Foundry endpoint not working

Question

Foundry endpoint not working

Carlos Albarran 0

Hello:

I activated the following foundry endpoint:

https://REDACTED.openai.azure.com/openai/deployments/Kimi-K2.5/chat/completions?api-version=2024-02-01

https://ai.azure.com/nextgen/r/[RESOURCE_ID_REMOVED]

However, I am getting the following error. What is missing? I activated everything... kindly asking for your help.

{"error":{"code":"404","message": "Resource not found"}}

Note: PII Redacted at support side.

Anshika Varshney 9,740 Reputation points Microsoft External Staff Moderator

2026-04-07T08:58:31.6833333+00:00

Hi Carlos Albarran,

When a Foundry endpoint does not work, it is usually due to a small setup or configuration issue rather than a bigger problem.

First, please check that the model deployment is running properly in Azure AI Foundry. If the deployment did not complete or is stopped, the endpoint will not respond.

Next, make sure you are using the correct endpoint URL. It is best to copy the endpoint again directly from the Foundry portal, as using a project or old endpoint by mistake is very common.

Also verify the authentication details. If the key was regenerated or the identity was changed recently, the old credentials will no longer work. Updating the key or identity usually fixes this.

Another thing to check is the API version. Using an unsupported or older API version can cause the endpoint to fail. Trying with the latest supported version often resolves the issue.

If networking is restricted, confirm that the calling service has permission to reach the endpoint. Private or restricted network settings can block requests.

To narrow it down, try sending a simple test request directly from the portal or a basic REST call. This helps confirm whether the issue is with the endpoint itself or the application code.

Most of the time, once the endpoint, authentication, and API version are aligned, the Foundry endpoint starts working as expected.

Official references that may help

Azure AI Foundry endpoints overview https://dori-uw-1.kuma-moon.com/azure/ai-studio/concepts/endpoints

Deploy and manage models in Azure AI Foundry https://dori-uw-1.kuma-moon.com/azure/ai-studio/how-to/deploy-models

Authentication for Azure AI services https://dori-uw-1.kuma-moon.com/azure/ai-services/authentication

Hope this helps, and feel free to share what you observe after trying these steps and Do let me know if you have any further queries.Thankyou!
Rayyan Fawad 0 Reputation points

2026-04-12T11:40:39.6233333+00:00

A 404 Resource Not Found error in Azure OpenAI or Azure AI Foundry typically means the endpoint URL is incorrectly formatted or targeting a deployment name that doesn't exist at that specific base URL.
Deleted

This comment has been deleted due to a violation of our Code of Conduct. The comment was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.

1 answer

Your answer

Anshika Varshney 9,740 Reputation points Microsoft External Staff Moderator

2026-04-07T08:58:31.6833333+00:00

Hi Carlos Albarran,

When a Foundry endpoint does not work, it is usually due to a small setup or configuration issue rather than a bigger problem.

First, please check that the model deployment is running properly in Azure AI Foundry. If the deployment did not complete or is stopped, the endpoint will not respond.

Next, make sure you are using the correct endpoint URL. It is best to copy the endpoint again directly from the Foundry portal, as using a project or old endpoint by mistake is very common.

Also verify the authentication details. If the key was regenerated or the identity was changed recently, the old credentials will no longer work. Updating the key or identity usually fixes this.

Another thing to check is the API version. Using an unsupported or older API version can cause the endpoint to fail. Trying with the latest supported version often resolves the issue.

If networking is restricted, confirm that the calling service has permission to reach the endpoint. Private or restricted network settings can block requests.

To narrow it down, try sending a simple test request directly from the portal or a basic REST call. This helps confirm whether the issue is with the endpoint itself or the application code.

Most of the time, once the endpoint, authentication, and API version are aligned, the Foundry endpoint starts working as expected.

Official references that may help

Azure AI Foundry endpoints overview https://dori-uw-1.kuma-moon.com/azure/ai-studio/concepts/endpoints

Deploy and manage models in Azure AI Foundry https://dori-uw-1.kuma-moon.com/azure/ai-studio/how-to/deploy-models

Authentication for Azure AI services https://dori-uw-1.kuma-moon.com/azure/ai-services/authentication

Hope this helps, and feel free to share what you observe after trying these steps and Do let me know if you have any further queries.Thankyou!
Rayyan Fawad 0 Reputation points

2026-04-12T11:40:39.6233333+00:00

A 404 Resource Not Found error in Azure OpenAI or Azure AI Foundry typically means the endpoint URL is incorrectly formatted or targeting a deployment name that doesn't exist at that specific base URL.
Deleted

This comment has been deleted due to a violation of our Code of Conduct. The comment was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.

Answer 1

Hi Carlos Albarran,

Endpoint syntax does not seem to be correct.

I just tested below code copied from deployment/code with Owner (Sub inherited)/Azure AI user (RG inherited) and it is working for me.

User's image

from openai import OpenAI

endpoint = "https://<pleaserefereploymentcode>.openai.azure.com/openai/v1/" deployment_name = "Kimi-K2.5"
api_key = "<your-api-key>"

client = OpenAI(
    base_url=endpoint,
    api_key=api_key
)

completion = client.chat.completions.create(
    model=deployment_name,
    messages=[
        {
            "role": "user",
            "content": "What is the capital of France?",
        }
    ],
)

print(completion.choices[0].message)

Output

 ChatCompletionMessage(content='The capital of France is **Paris**.', refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=None, reasoning_content='The user is asking a very simple factual question: "What is the capital of France?" This is a straightforward question with a clear, unambiguous answer. The capital of France is Paris.\n\n I should provide a direct, concise answer. I don\'t need to overcomplicate this with unnecessary details unless the user asks for more information, but a brief, friendly confirmation is appropriate.\n\n Answer: Paris.')

Attached screenshot for reference

User's image

Note:

Found a discrepancy in endpoint syntax creation though. Created a feedback item and checking with PG team internally on it.

Please copy the code suggested in deployment as workaround


endpoint = "https://<project-name>-resource.openai.azure.com/openai/v1/

Please let us know the proposed approach helps address the issue

Thank you.

Manas Mohanty 16,190 Reputation points Microsoft External Staff Moderator

2026-04-16T01:14:36.3133333+00:00

Hi Carlos Albarran,

Could you share your observation with above approach.

Thank you.

Share via

Foundry endpoint not working

1 answer

Your answer