Back to all models

OpenAI

Generate images

GPT Image 2.5 Flare Text-to-Image

Generates or edits an image with top-tier text rendering, 1K/2K/4K output, reference images, and optional mask inpainting.

Cost depends on your settings.

Read docs
Input
Advanced settings

Set compression for JPEG or WebP output, from 0 to 100.

Output
GPT Image 2.5 Flare Text-to-Image: Crimson towers from the well
Example

Try an example

  • A low-angle upward shot in a city of blood-red skyscrapers, the towers converging toward a single vanishing point in a pale grey sky. The buildings are stepped and sharp-edged, ten to twelve of them in two arcs, their glass facades reflecting each other in deep crimson and near-black shadow. Thin horizontal window lines run up every face. The perspective is symmetrical and dizzying, like looking up from the bottom of a well. Overcast, even light. Photographic, ultra-sharp, wide-angle lens. No text or signage.

More from OpenAI

Browse allBrowse all models