ChatGPT Images 2.5 is out — I’ve been testing it for 24 hours, and these are the 3 new features you’ll actually use

It’s been a hot minute since ChatGPT last updated its Images feature .

<![CDATA[ <article> <p>It’s been a hot minute since ChatGPT last updated its <a href="https://www.techradar.com/ai-platforms-assistants/i-compared-chatgpt-images-2-0-and-googles-nano-banana-2-using-real-world-prompts-from-portraits-to-product-shots-and-the-ai-image-generator-that-came-out-on-top-genuinely-surprised-me">Images feature</a>. Its last update, to Images 2.0 in April 2026, was pretty solid and I’d say put it ahead of Google’s <a href="https://www.techradar.com/ai-platforms-assistants/i-tried-nano-banana-2-lite-googles-new-4-second-ai-image-generator-and-it-changes-how-you-use-ai-art">Nano Banana 2</a>, even if its images still took longer to create, which was a bit frustrating. </p><p>Today's new Images 2.5 update adds a whole bunch of extra features and tools to the existing image generation engine that make it a lot easier to use, and it's quicker too.</p><p>In fact, there are so many features now in Images that to say it’s overcrowded would be an understatement. Because of that, it can be a little hard to locate the best new features amongst everything else, so I’ve pulled out the ones you should definitely look out for, because they make a real difference. </p><h2 id="1-much-improved-editing-tools">1. Much improved editing tools</h2><figure class="van-image-figure inline-layout" data-bordeaux-image-check ><div class='image-full-width-wrapper'><div class='image-widthsetter' style="max-width:2560px;"><p class="vanilla-image-block" style="padding-top:56.25%;"><img id="Pskwomx5BDR3LY8kVBzoph" name="ios-portrait-single-mockup-blue-medium (4)" alt="New editing tools for ChatGPT Images 2.5." src="https://cdn.mos.cms.futurecdn.net/Pskwomx5BDR3LY8kVBzoph.png" mos="" align="middle" fullscreen="" width="2560" height="1440" attribution="" endorsement="" class="inline"></p></div></div><figcaption itemprop="caption description" class=" inline-layout"><span class="credit" itemprop="copyrightHolder">(Image credit: OpenAI)</span></figcaption></figure><p>I’ve pulled these features out first because, despite not sounding that exciting, the improved editing tools are actually a real game-changer for how I use ChatGPT Images, and probably will be for you, too. </p><p>OpenAI appears to have quietly been rolling out its new image-editing toolbar in ChatGPT during August, although the company hasn’t formally announced it, so it might not be totally accurate to call them part of the new ChatGPT Images 2.5, but I’m including them because they are, effectively, new.</p><p>At the moment, we’re used to generating an image, then when it’s not exactly what we want, we type in another description using the prompting tool, like ‘make it brighter’ or ‘remove the people in the background’, and try again. After ChatGPT hasn't got quite right a number of times, I usually give up and start again.</p><p>Now I don’t need to. The new editing tools let you pinpoint the exact spot in the image you’d like to change, and tell ChatGPT exactly what you’d like to do with it.</p><p>The new <strong>Edit</strong> toolbar contains <strong>Markup</strong>, <strong>Comment</strong>, <strong>Remove BG</strong>, <strong>Erase</strong> and <strong>Resize</strong>. </p><p>In the web version of ChatGPT, just click on the new <strong>Edit</strong> button that appears over a generated image and you'll enter the Edit mode, where you’ll see this new toolbar. In the app version of ChatGPT, just tap on any image to enter Edit mode.</p><p>Let’s start with the most obvious tool: <strong>Remove BG</strong>. This keeps the foreground elements of your image and removes the background, making a transparent PNG file when you download it. It’s perfect for dropping your image into another document, such as a presentation you’re making.</p><p><strong>Resize</strong> is a quick way of changing the aspect ratio of the image and contains all the usual suspects, including Portrait 3:4, Widescreen 16:9 and Landscape 4:3.</p><p><strong>Erase</strong> makes it easy to remove elements from your image. To remove a person from the background, for example, just brush over them with the <strong>Erase</strong> brush and ChatGPT removes them entirely. You don’t need to be super-accurate with your brush — just cover enough of the object so that it’s obvious what you want to remove.</p><p><strong>Markup</strong> is for adding markup elements to an image, but the <strong>Comments</strong> feature is where it gets really interesting. Instead of doing what you’d expect — just adding comments to an image — these are actually points for ChatGPT to act on. Click on any element of an image and add a comment like ‘remove this’ or ‘change this to blue’, and ChatGPT will follow your instructions when you tap the <strong>Send</strong> button. This makes editing images so much easier.</p><p>The final feature worth noting with the Edit mode is that you can easily roll back to every previous version of your image — each edit is saved as a separate file, so just scroll up or down in Edit mode to access previous versions of your image.