Use GPT-4o Image to create a photograph, illustrated concept or lettered poster from a written brief, or provide a reference image and describe the changes you want to make. The existing task page brings generation and editing together: select gpt-4o-image for text generation or gpt-4o-image-edit for reference editing with an optional mask. Set a square, portrait or landscape ratio; generation can request one to four images, while an editing request returns one image.
GPT-4o Image is developed by OpenAI. Vidgo provides access to the model. Model names and trademarks belong to their respective owners.Open the GPT-4o Image task to choose generation or editing. Start from text when the composition is still a written idea; use an existing image when a subject, setting or design already provides the visual starting point.
For text generation, describe the subject, composition, materials, lighting and any exact words that should appear. For editing, upload the reference image, describe the requested changes and identify the details to preserve; add a mask when the change should target a selected region. Both modes accept prompts up to 1,000 characters and the ratios 1:1, 2:3 or 3:2. The current online editing form accepts one reference image.
| Model / task | Input and workflow |
|---|---|
| GPT-4o Image | OpenAI's native multimodal image generator with exceptional text rendering, precise prompt following, and conversational editing capabilities. |
The published PoYo examples include a farmers-market photograph, a typography poster and a cat edited with detective accessories. Each keeps its original prompt and media pairing. Use them to inspect how subject descriptions, exact text and reference-based instructions appear in a result; these independent demonstrations are not a controlled comparison.
Describe the person, activity, location, light and camera treatment in one brief. The market example requests a girl drinking pink boba tea, sunlit tents, a blurred crowd and film grain. Inspect the scene arrangement and photographic treatment against those instructions.
Write the required text in quotation marks and give each line a position. The poster prompt places a title at the top, a phrase at the lower right and a credit at the bottom, alongside a sculpture transitioning from wireframe to photorealism. Review both lettering and placement.
Upload a clear source image in the editing task and describe the additions or changes. The cat example uses a supplied cat photograph as the starting point, giving the model a concrete subject on which to place the requested accessories.
The published edit asks for a detective hat and monocle. Compare the source and result to inspect those additions and the surrounding subject. For your own local edits, use the optional mask_url with gpt-4o-image-edit to identify the region, and describe the intended change in the prompt.
{
"size": "3:2",
"n": 1
}Choose 1:1 for square assets, 2:3 for portrait work or 3:2 for landscape compositions. Set n from one to four for text generation when exploring several results. Use n: 1 for editing, which returns and bills one image per request.
These two public PoYo playground demonstrations retain their complete prompts and original CDN results. The poster request specifies three text placements and a wireframe-to-photograph sculpture. The editing request starts from the displayed cat image and asks for a hat and monocle. Compare each result with its own prompt and, for the edit, with its source image.
Create a vertical poster with the words 'ChatGPT image generation' at the top, and the phrase 'Form follows function' on the lower right. In the background, put a photo of a really sleek, modern sculpture. It should gradually transition from a wireframe sketch on the left to the fully photorealistic version on the right. At the very bottom, in medium-small text, say 'This entire poster was generated by ChatGPT image generation.'
Give this cat a detective hat and a monocle
Start with a complete brief or an existing visual, then choose generation or editing for the work you need. Review text accuracy, subject appearance, framing and the requested changes before adding the image to a presentation, content page or design draft.
Prepare a subject, action, location and lighting brief, then generate a scene to explore the intended visual story. Use the result in an editorial mood board or as a reference for a subsequent photo shoot.
Prepare the exact headline, supporting lines and intended positions, then describe the background image and visual hierarchy. Review the generated typography and composition for an event poster, campaign draft or presentation cover.
Upload an existing character or animal image and describe the accessory, clothing detail or prop you want to add. Compare the edited subject with the source when developing illustration concepts and character variations.
Prepare the source image and a mask for the region you intend to change, then explain the new content or appearance. Review the modified region and its surroundings before using the result in a design iteration.
Describe the subject and space needed for the composition, then select 1:1 or 2:3. Inspect the crop and text placement for social content, article illustrations or portrait-format promotional drafts.
Write a clear text-generation brief and choose an output count from one to four. Compare the returned images for composition and subject treatment, then use the selected direction as the basis for the next creative iteration.
The table reads the current catalog rate for this task. Text generation uses n to request and bill one to four images, so the requested count affects the total charge. The editing model returns and bills one image per request; use n: 1 when editing. Use the dynamic starting rate below and the selected task’s estimate to budget the complete request.
| Model | Provider | Workflow | Input material | Output specification | Starting rate |
|---|---|---|---|---|---|
| GPT-4o Image | OpenAI | Image generation / editing | Text prompt; one reference and optional mask in online editing | 1:1 / 2:3 / 3:2 | $0.02 / generation |
Test a complete creative brief online, then bring the verified request into your application.
Enter the complete brief, select the mode and size, and review the generation estimate. For online editing, upload one JPEG, PNG or WebP reference within 10 MiB, and add a mask when needed. Generate the result and check the exact text, subject details and requested changes.
Use the existing task page, current price information, task documentation and generation history to move from a creative test to an API workflow.
Open the GPT-4o Image task and select text generation for a new composition or editing for an existing image. Prepare a prompt up to 1,000 characters, choose 1:1, 2:3 or 3:2, and use the editing mode when supplying a reference or mask.
Enter the complete brief, select the mode and size, and review the generation estimate. For online editing, upload one JPEG, PNG or WebP reference within 10 MiB, and add a mask when needed. Generate the result and check the exact text, subject details and requested changes.
The table reads the current catalog rate for this task. Text generation uses n to request and bill one to four images, so the requested count affects the total charge. The editing model returns and bills one image per request; use n: 1 when editing. Use the dynamic starting rate below and the selected task’s estimate to budget the complete request.
Open the task documentation and review authentication, the model identifier, required inputs, accepted output settings and response fields. For reference inputs, prepare accessible, directly downloadable HTTP(S) image URLs. Carry the tested prompt and settings into your request.
Send the authenticated request with the matching model identifier and input object. Retain the returned task ID so your application can track this generation. Provide a callback URL when your application is ready to receive task notifications.
Use the task ID to query status, or handle the configured callback. When the task finishes, retrieve the generated file URLs, save the result for your project and compare it with the original brief. Revisit online generations in your account history when refining the next request.
GPT-4o Image is an OpenAI image-generation and editing family accessed through the existing Vidgo task page. Text generation uses gpt-4o-image; reference editing with optional masks uses gpt-4o-image-edit. The page provides online testing, API documentation and access to generated files.
Read model introductions, API guides and platform comparisons to understand creative inputs, integration choices and generation costs.
