OpenAI’s viral image generator is now available to developers through its API
Developers can now bring the same image-generation capabilities into their own apps and services.
Earlier this month, OpenAI’s upgraded image generator in ChatGPT went viral thanks, oddly enough, to its knack for pumping out interesting Studio Ghibli-style visuals and creepy, expressionless dolls. That weirdly powerful combo pulled in millions of users and sent demand for the tool skyrocketing.
According to OpenAI, over 130 million users created more than 700 million images in the first week alone. It was a breakthrough and a logistical headache all at once as it put serious strain on OpenAI’s infrastructure.
Rather than scale back, the company is leaning into the demand. And as of this week, the company says the same “natively multimodal model”— called gpt-image-1 — powering the image generator will be accessible through the company’s API.
Oluwaseun Bamisile
That means developers can now bring the same image-generation capabilities into their own apps and services. And some major names already have.
Per OpenAI's blog post, Figma is already leveraging the latest model to bring advanced image generation and editing capabilities into its platform. Rolling out starting today, users can use ‘gpt-image-1’ in Figma Design to generate and edit images from a simple prompt – adjusting styles, adding or removing objects, expanding backgrounds, and more. This new integration lets designers rapidly explore ideas and iterate visually, all in Figma.
Meanwhile, Adobe is also tapping in. Its suite of creative tools, including Firefly and Express, will soon support image generation through OpenAI’s model, giving creators more flexibility and control over style and content. Instacart is testing it too, using the model to generate images for recipes and shopping lists. Canva, Wix, GoDaddy, and Airtable are also exploring integrations, with more on the way.
OpenAI says gpt-image-1 stands out for its versatility. It can generate images across a wide range of styles, follow detailed instructions, apply its understanding of the world, and even produce legible, styled text within visuals. Developers can customize the number of images generated, adjust quality for faster results, and tweak content moderation settings based on how strict or relaxed they want the filters to be.
Emmanuel Oyedeji
To prevent misuse, the model includes the same safety guardrails as the version in ChatGPT, blocking content that violates OpenAI’s policies. Every image is also stamped with C2PA metadata — a watermark that helps platforms identify AI-generated media and adds a layer of transparency.
OpenAI’s expansion into API-driven image generation is a strategic move in its ongoing generative AI arms race with Google. Just weeks ago, Google showed off Imagen 3, the latest version of its own image model, during Cloud Next 2025.
At the same time, OpenAI rolled out new text models, including GPT-4.1 and a new reasoning-focused line of models called OpenAI o3, promising better performance in coding, math, and visual tasks.
OpenAI’s move to open up gpt-image-1 to developers signals a clear intent to make its most powerful tools available to builders everywhere and dominate the conversations around AI image generation.
