How to Make a GIF With ChatGPT: Steps, Access and Limits

Since November 4, 2026, ChatGPT can make looping animated GIFs. You describe what should move, or upload a still image (a photo or a sketch) to animate; ChatGPT returns a GIF you can speed up or slow down, save and share. The feature ships with ChatGPT Images 2.6 and is still rolling out, so it may not be on your account yet; OpenAI hasn’t published a schedule (ChatGPT release notes, checked Nov. 5, 2026). Images 2.6 itself is available on every plan, including Free, worldwide, on desktop, mobile and web (OpenAI announcement, Nov. 4, 2026, checked Nov. 5, 2026).

No plan is singled out for GIFs: Images 2.6 is on every plan, including Free, subject to the rollout. Paid plans (Go, Plus, Pro) raise image generation limits; current US prices are in our ChatGPT pricing guide. This tutorial walks through the four steps, with copy-ready prompts, then covers the limits. It’s based on OpenAI’s announcement, release notes, the Images in ChatGPT help article (updated on the day of the announcement) and the API guide, linked throughout. Génie Artificiel earns no commission on ChatGPT.

What Does ChatGPT’s GIF Feature Do?

It makes a short animation that loops, in GIF format, from:

• a text prompt: you describe the subject and the motion;
• an uploaded still image: a product shot, a drawing or a character that ChatGPT animates.

OpenAI’s help article also covers what comes next: to change a GIF ChatGPT created, you describe the edit in the conversation; then you open the GIF to adjust its playback speed, save it to your device or share it. OpenAI’s suggested uses include animating a brand mascot for a campaign, reaction GIFs and memes for social media, turning a product shot into a loop, and animated storyboards to explore character movement before a video shoot.

What it doesn’t do:

• Edit an animated GIF you upload: the help article rules that out explicitly.
• Make a video: GIFs carry no sound, and OpenAI has shut down Sora, its video generator: the web and app versions on April 26, 2026, and the API on Sept. 24, 2026 (OpenAI Help Center on the Sora discontinuation, checked Nov. 5, 2026). For actual video from inside ChatGPT, you need a connector: see how to make AI videos with ChatGPT or Claude using MCP.

The same day, OpenAI added saved reference images: you save a photo under a name (your dog, say), then call it up in a later chat by typing @ followed by that name. The help article describes this for images; it doesn’t say whether it works for GIFs.

Who Can Make GIFs in ChatGPT?

So there’s no way to tell whether a GIF counts as one image against your plan’s limit, or uses more. Any GIF quota figure that doesn’t point to an OpenAI page has no official source. For plan details and prices, see our ChatGPT pricing guide.

One special case: on a teen account linked to a parent, the parent can turn off image creation and editing (Images in ChatGPT).

How to Make a GIF With ChatGPT, Step by Step

Based on OpenAI’s Images in ChatGPT help article (checked Nov. 5, 2026):

1. Open a chat on chatgpt.com or in the iOS or Android app. You can also select More, then Images.
2. Ask for a “looping GIF” in so many words and describe what should move. The help article’s example is a paper boat bobbing on the water. OpenAI’s API guide recommends describing three things: the subject, the motion and the loop you want.
3. Wait: generation can take a few minutes depending on how complex the request is, and you can keep using ChatGPT in the meantime.
4. Open the GIF to adjust its playback speed, save it to your device or share it.

To change the result, describe the fix in the same chat (“the motion is too fast,” “keep the same framing from start to finish”).

Copy-Ready Prompts

Three prompts written for this guide, following OpenAI’s advice (subject, motion, loop):

Make a looping GIF: a steaming cup of coffee on a wooden table, the steam rising slowly and curling. Fixed camera, warm morning light. The last frame flows back into the first for a seamless loop. No text.

Make a looping reaction GIF for social media: a small round robot gives a thumbs-up, winks, then returns to its starting pose. Plain light-blue background, soft 3D illustration style, same framing in every frame.

Make a looping GIF of our mascot: an orange fox in flat illustration style waves its right paw, then returns to its starting pose. White background, identical framing in every frame, no text or logo.

The examples in OpenAI’s announcement go further, but they were written by OpenArt, a platform that uses the model, not for ChatGPT. They ask for a still sprite sheet of six same-size frames in a three-column, two-row grid, with a fixed camera, the same character and setting throughout, and a sixth frame that leads back into the first. ChatGPT’s help article doesn’t mention this method: if ChatGPT returns a still grid, ask again explicitly for a looping GIF.

