ChatGPT Adds Animated GIF Creation Using OpenAI's Images 2.6 Model

What happened

OpenAI has added a GIF feature to ChatGPT, letting users create animated GIFs from a text prompt or a photo using its new Images 2.6 model. According to The Decoder, users can also adjust playback speed and edit individual frames once a GIF is generated.

The reporting highlights image consistency as the key capability. The model keeps characters recognisable from one frame to the next, which is what makes flipbook-style animation possible rather than a series of loosely related images.

OpenAI also says users now generate more than four billion images per week, a figure that gives a sense of the scale at which an update like this could reach people.

Why it matters

The significance lies less in the GIF format itself than in what consistency across frames makes possible. Generative image tools have largely been judged on single outputs. Once a model can hold a character's identity steady over a sequence, the same technology starts to support short-form motion, storyboarding and simple animation, with editing controls that let people shape the result.

For leaders in experience, AI and digital transformation, this points to a broader pattern: generative tools are moving from producing a first draft to supporting iteration. Any organisation building on these models will need to plan for multi-step workflows, where users expect to adjust, correct and refine rather than accept a single result.

By the numbers

• More than four billion images generated by users per week, according to OpenAI as reported by The Decoder.

The Renascence take

Most coverage will focus on the novelty of animated GIFs. The more useful question for experience teams is what happens when generated output is almost right.

The headline is not the GIF; it is the editing layer that sits on top of generation. Frame-level and speed controls turn an unpredictable output into something a person can steer, which is where most generative tools still lose users. The behavioural principle is simple: people trust a tool more when they can fix its mistakes in the moment. Operators should design for the edit loop from day one, rather than treating the first generated result as the product.