[BUG] Structured Output throws error from Server

[bharathm03]

Library name and version

Azure.AI.Inference

Describe the bug

When I set Custom JSON Response format for model output, server throwing error that value is not supported. Tried it with Deepseek and Llama 4 models didn't work

Same code working Azure.AI.OpenAI packages

Expected behavior

Model should return structured output based on schema provided

Actual behavior

server throwing error that value is not supported.

Reproduction Steps

Call DeepSeek model with ResponseFormat set

Environment

dotnet core 9

[jsquire]

Hi @bharathm03. It is not clear from the context of your question whether the error that you're observing is the result of service behavior or the client library. Please provide:

• The version of the package that you're using
• A code snippet that demonstrates use and how to reproduce
• The exact and full error message and stack trace.

[github-actions]

Hi @bharathm03. Thank you for opening this issue and giving us the opportunity to assist. To help our team better understand your issue and the details of your scenario please provide a response to the question asked above or the information requested above. This will help us more accurately address your issue.

[bharathm03]

hi @jsquire I'm using latest version of Microsoft.Extensions.AI.AzureAIInference --version 9.5.0-preview.1.25265.7 which internally uses Azure.AI.Inference

Same example given in the documentation here but with AzureAIInference
https://learn.microsoft.com/en-us/azure/ai-services/openai/how-to/structured-outputs?tabs=python-secure%2Cdotnet-entra-id&pivots=programming-language-csharp#getting-started

Exception:

Microsoft.Extensions.AI.LoggingChatClient: Error: GetResponseAsync failed.

Azure.RequestFailedException: response_format value as json_schema is enabled only for api versions 2024-08-01-preview and later
Status: 400 (BadRequest)
ErrorCode: BadRequest

Content:
{"error":{"code":"BadRequest","message":"response_format value as json_schema is enabled only for api versions 2024-08-01-preview and later"}}

I think AzureAIInference client is not correctly sending the expected API version.