What's changed, and why does it matter?
On September 8, 2026, OpenAI introduced ChatGPT Images 2.5, a model designed to make image creation and editing more precise, realistic, and consistent.
The biggest improvements were in the details.
According to OpenAI, Images 2.5 produces more natural lighting, richer textures, and better preserves the appearance of subjects in reference photos. It also follows editing instructions more reliably, even when users make multiple changes to the same image.
In other words, you could make adjustments without constantly worrying about the AI changing details you wanted to keep.
For professionals, this meant less time fixing unwanted changes and more time refining the visuals they actually needed.
But OpenAI wasn't finished.
On November 4, 2026, the company introduced its latest upgrade: ChatGPT Images 2.6.
And this time, the improvements go beyond making images look better.
So, what's new with Images 2.6?
One of the biggest additions is something Images 2.5 couldn't natively do in the same way: create animated GIFs directly from a prompt or an existing image.
Think about what this means for your everyday creative work.
You can turn a static product photo into an animated visual, bring a brand mascot to life, create reaction GIFs for social media, or animate a character without having to build an animation from scratch.
You can even adjust playback speed and edit individual frames to get closer to the result you want.
But that's not the only improvement.
Images 2.6 also brings a major speed improvement.
OpenAI says Images 2.6 generates images 50% faster than before. That means less waiting between ideas and more opportunities to experiment with different concepts, compositions, and creative directions.
For someone working on marketing campaigns, website visuals, or product presentations, that extra speed can make a noticeable difference.
Instead of spending more time waiting for an image to finish generating, you can move through creative revisions more quickly.
The biggest distinction?
Image 2.5 made image creation more reliable. Image 2.6 makes the creative process more flexible.
People have been comparing the two models:
"GPT Images 2.6 Flare generates images almost twice as fast as GPT Images 2.5 Flare." — @ArtificialAnlys on X
"GPT Images 2.5 vs 2.6: We ran the same 9 edits in a row on both, in Vestra. Which one did better?" — @vestra_ai on X
The real question is no longer just how good your AI-generated images can look. It's what you can create with them next.