ChatGPT image prompts
Every picture here came out of GPT Image 2, the OpenAI model behind ChatGPT Images 2.0. We ran it on Leap, not in ChatGPT, and the words under each picture are the prompt that made it. Copy one to use in ChatGPT, or press Make this to open GPT Image 2 in the studio with the prompt in the box and the same shape.
7 prompts, each under the picture it made.
Film stills, live action and anime
Each opens with the kind of frame (a film still from an independent drama with Kodak 5219 grain, a frame from a hand-drawn 1990s Japanese animated feature), then the hour, the exact place and every person’s age and clothes. A “Palette:” line names three or four colors, and the drama adds “Real faces: liver spots, pores, wet hair” so the skin stays real.
Candid photos with animals
Both describe one exact moment as a photograph: a gull swooping in as the chip reaches his mouth, a cat’s paw on a full glass of tea while a hand reaches for it. They give the lens and the light, then the textures to keep real: “Real wind, salt spray, grain”, “Real fur, dust in the sunbeam, steam from the tea”.
Ads and magazine pages
Name the format first: a car commercial frame, an open magazine photographed from above. The magazine prompt puts its headline in quotes and lists what must stay off the page (brand names, photo credits, prices). The car is invented, with “No badges, no number plate”.
Write your own next.
Start free, no card needed. Every run is priced before you press Generate, and a run that fails costs nothing.
Open GPT Image 2





