OpenAI adds GIF creation and reference tools to ChatGPT Images 2.6

Bringing motion to ChatGPT

OpenAI has introduced Images 2.6, a new update that allows users to generate animated GIFs directly within the ChatGPT interface. This follows the company's decision to shut down its Sora video app earlier this year. While full video generation remains a separate space, GIFs offer a way to bring motion back to the platform in a more accessible format.

The new tool works by taking a user's input, such as a text description, a photograph, or even a rough sketch, and transforming it into a short animation. OpenAI demonstrated this capability by creating a dancing cat to show how the technology handles movement. These animations are designed to play on a loop and can be shared immediately with other users.

According to the company, these animations serve as a playful way for both individuals and businesses to communicate. They can be used for personal reactions or as brand mascots and product visuals. The update is currently rolling out to all ChatGPT and Codex users across web, desktop, and mobile platforms.

Customization and editing tools

The Images 2.6 update provides several specific controls for refining the output of these animations. Users can adjust the pace of the animation after it has been created to ensure the movement feels correct. If a specific part of the animation does not look right, the tool allows for the reworking of individual frames.

In addition to the GIF features, OpenAI added image markup to the suite of tools. This allows users to highlight and mark the exact spots on an image where they want to see specific changes made. This level of precision is intended to give users more control over the final aesthetic of their generated content.

The update also includes an SVG export option, which is a scalable vector graphics format. This is particularly useful for creators who need to move their generated images into websites, apps, or print materials. By providing these professional-grade export options, OpenAI is targeting a broader range of creative use cases beyond simple chat interactions.

Consistency for brands and devs

One of the most significant additions for professional users is a new library for reference images. This allows people to keep photos of themselves, their pets, or specific characters on hand to reuse in new creations. By using these references, users can ensure that characters and products look the same across multiple different images.

This functionality is specifically designed to help companies that need to produce brand assets in bulk. It solves a common problem in AI generation where a character might change appearance slightly from one prompt to the next. By keeping the core details consistent, businesses can create a more cohesive visual identity for their marketing efforts.

For developers using the API, OpenAI released two new models designed for speed and quality. The Flare model is built to generate images in as little as six seconds, while the Sunburst model focuses on higher-quality generation and consistency across various edits. These models are reported to be 50 percent faster than previous versions.

Market position and competition

The introduction of GIFs places OpenAI in direct competition with established platforms like Giphy and Google's Tenor. While those platforms typically host a library of ready-made animations, ChatGPT allows users to create the specific reaction they need on demand. This shift toward on-demand generation represents a different approach to how people consume and share moving imagery.

OpenAI is navigating a crowded landscape of AI image and video tools. Google is currently pushing image editing in Gemini using its Nano Banana models, and Meta has integrated its Muse Image model into Instagram. Meanwhile, Midjourney has begun offering short videos alongside its standard image outputs.

The company is also balancing these new capabilities against concerns regarding deepfakes. Because the new tools make it easier to animate photos of real people, some advocates have called for clearer safeguards. OpenAI has not yet specified how it will handle watermarks or specific protections for animated GIFs of real individuals.