ChatGPT adds animated GIF creation, OpenAI expands its image-gen cap
The Images 2.6 rollout brings prompt-based animation, frame-level editing and faster image generation to ChatGPT, opening new possibilities for digital publishers, marketers and content creators.
OpenAI is expanding ChatGPT’s visual capabilities with a feature that allows users to create animated GIFs from text prompts or uploaded photographs, marking a further step in the integration of image generation and lightweight animation into everyday AI tools.
The feature, reported by The Verge on 4 November, is being introduced alongside OpenAI’s Images 2.6 model. Users can describe an animation in ordinary language or upload an existing image and ask ChatGPT to turn it into a moving sequence. The resulting GIF can be adjusted for playback speed, examined frame by frame and edited before being shared.
The update changes the role of ChatGPT from a tool primarily associated with generating and editing static images into one that can also produce simple animated content. Previously, creating an animated GIF generally required a separate application or a multi-step workflow involving individual frames and animation software.
OpenAI’s announcement confirms that Images 2.6 is available globally through ChatGPT and Codex across web, desktop and mobile platforms. The company says the model generates images 50 per cent faster than its predecessor, while offering greater control over editing and visual consistency.
The new GIF capability allows users to turn a written instruction into a sequence of images designed to create the impression of movement. OpenAI demonstrated the feature using a request to generate a dancing cat, illustrating how a simple prompt can produce an animation without requiring the user to construct each frame manually.
The ability to inspect and edit individual frames is particularly relevant to practical production. Animation often requires consistency in the appearance of characters, objects and backgrounds across successive images. Frame-level editing gives users a way to refine the result rather than relying entirely on the first generation.
The feature could be useful for social-media publishers, digital newsrooms, marketing teams and educators. News organisations, for example, may use short animations to explain a process, illustrate a market movement or make a data point more accessible. Businesses could produce product loops, campaign visuals and branded characters without commissioning a full animation project.
However, generated GIFs should not be confused with conventional video production. GIFs are generally short, looping visual sequences, and their suitability depends on the complexity of the movement, the intended platform and the need for accuracy.
Beyond animation, OpenAI is positioning Images 2.6 as an improvement in image-generation speed and editing precision. Its announcement says the model generates images 50 per cent faster than before and is designed to preserve important details, including the appearance of characters, products and brand elements across editing rounds.
The company is also introducing two API offerings, GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst. Flare is designed for faster generation, while Sunburst targets higher-quality output and greater consistency when more detailed visuals are required.
Developers receive additional controls for image creation and editing, including image markup through the API and SVG exports. A new complexity parameter allows developers to adjust how much planning the model applies to a task, while image-quality settings control the level of visual detail.
These capabilities are relevant to businesses integrating image generation into websites, applications and creative production systems. Rather than treating every image as an isolated output, developers can incorporate generation and editing into repeatable workflows.
OpenAI also says users now generate more than four billion images each week across ChatGPT and its API, up from three billion reported in September. The figure indicates the scale of demand for AI-assisted visual production, although it does not reveal how many users generate images or how frequently those images are used commercially.
The expansion of image generation into animation could lower production barriers for smaller publishers and independent creators. A newsroom with limited design resources may be able to produce simple animated explainers, while a small business could create promotional material without relying on specialist animation software.
For journalists, the potential benefit is not simply faster production. Short animations can help explain complex information, particularly when a story involves changing figures, sequences of events or processes that are difficult to communicate through a static image.
But speed also introduces editorial risks. AI-generated animations can make fictional or altered scenes appear convincing, particularly when they are presented without clear labels. Newsrooms using the technology will need verification procedures, disclosure standards and human editorial review to ensure that illustrative material is not mistaken for documentary evidence.
There are also practical limitations. A generated animation may require several rounds of correction, and frame-level consistency does not automatically guarantee factual accuracy. For news and public-interest content, visual quality must remain secondary to the reliability of the information being communicated.
OpenAI’s Images 2.6 rollout reflects a wider development in generative AI: tools that once produced isolated text or images are increasingly combining multiple creative functions within a single interface. The significance of the GIF feature will depend on how reliably users can control the output and how effectively it fits into professional workflows.
For digital publishers, the next challenge is therefore not merely producing more visual content. It is using faster and more accessible production tools without weakening accuracy, transparency and editorial responsibility.
Sources
