Today I’m thrilled to share that we’re releasing our latest image gen model (same model powering ChatGPT) in the API. The model’s versatility allows it to create images across diverse styles, faithfully follow custom guidelines, leverage world knowledge, and accurately render text—unlocking countless practical applications across multiple domains.
The model is available as gpt-image-1 and with pricing for text and image tokens:
Text input tokens (prompt text): $5 per 1M tokens
Image input tokens (input images): $10 per 1M tokens
Image output tokens (generated images): $40 per 1M tokens
In practice, this translates to roughly $0.02, $0.07, and $0.19 per generated image for low, medium, and high-quality square images, respectively. For detailed pricing by image quality and size, see thedocs.
Getting Started
We’ve published a bunch of content to help you get started:
A gallery showing [a tiny smattering of] the ways this model can be used
A blog highlighting some of the ways early partners have integrated the model
Updated API reference & SDK. Note the controls you have over moderation, quality, size, output format, and background transparency
To reduce abuse, access to this modelrequiresverifying your organization. You can check your verification status in the developer console and if you haven’t verified already, it should only take a couple minutes. We’ll also be adding this model to our Responses API soon. We’ll also continue supporting DALL-E in the API.
You can tell that “moderation” as a new parameter was meant to be the verification gate, and then they just said “screw it, we need more to agree to this personal invasion”.
This edits tool utility I made should be prepped for use already - just changing the name of the global variables at the start of the file from a blank string to the new model name will get you buttons for the new model to generate or edit images. You can try it out for me - as I simply will not agree nor submit and the result of the verify organization policy will thus be zero support.
@jeffsharris I’m not able to verify our org. We’re usage tier 4 and I’m seeing this in the settings console.
I didn’t have a chance to submit any docs. I simply clicked the “Verify Organization” CTA and it immediately (almost no delay) was replaced with “Organization could not be verified”. See screenshot here:
I completed the verification process but I’m still getting an error about needing to be verified back from the API. Is there a lag or delay in the verification system?
@markhurlburt@theli0nheart@admin136 Team’s looking into the verification issues now. (It does take about ~30 min after completing verification for it to take effect on an existing API key.)
I will NEVER comply with forced ID or organization verification just to access APIs—especially for something like image generation. This is a massive breach of trust, a blatant disregard for user privacy, and an unnecessary security risk.
Tying creative and development tools to invasive verification methods is exactly how you push devs like me straight into the arms of your competition and the growing open-source ecosystem. If OpenAI continues down this path of gating innovation behind identity harvesting, you’re not “reducing abuse,” you’re alienating the very community that helped build your success.
Plenty of us value freedom, privacy, and control over our workflows more than shiny new models. Keep this up, and I (along with many others) will gladly move to alternatives that don’t treat developers like potential criminals by default.
I have my org verified (and have access to the o3 model) but the new image generation model is not showing up in the list of accessible models. How can I get access? Thanks!
I don’t work with or for OpenAI but I like how they are handling the situation. It makes sense to add some friction to the process, elseways this would get out of hands.
Just chill and enjoy whatever else you find out there, wish you all the best!
@Bittabola@markhurlburt@admin136 Can you try now? Looks like there’s a caching delay right now (although it was immediate yesterday!), and image gen should be available ~45 mins after completing verification.
