OpenAI opens image model API to global businesses

OpenAI has rolled out its image-generation model, gpt-image-1, to developers and businesses worldwide via API access, expanding the reach of its multimodal capabilities. The launch was confirmed by CEO Sam Altman in a post on X.

This comes a month after ChatGPT’s image-generation feature went live for users, with over 700 million images created by more than 130 million people in its first week.

Now, organisations can plug the API into their own products to generate high-quality images from text prompts—spanning use cases across design, ecommerce, marketing, education, and gaming.



Several tech firms are already experimenting with the model:

• Canva is integrating it into its design tools to auto-generate visuals from rough sketches.
• GoDaddy is exploring logo generation, background removal, and social media content creation.
• HubSpot is testing it for automated marketing imagery.
• Instacart is using it to visualise recipes and grocery lists.
• InVideo is incorporating it to enhance AI-driven video editing.

How it works

The gpt-image-1 model supports detailed text-to-image generation, multiple visual styles, and even renders embedded text—making it valuable for creative professionals and businesses alike.

Pricing & safety

The API follows a tiered pricing model:

• $5/million tokens for text input
• $10/million for image inputs
• $40/million for image outputs

Depending on output quality and size, this works out to around $0.02–$0.19 per image.

To address safety, OpenAI has built-in content filters and C2PA metadata tagging, which flags images as AI-generated. Developers can also calibrate how strict the filtering should be. OpenAI said it will not use customer data or generated images from the API to train future models.

The API is now available globally, although verification may be required for access in some cases.