The move comes after the feature’s debut last month, where over 130 million users generated more than 700 million images in just one week.
The API version makes it possible for organisations to incorporate the model’s image generation capabilities into their own products and services. OpenAI says the model is designed to support practical applications in fields like design, marketing, ecommerce, education, and gaming.
Technical know-how
gpt-image-1 is a multimodal AI model that can generate images from text prompts. It supports a range of visual styles, allows for detailed customisations, and is able to render text within images—features that make it useful across several industries. Users can apply it to tasks such as visual design, content creation, product marketing, and more, the company said.
Use cases
Discover the stories of your interest
As image generation capabilities become more accessible through the gpt-image-1 API, a number of companies across different sectors are already putting the technology to use.
Several companies have begun experimenting with the model:
- Canva is testing integration with its AI-powered design tools to help users turn rough sketches into detailed visuals and perform edits more efficiently.
- GoDaddy is exploring image generation for logo creation, background removal, and producing social media content for small businesses.
- HubSpot is investigating how the model might support customers in creating marketing materials such as images for emails, social posts, and landing pages.
- Instacart is testing image generation for recipes and grocery lists.
- invideo, a video editing platform, has added the model to improve its AI-based visual editing features.
Pricing and safety concerns
The model uses a token-based pricing structure: $5 per million tokens for text input, $10 per million for image inputs, and $40 per million for image outputs. Based on typical usage, this amounts to approximately $0.02 to $0.19 per image, depending on quality and size.
The API is now available globally, although some developers may need to verify their organisation before gaining access.
OpenAI says the image generation tool includes built-in safety measures to help prevent harmful or inappropriate content from being created. These include filters that screen out unsafe material, and digital “tags” (called C2PA metadata) that show the image was made by AI.
Developers who use the tool can also choose how strict the content filtering should be. Importantly, OpenAI doesn’t use images or data from the API to train its models, helping protect users’ privacy.