How to Animate a Photo or a Sketch

OpenAI’s help article covers this: upload a still image and ask ChatGPT to animate it. The Nov. 4 release notes give the example of a sketch whose character you make wave, and the announcement shows a hand-drawn owl turned into an animated character.

1. Attach your image to the message (a photo of a paper sketch, a product shot, an illustration).
2. Describe the motion and what should stay still.
3. Refine it in the chat, just like a GIF made from text.

Animate this sketch as a looping GIF: the character waves, then returns to its starting pose. Keep the line work, colors and background of the drawing.

Animate this photo as a looping GIF: only the product turns slowly in place; the background and lighting stay the same. Smooth motion, no jump at the loop point.

On mobile, ChatGPT also lets you draw a sketch inside the app (type @, then choose Sketch) and turn it into an image. The help article doesn’t say whether that sketch can go straight to a GIF; the documented route is uploading a still image.

Photos of people: OpenAI’s usage policies prohibit using someone’s likeness, including a photorealistic image of them, without their consent in a way that could confuse authenticity. Animating a photo of a friend or a stranger without permission may be refused by ChatGPT and can raise right-of-publicity issues under state law.

How to Edit, Slow Down, Save and Share Your GIF

• Edit: describe the change in the chat. This only works for a GIF ChatGPT made; an uploaded animated GIF can’t be edited.
• Speed up or slow down: open the GIF and adjust the playback speed. OpenAI doesn’t publish the available values, or whether the saved file keeps the speed you chose.
• Save: from the open GIF, save it to your device.
• Share: the same view offers sharing.

Images you create in ChatGPT are saved automatically under Images, on web and mobile, and disappear if you delete the chat they came from; the help article doesn’t say whether GIFs follow the same rule. For other image settings (aspect ratio, transparency, editing one area), see our guide to creating images with ChatGPT Images.

Why Is the GIF Option Missing or Not Working?

• The feature hasn’t reached your account yet: OpenAI has been rolling it out gradually since Nov. 4, 2026, with no schedule. Our tip: try again later and update the app.
• Your prompt doesn’t say “GIF”: the help article recommends asking for a “looping GIF”; without that word, you’re likely to get a still image.
• You’re trying to edit an existing GIF: ChatGPT can’t edit an animated GIF you upload. Upload a still frame from it instead and ask for a new animation.
• You’ve hit your image limit: Free has limited image creation, with no published number.
• The request is refused: ChatGPT filters prompts and images under OpenAI’s usage policies (someone’s likeness without consent, non-consensual sexual content or sexualizing minors, infringing other people’s intellectual property). Describe an original character instead.
• Teen account: a parent may have turned off image creation.
• It’s taking a while: a few minutes is normal for a complex request, according to the help article.

How Many Frames and How Long Is a ChatGPT GIF?

For ChatGPT, OpenAI publishes no numbers: no frame count, length, resolution or file size (help article and release notes checked Nov. 5, 2026).

The only official reference comes from the API, which runs the GPT-Image-2.6 models (Image generation guide, checked Nov. 5, 2026):

OpenAI’s own example, 9 frames at 200 ms, returns to its starting pose after 1.8 seconds. Our math: at the default frame duration, 16 frames make a 3.2-second loop; the longest possible loop, 16 frames at 2 seconds each, runs 32 seconds, but at that point it’s a slideshow rather than an animation.

Nothing says ChatGPT uses the same settings; these are the API’s. If ChatGPT works within similar limits (OpenAI hasn’t confirmed it), a GIF has only a few frames, so aim for a simple repeated gesture (a wave, rising steam, a turn) rather than a scene that tells a story.

Usage Rights and AI Labels

Who owns the GIF? Under OpenAI’s Terms of Use (effective Jan. 1, 2026, checked Nov. 5, 2026), you own the output, to the extent permitted by applicable law, and OpenAI assigns you any rights it has in it. No plan is excluded. Two caveats: another user may get a similar result, and the assignment doesn’t protect you against third-party rights (an existing character, logo, brand or face). Our story on AI-generated Studio Ghibli-style images covers that debate. This isn’t legal advice.

