OpenAI's gpt-image-1 enables image generation and image editing via OpenAI's image API, ideal for creating and refining images. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.
Entrée
Sortie
InactifVotre requête coûtera 0.011 par exécution. Avec $10, vous pouvez exécuter ce modèle environ 909 fois.
Vous pouvez continuer avec :
Schema d'entrée
Les paramètres suivants sont acceptés dans le corps de la requête.
Aucun paramètre disponible.
Exemple de corps de requête
{
"model": "openai/gpt-image-1/edit"
}Veuillez vous connecter pour voir l'historique des requêtes
Vous devez vous connecter pour accéder à l'historique de vos requêtes de modèle.
Se ConnecterOpenAI GPT-Image-1-Edit
Edit and transform images with natural language using OpenAI's GPT-Image-1-Edit. This versatile model understands your instructions to apply style changes, modifications, and creative transformations — with optional mask support for precise regional editing and multiple quality tiers to match your needs and budget.
Why It Looks Great
- Natural language editing: Describe transformations in plain text — style changes, modifications, enhancements.
- Mask support: Use mask images for precise control over which areas to edit.
- Quality tiers: Choose from
low,medium, orhighquality based on your needs. - Multiple sizes: Output in square (
1024x1024) or rectangular (1024x1536,1536x1024) formats. - Style transformation: Excels at converting images to different artistic styles.
- OpenAI quality: Powered by advanced vision-language understanding.
Parameters
| Parameter | Required | Description |
|---|---|---|
prompt | Yes | Text instruction describing the edit or transformation you want. |
image | Yes | Source image to edit (upload or public URL). |
quality | No | Output quality: low, medium, or high. Default: medium. |
mask_image | No | Optional mask to specify edit regions (upload or URL). |
size | No | Output dimensions: 1024x1024, 1024x1536, or 1536x1024. |
enable_sync_mode | No | API only: Waits for result and returns it directly. |
enable_base64_output | No | API only: Returns base64 string instead of URL. |
How to Use
- Write your edit instruction — describe the transformation you want (e.g.
"Become a comic style"). - Upload your image — drag and drop or paste a public URL.
- Choose quality — select
low,medium, orhighbased on your needs. - Add mask (optional) — upload a mask image to limit edits to specific areas.
- Select size — choose your desired output dimensions.
- Run — click the button to apply the edit.
- Download — preview and save your transformed image.
Quick Reference
| Quality | 1024x1024 | 1024x1536 / 1536x1024 |
|---|---|---|
| Low | $0.011 | $0.016 |
| Medium | $0.042 | $0.063 |
| High | $0.167 | $0.250 |
Best Use Cases
- Style Transfer — Convert photos to comic, cartoon, painting, or other artistic styles.
- Creative Transformation — Reimagine images with different aesthetics or themes.
- Regional Editing — Use masks to edit specific areas while preserving the rest.
- Content Enhancement — Improve or modify specific aspects of images.
- Artistic Interpretation — Transform photos into various art forms.
Example Prompts
"Become a comic style"
"Transform into a watercolor painting"
"Make it look like a vintage photograph from the 1950s"
"Convert to anime style illustration"
"Apply a cyberpunk neon aesthetic"
"Turn into a pencil sketch"
Quality Guide
| Quality | Best For | Trade-off |
|---|---|---|
| Low | Quick previews, testing concepts, high-volume processing | Fastest, most affordable, lower detail |
| Medium | General use, social media, balanced needs | Good quality/cost balance |
| High | Professional work, final deliverables, maximum detail | Highest quality, premium price |
Pro Tips for Best Results
- Start with
mediumquality to test your prompt, then upgrade tohighfor final output. - Use masks when you want to preserve specific areas untouched.
- Be specific about the target style —
"comic style","oil painting","anime". - For style transfers, simpler source images often produce cleaner results.
- Rectangular sizes work well for portraits (
1024x1536) or landscapes (1536x1024). - The model interprets style instructions creatively — embrace the artistic interpretation.
Notes
- If using URLs for images or masks, ensure they are publicly accessible.
- The
enable_sync_modeandenable_base64_outputoptions are only available through the API. - Mask images should be black and white, where white indicates areas to edit.
- Processing time varies by quality level — higher quality takes longer.
