Back to models
OpenAI
Image modelHotGPT Image 2.5
GPT Image 2.5 is OpenAI's image generation and editing family, with Sunburst focused on quality and Flare focused on speed. For website users, the name represents a shared set of creative capabilities: generating new images from text, or providing images with prompts to modify, rearrange, or preserve subjects. Both variants support five quality levels and output from 1K to 4K. Sunburst favors detailed content, while Flare favors rapid exploration, and editing accepts multiple reference images. It suits combining people and products, revising brand assets, changing styles, reconstructing scenes, and preserving important visual information across revisions. For more consistent results, describe subjects, composition, text, style, and elements that must not change separately. Focus on a few edits at a time to make progress toward the target easier to assess. If a website presents a single GPT Image 2.5 entry point, clearly identify whether it uses Sunburst or Flare in the creation interface or specifications, so users do not assume identical speed and image quality.
- Precise Editing
- Two Model Variants
Key features
Sunburst and Flare
Choose a variant for detailed imagery or rapid iteration.
Editing with multiple images
Use one or more reference images and text instructions to adjust content while preserving important subjects.
Five quality levels and 4K
Offers quality settings from low to max and 1K, 2K, and 4K sizes for different delivery requirements.
