OpenAI Launches ChatGPT Images 2.6 with GIF Creation and Saved Reference Images

OpenAI Upgrades Its Image Tools with Animation and Faster Generation

A new chapter in AI image creation opened on 4 November 2026, when OpenAI announced ChatGPT Images 2.6 — an update that pushes its image tools beyond still pictures and into animated GIFs for the first time.

The announcement came via OpenAI’s official account on X, where the company posted: “Images, upgraded—introducing ChatGPT Images 2.6.” A follow-up line made the headline features plain: “Now with GIFs and reference images.”

The update applies across ChatGPT, Codex and OpenAI’s API. And it’s not a small tweak — it introduces GIF creation, a saved reference image system, and two brand-new API models, all at what OpenAI claims is a meaningfully faster generation speed.

What’s Actually New in Images 2.6

The GIF feature is the most visible addition. Users can now generate looping animated GIFs from a text prompt, an existing photograph, or even a hand-drawn sketch. Once a GIF is created, you can adjust its playback speed, edit individual frames, and share it directly. That puts a basic animation workflow inside a tool that, until now, only produced still images.

Meanwhile, the saved reference image system is a quieter but potentially very useful change. You can store photographs of people, pets or recurring characters inside the platform, then call on them in later requests without having to upload the same image again. The idea is consistency — if you’re creating a series of images featuring the same character or person, the model can refer back to that stored reference rather than starting from scratch each time.

Both features are aimed at users who do repeated creative work. Think social media managers building a visual brand, or developers generating character art across a game project.

Two New API Models: Flare and Sunburst

Alongside the consumer-facing changes, OpenAI is introducing two new models for developers working through the API. GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst serve different purposes.

Flare is the speed option. OpenAI says it can generate an image in as little as six seconds when quality is set to low — which, in a production pipeline, could make a real difference. Sunburst, by contrast, is positioned for more professional visual output, trading some of that raw speed for higher-quality results.

OpenAI posted: “Plus, two new models are coming to the API with 50% faster image generation.” That 50% figure is the company’s own estimate, and it’s worth being clear — no independent audit of that claim has been published. Real-world performance will vary depending on quality settings, prompt complexity and server load.

Speed Claims and What to Make of Them

The 50% speed improvement is a headline number, and it’s a company claim. That doesn’t make it wrong, but it does mean you should treat it as a benchmark figure rather than a guaranteed experience.

What is confirmed is the six-second floor for Flare at low quality. For developers building tools where fast turnaround matters — think rapid prototyping or high-volume content generation — that’s a concrete number to test against.

OpenAI’s announcement was published on 4 November 2026 and updated the following day. Availability, rollout timing and pricing may vary depending on which product you’re using and what subscription tier you’re on. No GBP pricing was confirmed in the available sources, so users and developers in the UK should check OpenAI’s pricing pages directly before building costs into any project.

The Bigger Picture

This update marks a clear shift in how OpenAI is positioning its image tools. Still images were already part of ChatGPT’s toolkit, but GIF creation moves the platform into territory previously occupied by dedicated animation software and specialist AI tools.

The saved reference feature addresses one of the more persistent frustrations with AI image generation — the difficulty of keeping a character or subject looking consistent across multiple images. By letting users store and reuse references, OpenAI is making a direct pitch to anyone doing serial or ongoing visual work.

Whether that’s enough to pull users away from established tools is uncertain. But the combination of speed, animation and reference consistency in a single platform is a meaningful step forward in what these tools can do day-to-day.

What This Means for Kent Residents

For anyone in Kent using AI tools for work or creative projects — whether that’s a small business putting together social media content, a school creating educational materials, or a local organisation producing promotional graphics — Images 2.6 adds capabilities that were previously out of reach without specialist software. The GIF creation and saved reference features could reduce the time spent on repeated visual tasks, though access will depend on your subscription tier and API costs. Before uploading photographs of identifiable people or third-party artwork, users in Kent — as anywhere in the UK — should check their obligations under data protection law and consider copyright implications.