Does the GIF carry a watermark? Here’s what OpenAI documents (Help Center on provenance signals, updated in late October 2026, checked Nov. 5, 2026):

• images generated with ChatGPT, Codex or the API carry C2PA metadata (Content Credentials, which record the tool and creation date) and an invisible SynthID watermark;
• coverage “may vary” by product, model, export path and file type;
• openai.com/verify checks “supported” image and audio files.

What isn’t documented: as of Nov. 5, 2026, no OpenAI page says whether an animated GIF carries these signals or whether the verification tool accepts GIFs. A negative result would therefore prove nothing. OpenAI also notes that these signals don’t replace a visible label: depending on where you post (ads, social media), you may need to disclose yourself that an animation is AI-generated.

For Developers: How to Generate a GIF With the API

OpenAI’s API guide (checked Nov. 5, 2026) documents GIF output for both version 2.6 models: gpt-image-2.6-flare, built for speed, and gpt-image-2.6-sunburst, for detailed work.

• Set output_format: "gif" on a POST /v1/images/generations request, along with frame_count and frame_duration_ms (table above).
• The edit endpoint (/v1/images/edits) doesn’t return GIFs.
• GIF output doesn’t support streaming, partial images, output_compression or prompt_rewriter.
• Frame count, frame size and the number of animations requested must fit the model’s pixel budget; if not, lower frame_count, the size or n.
• The response contains the animation as base64 (b64_json); save it with a .gif extension.
• Your organization may need to complete API Organization Verification in the developer console before using GPT Image models.

curl https://api.openai.com/v1/images/generations \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2.6-flare",
    "prompt": "An orange fox in flat illustration style waves its paw, then returns to its starting pose. White background, fixed framing, smooth loop.",
    "output_format": "gif",
    "frame_count": 9,
    "frame_duration_ms": 200,
    "size": "auto"
  }' | jq -r '.data[0].b64_json' | base64 --decode > fox.gif

Pricing: both models cost $8 per million image input tokens and $30 per million image output tokens; text input is $5 per million (API pricing, checked Nov. 5, 2026). The lower cached-input rate probably doesn’t apply to GIFs: for the 2 and 2.5 models, the API guide says it doesn’t apply to direct Image API requests, the only route to a GIF, and it doesn’t restate this for version 2.6. That’s the same as GPT-Image-2.5. OpenAI doesn’t publish a per-GIF price: cost depends on the tokens used, which the response reports under usage. API usage is billed separately from any ChatGPT subscription.

Other Ways to Make an Animated Loop

• Midjourney animates an image (yours or a Midjourney one) into a 5-second video you can extend up to 21 seconds. A Loop checkbox reuses the first frame as the last, and the download menu offers an animated GIF alongside MP4 (Midjourney video docs, checked Nov. 5, 2026). Plans and prices are in our Midjourney guide.
• An AI video generator gives you longer, smoother clips, sometimes with sound, usually as MP4, which you then convert to GIF with an editing or conversion tool. From inside ChatGPT, that means an MCP connector (see above).

FAQ

Can I make a GIF with ChatGPT for free?

In principle, yes: Images 2.6, which brings GIFs, is available on every plan, including Free. But GIF creation has been rolling out gradually since Nov. 4, 2026, and Free has limited image creation with no published number.

Why don’t I see the GIF option in ChatGPT?

Most often because the rollout hasn’t reached your account yet. Also check that your prompt actually says “GIF,” that the app is up to date, and that you aren’t trying to edit an uploaded GIF.

How long is a GIF made with ChatGPT?

OpenAI doesn’t say for ChatGPT. In the API, a GIF has 2 to 16 frames of 20 to 2,000 ms each; the official example, 9 frames at 200 ms, loops in 1.8 seconds.

Can ChatGPT edit a GIF I upload?

No. OpenAI’s help article says ChatGPT can’t edit an animated GIF you upload. It can animate a still image, and edit a GIF it made itself.

Do ChatGPT GIFs have a watermark?

ChatGPT images carry C2PA metadata and an invisible SynthID watermark. For GIFs, OpenAI doesn’t yet say whether those signals are present: coverage varies by file type.

Can I use a ChatGPT GIF commercially?

OpenAI’s terms give you ownership of the output, to the extent permitted by applicable law, on every plan. You’re still responsible for third-party rights: no existing characters, logos or real people without permission.