📢 New Model Announcement: GPT-image-1 Now in Azure AI Foundry (Limited Access Preview)

[nicholasdbrady]

We're excited to share that GPT-image-1, OpenAI's latest and most advanced image generation model, is now available in limited access preview on Azure OpenAI Service.

---

🔍 What's New

GPT-image-1 builds upon the capabilities of DALL·E, introducing significant enhancements:

• Improved Instruction Following: Generates images that more accurately reflect detailed prompts.
• Reliable Text Rendering: Capable of rendering text within images, useful for educational materials and storybooks.
• Image Input Support: Accepts images as input, enabling image-to-image transformations and inpainting.
  These features expand the model's versatility for various creative and professional applications.

🛠 Capabilities

GPT-image-1 supports multiple functionalities:

• Text-to-Image: Generate images from text prompts.
• Image-to-Image: Transform existing images using text prompts.
• Inpainting: Edit specific areas within images using bounding boxes and text.
  These capabilities allow for a wide range of creative and practical uses, from designing user interfaces to developing game assets.

🌍 Availability

GPT-image-1 is currently available in the following Azure regions:

• West US 3 (Global Standard)
• UAE North (Global Standard)
  Access is limited and requires an application.

🚀 Getting Started

To request access:

1. Apply via the GPT-image-1 Limited Access Application](https://aka.ms/oai/gptimage1access).
2. Once approved, deploy the model in your resource in westus3 or uaenorth
3. Refer to the Image Generation How-To Guide

---

Feel free to share your experiences or ask questions below. Happy building!

AI prompt: A bustling metallurgical foundry is portrayed in an isometric illustration. The scene features developers operating computers, typing into keyboards, speaking into microphones and overseeing an "ai agent factory" creation process. The environment is filled with datacenter elements such as long rows of server racks, neatly organized networking cables, and large cooling apparatuses.​ Above the main workspace, a prominent banner stretches across the foundry floor, displaying the words "Happy Building" in bold, industrial-style lettering. The sign adds a touch of positivity and motivation amidst the intense industrial setting.

[arafattehsin]

I like these crisp details (both in the image and the post 😀) - thanks for sharing.

[nrobert]

Thanks @nicholasdbrady , it's the 1st place where I finally got the details about the regions availability.
And I can confirm that it's now visible under Azure AI Foundry, yes!!

[nrobert]

@nicholasdbrady , trying to go deeper this morning, it brings me 2 additional questions that might be interesting to add when publishing a new model:

• can it be used in AI Foundry playgrounds? (looks like it can't at the moment for GPT-image-1, am I right?)
• does it needs a new version of the inference API (preview or GA)?

[arafattehsin]

@nrobert I found something for you: https://github.com/Azure-Samples/visionary-lab

[nicholasdbrady]

Great points!

• can it be used in AI Foundry playgrounds?

Coming tomorrow to the image playground 😊

[nrobert]

Great points!

• can it be used in AI Foundry playgrounds?

Coming tomorrow to the image playground 😊

Hi @nicholasdbrady , I had a quick look this morning but can't see a change in the image playground: do you know if the change has been released?

[nicholasdbrady]

The image playground is now live: https://devblogs.microsoft.com/foundry/images-playground-may-2025/

[CatStudioApp]

any news about the model update here?

gpt-image-1 recent has an update https://x.com/OpenAIDevs/status/1945538534884135132

[guygregory]

Hi @CatStudioApp, thanks for your interest in these enhancements to the gpt-image-1 model on Azure OpenAI. The team are currently working on this, so expect some news very soon. Watch this space!