Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Back in March, while I was Testing and review In the first iteration of the OpenAI image model, I gave the chatbot what I thought would be a simple test: Make me a graph that explains the water cycle. This test ended up being a disaster. So when OpenAI Announce With its new 1.5 model and ChatGPT Images interface, I was ready for another try. Sadly, what I found was a symbol of progress and the ongoing issues with generative media in 2025.
Stay with me for a minute and remember when you first learned about the bathroom. For me, it was in elementary school, and we had a rhyming song to help us remember how water is transformed throughout the environment, from water vapor to clouds to rain. I clearly remember looking at the different pictures and videos in class explaining the different stages.
So, for a company that says their chatbot is GPT-5 Smart as a Ph.DWhy can’t their image models provide a simple diagram of one of the most fundamental (or should we say, elementary) and well-documented geological processes?
(Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging that it infringed Ziff Davis’s copyrights in training and operating its AI systems.)
This has been much improved for ChatGPT, but it is still not perfect. Stocks are not facing the right directions.
If I were to defend ChatGPT, I would point out that this newer graph is much better than previous versions. The text is finally legible, which is a big step up from just a month ago. It included all stages and their sources, as requested. But the arrows are still wrong, as the vertically facing precipitation arrow gives the impression that rain can fly up from the ground. I asked him three more times after his initial failure, and the results were almost identical.
Using the original OpenAI image model, this took over 20 prompts and 2 hours to create. There are still errors.
I’m sure no one using the new ChatGPT Images feature is as invested as I am in its ability to properly display the bathroom. Maybe I’m being overly picky. But the issue highlights an ongoing problem with OpenAI: It may have one of the most widely used chatbots, but its creative services leave a lot to be desired. This wasn’t a big deal before, until Google changed the game this summer.
This holiday card made with ChatGPT preserves my friends’ features and mine from the original image (left) well, but our skin still has that shiny AI look.
ChatGPT images from OpenAI vs. Nano Banana from Google
OpenAI and Google have been competing throughout 2025 in the race for the most advanced AI. Google has often taken the lead, but OpenAI has never lagged behind. When Google dropped its industry-leading video and photo models, e.g I see 3 and Nano bananaOpenAI will follow up a few weeks later with Sora And now Picture 1.5. This image model in particular, along with GPT 5.2likely the result of an internal OpenAI “code red” issued out of concern over the technical prowess Google has displayed with its recent suite of models in Gemini 3.
In other words, the two largest technology companies, while not traditionally known for their creative histories, have been investing heavily in creative AI. It has led to tremendous technical progress – and So many fears.
ChatGPT images (right) are improved over previous versions, but they don’t fool anyone like Gemini’s AI-powered images (left).
Google’s Nano Banana Pro was recently crowned the best AI image creator of 2025, and it was one of the most impressive and inspiring AI programs I’ve ever used. It wasn’t perfect, and it definitely left me feeling annoyed, but it did the job it should have been able to do. In comparison, the ChatGPT images with the new 1.5 model were a huge leap forward for the company but didn’t wow me in the same way.
Gemini (left) managed to create a more realistic version of my family at Appalachian State University’s Kidd Brewer Stadium, while ChatGPT’s is clearly fake due to the snow falling over the top.
Gemini generates images faster than ChatGPT, and these Gemini images tended to perform better in my tests. I also reached my generation limit with the free version of ChatGPT faster than I did with the free version of Gemini, which was annoying.
One area where ChatGPT does well is in less realistic, more creative work. OpenAI’s AI images have become known for the way they tend to have cartoon-like patterns and a dull yellow color in some images. The yellow tint is gone now, but the animation style remains where ChatGPT excels. She certainly has a lot of experience in this field, since her first model went viral among people Studio Ghibli-esque versions themselves.
In the short time I spent using the new model, I could see where OpenAI was headed. It has already shown significant improvements, especially with realistic text and photography. But I can’t help but wonder if this 1.5 model is a half-measure to prove it can compete with Google and that there are better capabilities waiting in the full second-generation model. We’ll have to wait and see. For now, my advice remains the same: ChatGPT Images are great for ChatGPT users, but Nano Banana Pro is truly exceptional.