</p><div style="min-height: 250px;"> <div class="kwizly-quiz kwizly-exgklX"></div> </div> <script src="https://kwizly.com/embed/exgklX.js" async></script><h2 id="2-sketching-out-an-image-first">2. Sketching out an image first</h2><figure class="van-image-figure inline-layout" data-bordeaux-image-check ><div class='image-full-width-wrapper'><div class='image-widthsetter' style="max-width:2560px;"><p class="vanilla-image-block" style="padding-top:56.25%;"><img id="92cWt6k8eevmXchfZqXbJ6" name="ios-portrait-dual-mockup-blue-medium (1)" alt="The Sketch feature in Images 2.5" src="https://cdn.mos.cms.futurecdn.net/92cWt6k8eevmXchfZqXbJ6.png" mos="" align="middle" fullscreen="" width="2560" height="1440" attribution="" endorsement="" class="inline"></p></div></div><figcaption itemprop="caption description" class=" inline-layout"><span class="caption-text">Scroll down to see what my sign sketch on the right turned into… </span><span class="credit" itemprop="copyrightHolder">(Image credit: OpenAI)</span></figcaption></figure><p>The new <strong>Sketch</strong> tool enables you to sketch out an image, then ChatGPT transforms your rough sketch into a proper AI-generated image. It's useful when it would be quicker to draw your idea out than describe it. </p><p>Choose <strong>Images</strong> from the main menu and then <strong>Sketch</strong> at the top of the available options. </p><p>The Sketch tool does look a little bit like MS Paint from 1995, but it’s good enough for doing a quick sketch with your finger, or mouse. Once you’ve done that click or tap the blue tick button and ChatGPT loads in a prompt telling it to convert the sketch into a full image. You can add any details you want here.</p><p>Of course, once you’ve got your finished image you can still enter Edit mode to tweak it further.</p><figure class="van-image-figure inline-layout" data-bordeaux-image-check ><div class='image-full-width-wrapper'><div class='image-widthsetter' style="max-width:2560px;"><p class="vanilla-image-block" style="padding-top:56.25%;"><img id="F6pdMHyeSJ3JkgukvCm68Q" name="ios-portrait-single-mockup-blue-medium (3)" alt="Using the Sketch tool in ChatGPT Images 2.5" src="https://cdn.mos.cms.futurecdn.net/F6pdMHyeSJ3JkgukvCm68Q.png" mos="" align="middle" fullscreen="" width="2560" height="1440" attribution="" endorsement="" class="inline"></p></div></div><figcaption itemprop="caption description" class=" inline-layout"><span class="caption-text">Here's the finished image </span><span class="credit" itemprop="copyrightHolder">(Image credit: OpenAI)</span></figcaption></figure><h2 id="3-templates-have-finally-arrived">3. Templates have finally arrived</h2><figure class="van-image-figure inline-layout" data-bordeaux-image-check ><div class='image-full-width-wrapper'><div class='image-widthsetter' style="max-width:2560px;"><p class="vanilla-image-block" style="padding-top:56.25%;"><img id="9KnAifKsDDcvdYB5qX2SdJ" name="ios-portrait-dual-mockup-blue-medium (2)" alt="New Templates in ChatGPT Images 2.5" src="https://cdn.mos.cms.futurecdn.net/9KnAifKsDDcvdYB5qX2SdJ.png" mos="" align="middle" fullscreen="" width="2560" height="1440" attribution="" endorsement="" class="inline"></p></div></div><figcaption itemprop="caption description" class=" inline-layout"><span class="credit" itemprop="copyrightHolder">(Image credit: OpenAI)</span></figcaption></figure><p>The new Templates make it a lot easier to create the sort of things people actually want to do with ChatGPT Images, such as make posters, logos, icons, stylish product shots and much more. Each different task now has its own template, which you can work through with ChatGPT.</p><p>Each template starts with a basic prompt, then it asks a series of follow up questions so that it can find out exactly what you want. </p><p>For example, choose the <strong>Product photo</strong> template and it asks you to upload a photo of the product you want to feature, then when you’ve done that it asks for a style, asking you to choose between styles like Clean Studio, Editorial, Lifestyle and Dramatic, amongst other choices. Then it gives you a choice of a setting, like Dark Studio, then finally it starts to create your image.</p><p>If you’re going to ChatGPT to create something specific then the new Templates are going to be a great place to start.</p><h2 id="and-there-39-s-more">And there's more!</h2><p>That’s not all that’s changed in the new ChatGPT Images 2.5. The images it creates are more personalized, so you get better results when working with recognizable people and portraits. It’s better at infographics than before, and it’s now faster than ever, which is great because I did find it a bit on the slow side previously.</p><p>Finally, when you tap to share an image, you now also have the option of sharing the prompt that created the image, too, so people can make their own versions of whatever image you created more easily.</p><p>All in all these improvements are welcome, if not groundbreaking. I can see why they didn’t call this Images 3.0. It’s essentially a much improved version of the previous Images, but it’s certainly going to change the way I use ChatGPT images from now on.</p> </article> ]]>