Getting your Trinity Audio player ready...
OpenAI has launched in recent days an advanced new tool for creating images directly within ChatGPT – and the internet is abuzz. According to the company’s CEO, Sam Altman, the tool has even pushed OpenAI processors to their limits due to the heavy workload.
Indeed, the platform's graphical capabilities have received a significant upgrade, thanks in large part to the new GPT-4o model, which better understands what users imagine and turns it into highly accurate images.
ChatGPT is not just a single model but a comprehensive platform. Within it, you can find tools such as reasoning models from the o1 series, Canvas for editing texts and files, an advanced voice system, web search capabilities, data analysis tools, and, of course, the well-known image generator DALL·E, which is integrated directly into the chatbot.
So how does it work?
When users request an image from ChatGPT, the GPT-4o model processes, refines and enhances the request, then sends it to DALL·E 3, which generates the image and delivers it back to the user. The more creative the model, the better and more accurate the result will be. Similarly, the more detailed the user’s prompt, the more precise the final output will be.
Currently, free-tier users have a limit on the number of images they can create, but paid subscribers ($20 per month) can use the new tool almost without restriction. To truly test how much the tool has improved, we tried some particularly challenging prompts – and the results were simply stunning:
Prompt example 1:
"Create a 16:9 atmospheric image of a lone lighthouse standing on the edge of a cliff in a post-apocalyptic world. The sky is dark with storm clouds, the light from the lighthouse cuts through the fog, and the sea below is stormy. In the surroundings, remnants of abandoned technology are visible."
10 View gallery
A lone lighthouse, a post-apocalyptic world and a beam of light piercing the darkness
(Photo: ChatGPT)
Get the Ynetnews app on your smartphone: Google Play: https://bit.ly/4eJ37pE | Apple App Store: https://bit.ly/3ZL7iNv
Prompt example 2:
"Design a futuristic simulation in a 16:9 format of a central street in Tel Aviv in the year 2125 – with illuminated skyscrapers, airborne transportation filling the skies, a futuristic local market with floating stalls, and people in high-tech clothing."
Prompt example 3:
"Create a digital fantasy-style illustration of a giant, mysterious cat with glowing eyes and stars in its fur. The cat hovers above a small planet, while galaxies swirl in the background."
10 View gallery
The galactic cat—after we asked the chatbot to add a splash of color
(Photo: ChatGPT)
Prompt example 4:
"Create a colorful 16:9 fantasy image of an enchanted forest lit by glowing fairies. At its center, an old wooden table surrounded by small robots with human-like expressions, all drinking tea together."
Prompt example 5:
"Draw a Renaissance-style portrait of Kratos from the God of War video game, wearing classical clothing, dramatic lighting and European castle in the background."
Prompt example 6:
"Design a dramatic poster for a horror movie titled "The House Beyond the Fog." The poster shows a Victorian house in the middle of a misty forest, with mysterious shadows in the windows and bold red lettering."
Prompt example 7:
"Draw a 16:9 atmospheric image of a massive library built inside the crater of an extinct volcano. Towering bookshelves stretch to the ceiling, suspended bridges connect the stacks, and soft lava light glows through the space."
And finally—because we couldn’t resist—an image of yours truly, originally uploaded as a regular photo and processed in a Studio Ghibli style.
10 View gallery
Digital section editor Daniela Ginzburg, in an image processed by ChatGPT
(Photo: ChatGPT)
First published: 13:44, 04.01.25