ChatGPT can now generate animated GIFs from text or photos and let users edit playback speed and individual frames, The Verge reported on November 4, 2026. According to the outlet, the feature is rolling out alongside a new image model, "Images 2.6." A GIF is an image format that conveys motion by displaying multiple still images in sequence. With the new feature, users can change the playback speed of a generated GIF or fix specific frames. The main change is that creating and editing GIFs, which previously required separate tools, can now be handled within ChatGPT.
GIF generation from text or photos, with speed and frame-level editing
According to The Verge, users can generate a GIF by describing what they want in text or by uploading a photo. After generation, they can adjust playback speed, edit individual frames, and share the finished GIF with others. The feature is notable because it lets users adjust the tempo of the motion and the content of each frame after the fact, rather than simply producing a moving image.
In an example OpenAI presented, asking ChatGPT to "make a GIF of a dancing cat" produces a GIF whose playback speed can be changed and whose frames can be viewed one by one. However, this is a demo shown by OpenAI, not a result verified by a third party. For instance, someone making a short reaction GIF for social media could adjust the tempo of the motion after generation or correct only the frames that look unnatural.
The Verge explains that earlier versions of ChatGPT did not offer a standard feature for generating GIFs directly. That does not mean GIFs were impossible before: it was possible to combine multiple images into a GIF using methods such as code execution. The difference this time is that everything from generation to finishing is integrated into ChatGPT's standard workflow.
The outlet says Images 2.6 is rolling out to all users on ChatGPT's web, desktop and mobile versions, as well as in Codex. However, this refers to the rollout of the image model itself; it has not been confirmed that the GIF generation and editing features are available in every environment at the same time. OpenAI says Images 2.6 has also made ChatGPT's image generation faster, but it has not disclosed specific processing times or the rate of improvement.
Character consistency is key to GIF quality
When creating an animated GIF, it is important to keep a subject's appearance consistent across successive frames. For example, in a GIF of a dancing character, features such as the face and clothing need to stay the same even as poses and expressions change. If the character's look shifts unintentionally from frame to frame, the motion appears unnatural. The ability to generate high-quality individual still images is not necessarily the same as the ability to connect them naturally as continuous motion.
The Decoder, discussing the new feature, suggests that consistency in keeping the same character across images, with slightly different character images strung together like a flipbook, may have improved to a level usable for making GIFs. However, this is the outlet's interpretation; it has not confirmed how Images 2.6 actually generates GIFs. It is also unclear at this point whether the model generates each frame with temporal continuity in mind or whether a separate process stitches still images together.
Image consistency is also an issue OpenAI has been working to improve. When announcing Images 2.5 on September 8, 2026, the company said it had improved the ability to preserve the features of a subject in a reference photo, as well as to maintain instruction fidelity and image quality across repeated edits. Such improvements matter when changing part of an image based on a photo while keeping everything else intact.
OpenAI also said that Images 2.5 reduced image generation latency by up to 50% compared with Images 2.0. However, that figure is the company's claim about Images 2.5 and does not indicate the performance of Images 2.6. The official API guide also notes, as a limitation of GPT Image models in general, that it can be difficult to keep the appearance of the same character or brand elements consistent across multiple generations. This is not a test result for GIF generation with Images 2.6, but it suggests consistency between frames should not be assumed to be guaranteed.
GIF generation in ChatGPT and API support are separate questions
In evaluating the new feature, it is necessary to distinguish between being able to create GIFs in ChatGPT and developers being able to generate GIF files directly through the API. In addition to the GIF generation and editing features in ChatGPT, The Verge reports that "GPT-Image-2.6 Flare" and "Sunburst" have been introduced for the API. However, OpenAI's official image generation API guide, retrieved on November 4, 2026, describes the Images 2.5 generation of models and output formats, and is not a document that confirms Images 2.6 specifications.
What can be confirmed at this point, and its limits, is summarized below.
| Item | Public information | What can be determined now |
|---|---|---|
| ChatGPT GIF feature | The Verge reports GIF generation, playback speed changes and frame editing | Creation and editing in ChatGPT has been reported |
| API models | The Verge reports the introduction of 2.6 Flare and Sunburst | Official API model identifiers and conditions for direct GIF output are unconfirmed |
| API output formats | The official guide lists PNG, JPEG and WebP for the 2.5 generation | Cannot determine whether Images 2.6 can output GIFs directly |
The "Output format" section of the official API guide says that, in addition to the default PNG, JPEG and WebP can be specified. GIF is not included, but because it cannot be confirmed that the guide reflects Images 2.6 specifications, the absence of GIF cannot be taken as proof that it is unsupported. Conversely, a report that ChatGPT can generate GIFs does not by itself mean GIF files can be obtained directly from the API. Developers building this into their own apps will need to separately check the official identifier of the target model and the output formats it supports.
Over 4 billion images a week, and limits still to be confirmed
According to The Verge, OpenAI says the number of images generated across ChatGPT and the API has exceeded 4 billion per week. When announcing Images 2.5 in September 2026, the company said images generated by ChatGPT Images and the API's GPT-Image models exceeded 3 billion per week. However, both are lower bounds rather than exact figures, so an accurate growth rate cannot be calculated from this information alone.
These figures also represent the usage scale of image generation overall; they do not indicate the number of GIFs generated or GIF users. Nor can the number of images per user or the degree to which the GIF feature is used be inferred, so growth in image generation as a whole should be considered separately from demand for GIF creation.
For practical GIF creation, the length of animation that can be generated, the number of frames, and the resolution are important. However, the materials reviewed do not specify limits such as maximum playback duration, maximum frame count, resolution or usage quotas by plan. API pricing for Images 2.6 has not been confirmed either.
Being able to handle GIF generation, editing and sharing within ChatGPT should make it more convenient for users who create short animations and social media content. At the same time, judging its practicality will require checking consistency between frames and editing precision, as well as generation conditions and usage limits. For app developers in particular, whether GIFs can be output directly through the API, and how that is priced, will be key questions to watch.