OpenAI adds GIF creation and faster image generation with Images 2.6

OpenAI has released Images 2.6, an image generation update for ChatGPT, Codex, and its API that adds GIF creation, faster generation, and more controls for editing images. The company says the model produces images 50 percent faster than its previous version and is available globally across ChatGPT subscription tiers.

The most visible addition is native GIF creation. Users can ask ChatGPT to animate an idea from a text prompt or turn an uploaded photo into a short animation. According to Emma Roth’s report for The Verge, ChatGPT also lets users adjust a GIF’s playback speed and edit individual frames after generation.

That workflow could make basic animation more accessible for social media teams, marketers, and other content creators. Rather than creating a sequence of still images and assembling it in separate software, users can produce and revise an animated asset within ChatGPT. OpenAI shows examples involving animated mascots, product imagery, reactions, and storyboard-style character movement.

Faster generation and more consistent edits

OpenAI says Images 2.6 is designed to improve the speed of image generation without losing control over details. The company reports that people create more than 4 billion images each week across ChatGPT, Codex, and its API.

For developers, OpenAI offers two API model options: GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst. Flare is intended for speed. OpenAI says it can create an image in as little as six seconds when developers select low quality. The company also says the model improves layouts for work such as posters and product images.

Sunburst targets projects that need more detailed output and stronger consistency across multiple editing rounds. That is particularly relevant when a team needs to preserve the appearance of a character, product, or visual brand element while changing other parts of an image.

The update introduces a complexity setting in the API. This controls how much planning the model applies to a request. OpenAI recommends a low setting for most edits, where speed matters, and medium or high settings for scenes with many objects, instructions, or visual constraints. A separate image quality setting controls detail and visual fidelity.

Editing tools for production workflows

Images 2.6 also brings image markup to the API. This allows developers to identify a specific region of an image and request a change there, rather than describing the location only in text. The feature was already available in ChatGPT.

OpenAI is also adding SVG exports, which can help teams move suitable designs into websites, apps, or print workflows. SVG is a scalable vector format, meaning graphics can often be resized without the loss of sharpness associated with conventional pixel-based images.

OpenAI presents the release as a tool for faster creative iteration rather than a replacement for traditional animation or design software. The practical results will depend on whether the new consistency controls work reliably on complex brand assets and whether GIF editing can retain a coherent look across every frame.

Images 2.6 is available in ChatGPT and Codex on desktop, mobile, and web. The Flare and Sunburst variants are available through the API, with pricing listed separately by OpenAI.

Sources

• Introducing Images 2.6 – OpenAI
• ChatGPT will now make you animated GIFs – The Verge