the left.
- Two black dot eyes and a small smile.
- A thin line of steam above the lid.
- Plain cream background.
- Words on the cup: "TEA TIME".
Check your result
Yes or no: every line matches your picture, and any words are spelled exactly as printed. If not, fix the list by hand before moving on.
Ask for one simple motion with a fixed camera
Pick one motion that can loop: a wave, a blink, a bounce or a nod. Say the camera does not move. Keep the picture attached. The fallback line asks for four still frames if the GIF feature is missing. Decrypt could not use the feature on its accounts, so this may happen to you.
Using the attached picture, make a looping animated GIF. One simple motion only: [MOTION, for example: the character blinks and gives a small wave]. Fixed camera: no zoom, no pan, no rotation. Keep the character, colours, face and any words exactly as in the picture. Plain background. Use 6 to 8 frames and make the last frame join the first. If you cannot make an animated GIF in this chat, say so, and give me 4 separate still frames of the same motion instead.Example / expected result
Illustrative example; not a recorded model result.
Input: [MOTION] = the cup blinks and gives one small wave. Possible reply if the feature is on: "Here is your looping GIF. Steamy blinks, waves once, and returns to the start pose." Possible reply if it is not: "I cannot make an animated GIF here. These are 4 still frames: eyes open, eyes closed, wave up, back to start."
Check your result
Yes or no: you see one motion only, the camera never moves, and the last frame matches the first. If you got four stills, that counts as a pass for now.
Check the subject and the text, then label it
Ask ChatGPT to compare the result with your list, frame by frame. Then look yourself, because the AI can miss its own mistakes. Fix one frame at a time (see the lesson on editing one thing). Post it with "Made with AI" in the caption. Selling it or using it for a business? Read the rights lesson first.
Compare this result with my must-not-change list: [PASTE YOUR 5 LINES FROM STEP 1]. Go frame by frame. In a table, list any change in shape, colour, face, accessories or words. If you find none, write "no change found". Do not fix anything yet.Example / expected result
Illustrative example; not a recorded model result.
Frame 1: no change found. Frame 2: no change found. Frame 3: the words now read "TEA TIEM". Spelling changed. Frame 4: the handle is on the right. Position changed. Caption to post: "Steamy the tea-cup says hi. Made with AI."
Check your result
Yes or no: your table says "no change found" for every frame, you looked at each frame yourself, and the caption says "Made with AI".
When it goes wrong
If you do not see the feature yet, say: "Make 4 separate still frames of the same single motion, same character, same background." Try again in a few days. If the camera moves, say: "Redo it with a completely fixed camera and the character in the same place in every frame." If words turn into nonsense, say: "Keep the words exactly as in my picture, or remove them." You can add real words later in an editor. If the character changes between frames, ask for a new version. Do not post it as real footage.
Make it your own
Do this today in 15 minutes: make one GIF, or four stills, of a mascot you own with the step 2 prompt. Count the frames that kept the same shape, colours and words, and write the score. Then post it with "Made with AI" in the caption.
Useful next resources
Sources and review
- The Next Web: ChatGPT can now make animated GIFs with Images 2.6 (4 November 2026)
- Decrypt: OpenAI rolls out GIF creation in ChatGPT; feature not available on its accounts, reports the model returns frames (4 November 2026)
- AI Weekly, In the Wild (page undated; week taken from its web address; quotes Unite.AI)
Last checked Nov 9, 2026. Features and account limits can change. Verify important results against your original material.
Did this help you complete your task?
Tell us what needs improving. No account or private output required.
Report incorrect information or a broken linkGenius Forges Editorial
Practical exercises based on checked sources. Examples are illustrations, not recorded model results.
