After deploying GPT 4 mini in my Azure foundry i am tiring to test it but getting this issue. What might be the reason

Question

Answer 1

Hey there!

It looks like you’ve successfully deployed gpt-4.1-mini in your Azure AI Foundry/Studio environment but when you try a simple prompt (“What is AI?”) you get:

“Error An unexpected error occurred. Check your model's config and try resubmitting the request.”

That generic Playground error usually means the request never made it to the model inference backend. Here are the most common culprits and how to check them:

Deployment still provisioning or unhealthy

1. In the Foundry portal (or AI Studio), go to Models + Endpoints → your gpt-4.1-mini deployment.
2. Verify the status is “Succeeded” (not “Creating” or “Failed”).
3. If it’s still rolling out, wait a few minutes and retry.

Service depreciated on hub-based project (won't work beyond GPT-4o), Please switched to non-hub-based project if you are using foundry classic. Reference-

Out-of-quota

1. Some GPT-4 variants require extra quota.
2. In the Foundry portal, go to Quotas and confirm you have enough unused Provisioned Throughput Units (PTUs) for this deployment.

Mis-configured model settings

1. If you customized any advanced settings in your deployment (model ID, temperature, max tokens, etc.), double-check the JSON config.
2. A bad parameter can cause the service to immediately reject requests.

Follow up

Have you got Cognitive Service OpenAI contributor access

Have you tried adjusting max_token from model configuration once.

What is current region. Have you tested GPT 4 mini deployment in new foundry UI in any other region

What is your set TPM and max token used in model side.

Please share your temperature