Presenting the API’s most recent picture creation model GPT-Image-1
Last month, OpenAI added picture production to ChatGPT, and it immediately became one of the most popular features. In the first week alone, over 130 million users worldwide produced over 700 million images.
Through gpt-image-1, OpenAI exposing the natively multimodal model that underpins this ChatGPT experience to the API today, making it simple for developers and companies to include professional-caliber, high-quality picture production straight into their own tools and platforms. Because of its adaptability, the model can produce images in a variety of styles, adhere to certain criteria, use its knowledge of the environment, and display text appropriately, opening up a world of useful applications in a wide range of fields.
Image creation is already being used in products and experiences by top companies and startups in a variety of industries, including creative tools, e-commerce, education, corporate software, and gaming.
Additionally, OpenAI still collaborating with developers and companies to find other ways that picture creation via the API may support their use cases, such as:
Canva is investigating methods to incorporate OpenAI’s gpt-image-1 into its Canva AI and Magic Studio products in order to advance design creation and modification. By enabling high-fidelity visual changes and converting rough drawings into beautiful graphic components, Canva’s gpt-image-1 capabilities will enable its 230 million users to realise their ideas in whole new ways.
Logo creation and editing: GoDaddy is currently testing the integration of image generation to enable users to quickly and simply design editable logos, eliminate backgrounds, and produce expert typography. Additionally, as a novel experience with GoDaddy Airo, the team is assisting clients in producing marketing materials and social media posts that showcase their distinct brand identities for original brand and product content generated in response to prompts.
Creating marketing material: HubSpot is investigating how clients may use OpenAI’s recently released AI picture generating capabilities to produce marketing and sales collateral. These picture generating features might revolutionise how clients create excellent photos for social media, email marketing, and landing pages without the need for specific design knowledge as HubSpot keeps adding new AI-powered tools for growing businesses.
Making shopping lists and recipes: Instacart is experimenting with new methods to generate photos using the API, such as creating images for shopping lists and recipes.
Video editing: Invideo uses AI to help millions of people turn their ideas into videos. The platform now provides enhanced text creation, fine-grain editing tools, and sophisticated style advice to the inclusion of gpt-image-1.
Security
The API’s gpt-image-1 model makes use of the same security measures as ChatGPT’s 4o image creation, which include measures to prevent the model from producing dangerous photos and the addition of C2PA information to created images. More information on this topic may be found in the previous blog article here. With the moderation option, developers may additionally adjust the sensitivity of moderation. It can be set to low for less stringent filtering or auto (default) for standard filtering.
OpenAI never train on client API data by default, and all picture inputs and outputs are still governed by the API usage guidelines.
Pricing
The cost of using gpt-image-1 is token-by-token, with distinct costs for text and picture tokens:
- $5 for per million text input tokens (prompt text)
- Tokens for picture input ($10 for 1 million tokens)
- Tokens for picture production (produced images): $40 for each million tokens
For low-, medium-, and high-quality square pictures, this corresponds to around $0.02, $0.07, and $0.19 per created image, respectively.
Get started
The Responses API will soon support the gpt-image-1 model, which is now accessible worldwide through the Images API. Before using the model, certain developers might need to confirm their organisation.
To begin, check out the picture generating tutorial and explore the model’s possibilities in the playground.
Microsoft Azure excited to share that GPT-image-1, the newest and most sophisticated picture creation model, will soon be available to qualified users on Microsoft Azure OpenAI Service. This innovative model raises the bar for producing excellent photographs, responding to challenging cues, and providing zero-shot capabilities in a range of situations.
