Assistance Required – Issues with Azure OpenAI Image Edit API (gpt-image-1.5 and gpt-image-1.0)

Question

Dear Microsoft Azure Support Team,

We are currently integrating Azure AI models into our application and have encountered issues specifically with the Image Edit API. Below are the details:

gpt-image-1.5

• The images/generations endpoint works as expected.
  • However, when calling the images/edits endpoint, we consistently receive a 404 Resource Not Found error, despite following the official documentation and sample code.
• gpt-image-1.0
• The images/generations endpoint functions properly.
  • When invoking the images/edits endpoint, the request is rejected with the following error:
    “Your request was rejected by the safety system. If you believe this is an error, contact us at Please contact support via the official support channel and include the request ID [REQUEST_ID].”

Interestingly, the Image Edit functionality works correctly in the Azure AI Playground, but fails when we attempt to implement it programmatically in our code. This discrepancy makes it difficult to proceed with integration.

We kindly request your assistance in investigating:

1. Why the images/edits endpoint returns 404 for gpt-image-1.5.
2. Why the safety system rejects edit requests for gpt-image-1.0, even when the same prompt works in the Playground.
3. Whether there are alternative models or recommended deployments that support image editing functionality reliably, in case gpt-image-1.5 or gpt-image-1.0 have limitations.

Please let us know if additional information (such as request payloads, headers, or logs) is required for troubleshooting. We would greatly appreciate guidance on best practices for using the Image Edit API in production.

Thank you for your support and assistance.

Best regards,
Surya

Answer 1

Hello Mahendra, Talari Surya (Cognizant),

Welcome to Microsoft Q&A and Thank you for the questions.

You’re currently seeing two separate behaviors:

• gpt-image-1.5 → 404 error on /images/edits
• gpt-image-1.0 → Request blocked by safety system
• Playground → Image edit works
• API → Fails

These are not caused by a single issue, but rather by model capability differences and safety enforcement behavior in Azure OpenAI APIs.

1. gpt-image-1.5 - 404 on /images/edits

The 404 error you are encountering when calling the /images/edits endpoint with gpt-image-1.5 is expected with the current service behavior.

At present:

• gpt-image-1.5 supports image generation (/images/generations)
• It does not support image editing (/images/edits) via API

When an unsupported operation is invoked for a model, the service returns:

404 Resource Not Found, since that route is not enabled for that model

Why it appears to work in Playground

The Azure AI Playground can sometimes give the impression that a feature is supported because:

• It may internally use a different model or fallback configuration
• It may route requests through internal pathways not exposed publicly via API

So while the Playground is useful for testing, it does not always reflect exact API-level capabilities

2. gpt-image-1.0 - Safety rejection on edit requests

For gpt-image-1.0, the /images/edits endpoint is supported. However, your requests are being rejected by the content safety system.

The error 'Your request was rejected by the safety system…'

indicates that either the prompt, The input image or the combination of both

is being flagged under Azure’s safety policies.

Why it works in Playground but fails in API

This is a common point of confusion. The difference is due to how requests are handled:

In Playground

Prompts may be pre-processed, Slightly modified or normalized

Some guardrails may be Adjusted for testing scenarios

In API

Requests are evaluated as-is

Safety filters are applied strictly and consistently

So even if the same prompt is used, the actual payload reaching the model may differ slightly, leading to different outcomes.

Common reasons for safety rejection in image edits

Even prompts that seem valid can be blocked if they involve:

Editing images of people or faces

Modifying appearance, identity, or body features

Content that could be interpreted as:

• Sensitive
• Medical
• Biometric-related

Additionally, The input image itself is also evaluated, not just the prompt

Recommendations

For gpt-image-1.5

• Use it only for image generation
• Avoid using /images/edits as it is not supported currently

For gpt-image-1.0

To reduce safety rejections:

Keep prompts neutral and descriptive

Avoid References to specific individuals, Sensitive transformations (face, body, identity changes)

Start with Simple, non-sensitive images.

Please refer this

Call the Image Edit API (Python, REST, Java, PowerShell) https://learn.microsoft.com/azure/foundry/openai/how-to/dall-e?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#call-the-image-edit-api

Understanding Image Content Filtering https://learn.microsoft.com/azure/ai-services/openai/concepts/content-filter?tabs=definitions%2Cuser-prompt%2Cpython-new#image-content

I Hope this helps. Do let me know if you have any further queries.

---

If this answers your query, please do click Accept Answer and Yes for was this answer helpful.

Thank you!