GPT Image 2.5 is an online creative workspace for image generation and editing, where users can create product photos, posters, illustrations, and photorealistic-style images through text prompts or reference images.
Uses This tool is mainly used to generate and download images starting from text ideas or reference images. It can be used for product concept images, posters and event visuals, illustrations and story scenes, and other directions.
Main capabilities
- Text generation: Describe the subject, environment, lighting, and composition through detailed text prompts to guide image generation.
- Reference image editing: Upload one reference image and use text to explain what needs to be changed and what should be preserved.
- Flexible canvas: Supports six aspect ratios, as well as 1K, 2K, and 4K resolution tiers, with output as PNG.
- Unified workspace: Prompts, reference images, model selection, and output dimensions are managed in one place.
- Community showcase: You can browse community works selected by administrators; prompts, uploaded reference images, and account information are not shared.
How to use
- Write clear prompts: Explain the subject, environment, and visual style, and you may add lighting or framing instructions. Prompts support up to 5,000 characters; it is recommended to keep them concise and focused and avoid conflicting instructions.
- Edit from a reference image: Switch to reference editing mode, upload one PNG, JPEG, or WebP file no larger than 5 MB, and then describe what needs to be changed and what should be preserved. Only upload reference images you have the right to use.
- Choose model and canvas: In the model bar, select GPT Image 2.5 or another available model; choose one of the six aspect ratios and a 1K, 2K, or 4K resolution tier, and confirm the specific pixel dimensions before generating.
- Check the result: Inspect the preview for distracting objects, distorted details, or incorrect text, and download the PNG after confirmation. If the composition needs adjustment, modify one part of the prompt and compare the next result.
Applicable scenarios
- Product photography concepts: For example, a ceramic coffee cup on a light wooden table, soft window light, soft shadows, a clean background, and space reserved for a headline.
- Posters and event visuals: For example, a jazz concert scene in midnight blue and warm orange tones, with an abstract saxophone silhouette, and space reserved for the event title and date. Important text layout is recommended to be completed in a design editor.
- Illustrations and story scenes: For example, a small fox reading beside a forest stream, watercolor texture, moss-covered stones, and delicate wildflowers, specifying soft afternoon light and a vertical composition.
Limitations
- Each generation costs 5 credits, and failed tasks refund credits.
- Reference images support only single uploads, in PNG, JPEG, or WebP format, with a size no larger than 5 MB.
- Prompts support up to 5,000 characters.
- Text in the output may contain errors; important layout is recommended to be completed in an external design tool.
- Only upload reference images you have the right to use.
- Paid plans are divided into monthly, annual, and one-time credits; credits are issued after each payment. All plans use the same available models and sizes. For specific prices and credit tiers, please refer to the actual pages on the website.
- More details such as whether it can be used for free and whether created content is public are not fully provided in the materials.
