GPT Image 1.5 is OpenAI’s third new model in about a month, and it arrives in the wake of Nano Banana Pro‘s dominant launch (the original version won our Technical Excellence award). I’ve spent some time with the new ChatGPT model and found that it’s appreciably better than its predecessor. (This isn’t another GPT-5.2 situation, in which few people noticed any differences after the upgrade.) Generated images have fewer errors and less distortion, and they tend to look more realistic. That said, GPT Image 1.5. still loses out to Gemini’s leading model in some scenarios. I break it all down below.
What Is GPT Image 1.5, and What Can It Do?
GPT Image 1.5 is ChatGPT’s latest image generation model. If you generate an image with ChatGPT now, either as a free or paid user, you’re using GPT Image 1.5. OpenAI promises a range of improvements with the new model, including better text generation, closer adherence to instructions, editing that keeps details intact, faster overall performance, and higher-quality generations with fewer errors and less distortion.
Alongside the new model, OpenAI also introduced an Images tab in ChatGPT. This is a hub for both accessing the images you generate and generating new ones. It’s more of a new interface element than new functionality, but it’s a convenient way to see what you’ve created. It also provides easy access to different styles you can generate images with and sample prompts.
ChatGPT’s New Image Model Is a Definite Improvement
Considering how unapologetically uninteresting OpenAI’s last few model releases have been, I wasn’t expecting much from GPT Image 1.5. However, I’m happy to report I was wrong. GPT Image 1.5 might not be revolutionary, but its upgrades are obvious. Editing images results in less distortion, and generating ones from scratch feels much snappier. Results also just look a lot more realistic, too.
Generated Images Look More Lifelike
I can appreciate GPT Image 1.5’s improvements over OpenAI’s older model, GPT Image 1, but the more important test is how it stacks up with the current leader in AI image editing and creation, Gemini’s Nano Banana Pro. I always start my AI image generation testing by asking for a relatively basic image of the interior of a home. As such, I prompted GPT Image 1.5, GPT Image 1, and Nano Banana Pro to “Generate me an image of a cozy suburban home with an open floor plan. I want to see a nice living space with a dining room, kitchen, and living room. Nothing too fancy.”
GPT Image 1.5’s (first slide) improvements are substantial. GPT Image 1’s image (second slide) looks like a stock photo and is strangely bereft of furnishings and other details. The new model’s image has rich details and looks far more realistic. It’s roughly comparable to Nano Banana Pro’s image (third slide), though some minor errors in the background, such as conjoined chairs, are still noticeable.
I also test AI image tools by asking them to generate complex illustrations, such as comics. I used the following prompt: “Generate me a six-panel comic of a high fantasy world where magic users have assault rifles instead of wands or staves, but you’re going to spice it up: I want the magic users fighting robot knights wielding swords and shields. Make sure there’s a major twist by the final panel.”
Compared with GPT Image 1’s comic (second slide), which has very basic art and doesn’t deliver on the twist I asked for, GPT Image 1.5’s comic (first slide) looks excellent. Its art is much more detailed, and it actually manages to include a real twist. However, Nano Banana Pro’s comic (third slide) is still the best overall, since ChatGPT included seven panels (I asked for six) and awkwardly cuts off that last panel.
Get Our Best Stories!
Your Daily Dose of Our Top Tech News
By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use and Privacy
Policy.
Thanks for signing up!
Your subscription has been confirmed. Keep an eye on your inbox!
Diagrams often push AI image generators to their limits because of their complexity and heavy emphasis on text. So, I gave the bots the following prompt: “I’ve got an Ethernet switch, a modem, and a router. I want to connect my PC, PlayStation, smart light hub, and smart TV via Ethernet. I want the modem to connect to the router, and then the router to connect to the switch. Then, I want the switch to connect to the devices. Generate me a diagram showing this.”
ChatGPT has historically done well at this task, but GPT Image 1.5 (first slide) does even better. Its graphics have more detail, and I couldn’t spot any misspellings like in GPT Image 1’s diagram (second slide). Again, however, Nano Banana Pro (third slide) manages to produce the best image of the bunch. It isn’t missing any of the connections (unlike between ChatGPT’s Ethernet switch and PC) and maintains a more consistent font.
Editing Works More Reliably (But Still Imperfectly)
Editing is another core component of AI image models, so I tested the models’ abilities to remove a pesky hand from a nature shot (first slide). GPT Image 1.5’s (second slide) image has the best overall contrast and most vibrant colors, and it’s significantly sharper than Nano Banana’s Pro’s take (fourth slide). However, the area it generated underneath my hand doesn’t match the original scene quite as well as Nano Banana Pro’s image did. Furthermore, GPT Image 1.5’s image adds a sort of tessellation effect to finer details and doesn’t maintain the original image’s aspect ratio. GPT Image 1’s image has similar strengths and weaknesses.
Recommended by Our Editors
AI image generators can also combine multiple images into a cohesive scene. I tasked the models with combining images of chairs and a table with an image of my kitchen. GPT Image 1.5’s result (first slide) does a significantly better job of not distorting surrounding elements compared with GPT Image 1 (second slide). However, the lighting doesn’t match up well with the rest of the room. In Nano Banana Pro’s image (third slide), the chairs and table look more like they belong in the kitchen.
The above testing doesn’t account for every possible use case, and, considering the variance inherent to testing AI chatbots, it’s inevitable that GPT Image 1.5 will sometimes get things right when Nano Banana Pro stumbles. However, on average, Nano Banana Pro seems to do a slightly better job with both image editing and generation. That doesn’t take away from the fact that GPT Image 1.5 is an obvious upgrade over the model it replaces, however.
A Fine Upgrade If You Prefer ChatGPT to Gemini
If ChatGPT is your chatbot of choice, you should dive right in with its new image generation model because it’s just better than what you had access to before. Simply prompt ChatGPT to generate the same image you have in the past, and the differences should be clear.
If you are a habitual Gemini user, you don’t need to concern yourself much with GPT Image 1.5. It’s mostly competitive, but I haven’t seen it beat Nano Banana Pro with any level of consistency. It’s at least a good backup option if you use up all your free access to Nano Banana Pro and don’t feel like paying, or it can be worth trying out if you aren’t happy with the results you’re getting from a particularly troublesome prompt.
Disclosure: Ziff Davis, PCMag’s parent company, filed a lawsuit against OpenAI in April 2025, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.
About Our Expert
Ruben Circelli
Writer, Software
Experience
I’ve been writing about consumer technology and video games for over a decade at a variety of publications, including Destructoid, GamesRadar+, Lifewire, PCGamesN, Trusted Reviews, and What Hi-Fi?, among many others. At PCMag, I review AI and productivity software—everything from chatbots to to-do list apps. In my free time, I’m likely cooking something, playing a game, or tinkering with my computer.
Read Full Bio
