gpt-image-1 is OpenAI's native multimodal image model, designed for text-to-image generation and image-based visual editing. Its strengths include not only generating images in different styles, but also understanding composition requirements, following creative guidelines, and rendering text in images. It is suitable for posters, product visuals, illustrations, and design drafts, using dedicated image generation or editing endpoints for integration.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
api.acedata.cloudgpt-image-1Input parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
Model capabilities include image generation, text rendering, and visual editing; the input quantities and result formats above are invocation specifications for this platform's image endpoint.
gpt-image-1 can create images by combining requirements for theme, style, and composition, making it suitable for organizing scattered creative briefs into visual concepts. Prompts can separately describe the subject, environment, lighting, and negative space, then add brand colors and elements that must be avoided, making the generation goal clearer than simply specifying an art style.
Text rendering in images is a clear strength of this model, making it useful for headline posters, promotional graphics, and visual drafts with labels. When creating, list the text that needs to appear separately and specify the title hierarchy and placement; before final delivery, still check spelling, punctuation, and small text to avoid treating generated results directly as finalized layout.
In addition to generating from scratch, gpt-image-1 is also suitable for making visual modifications based on original images, such as changing the background, adjusting the color palette, or developing a sketch into a complete graphic. When editing, clearly specify what should be retained and what needs to change, and choose reference fidelity control appropriately; this conveys design intent more effectively than simply writing “optimize it.”
Enter the event theme, headline copy, brand color palette, and layout requirements to generate posters or social media graphics for discussion. This is suitable for comparing different visual directions first, then selecting a concept for further revision. When dates, prices, and event terms are involved, they should be checked item by item, and the necessary design review should be completed before official publication.
Provide a product image URL, describe the desired background, lighting, and visual atmosphere, and emphasize the appearance features that should be retained to explore product promotional images. After generation, focus on checking whether logos, materials, structure, and colors are accurate; when transparent assets are needed, choose an output format that supports transparency for further processing.
Submit sketches or reference images together with written instructions, describing the line style, color relationships, and intended use to create illustrations, graphic elements, or concept proposals. Each reference image should serve a clear purpose, for example, one explaining the composition and another explaining the color palette, reducing conflicting visual requirements and making subsequent selection and revision easier.
If a task includes both “generate an image first” and “continue modifying it based on the original image,” the generation and editing workflow of gpt-image-1 is worth prioritizing. Compared with DALL·E 3, which is primarily positioned for text-to-image generation, this is better suited to organizing modification requests around reference images. For text analysis or general Q&A alone, a conversational model should be selected.
gpt-image-1, gpt-image-1.5, and gpt-image-2 are different models, so you should not replace calls based on their names alone or assume their parameters are exactly the same. Existing gpt-image-1 workflows can continue to be optimized around text rendering and reference image editing; when considering a switch, it is recommended to compare results using the same set of posters, product images, and modification instructions before deciding whether to migrate.
Provide a prompt when generating; provide both an image and editing instructions when editing. Clearly specify the original text, subject preservation requirements, and target aspect ratio.
Use the image generation or image editing endpoint, explicitly specifying model=gpt-image-1; use auto or WIDTHxHEIGHT for size, and generate one image first before evaluating. Set masks, quality, and file format according to the guide for this endpoint.
Read the URL or Base64 image according to the response format, asynchronously save the task_id before querying results; check text, reference details, and the alpha channel, and record usage according to the current Pricing rules.
Create a horizontal community workshop poster, titled “Build Something,” featuring tools and materials on a desk, with space reserved on the right for registration information, in a blue-and-white color scheme with clear text.
First establish the visual design through the image generation API, then refine it with the image editing API; proofread the title and whitespace, and integrate according to the image responses for generation and editing.
No. gpt-image-1 is designed for image generation and editing; when calling it, you should select the corresponding image endpoint and model ID. It has technical ties to the ChatGPT image experience, but this does not mean that regular gpt-4o conversation calls are the same workflow, nor does it automatically include features such as voice or web access.
Call /openai/images/edits, explicitly specify model=gpt-image-1, and provide the image URL and prompt modification instructions. You can submit multiple reference images, but it is best to explain the purpose of each image while distinguishing what should be kept and what should be changed, making the editing goal easier to express.
Write the title text separately from the image description, clearly specify the exact text to appear, its position, and hierarchy, and try to avoid crowding long passages into complex backgrounds. gpt-image-1 excels at rendering text in images, but the completed image still needs to be proofread character by character; for formal posters, refine the typography after confirming the visual design.
The image endpoint provides background=transparent, which can be used to request a transparent background; PNG or WebP is recommended for output. The prompt should also describe requirements for the subject and edges. After downloading the result, check the alpha channel and edge quality, especially thin lines, semi-transparent materials, and shadow areas.
Use the image generation endpoint for creation from scratch; use the editing endpoint to modify existing images, and explicitly specify model=gpt-image-1. The application reads the image URL or Base64 result; for asynchronous requests, save task_id and query it later. Do not process image responses as ordinary conversation text.
Choose a development approach or application suited to this model and follow its setup guide.
↗Python SDK Integration Guide
SDK installation, authentication and examples↗TypeScript SDK Integration Guide
SDK installation, authentication and examples
↗↗Pricing
$0.019· per generation
You pay$13,584for 142,800 credits · $0.0951 / credits46% OFF
You're billed in credits. Pick a top-up tier to see its per-credit rate — the larger the top-up, the cheaper each credit.
This is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs