Resource Not found when using Azure Open Ai Model o1

[GodierLuna]

Hello,

I am trying to setup a service to access AzureOpenAI but no matter what i get : {"error":{"code":"404","message": "Resource not found"}}

I have tried to put the endpoint and/or the resourcename and deployment name but it always fails.

Belo is a mock of my code.

Does any body have an idea??

I would be so thankful!

Best,

Luna

application.yaml

quarkus:
  langchain4j:
    azure-openai:
      log-responses: true
      chat-model:
        deployment-name: "o1"
        resource-name: "resource"
        api-key: "key"
        api-version: "2025-01-01"

and my service

import jakarta.enterprise.context.ApplicationScoped;

import dev.langchain4j.service.SystemMessage;
import dev.langchain4j.service.UserMessage;
import io.quarkiverse.langchain4j.RegisterAiService;

@RegisterAiService
@SystemMessage("my message")
@ApplicationScoped
public interface AzureOpenAiServiceModel {

    @UserMessage("{prompt}. Respond in the following JSON format: {jsonTemplate}.")
    String queryAzureOpenAi(String prompt, String jsonTemplate);

}

[geoand]

cc @edeandrea

[sberyozkin]

@GodierLuna I did a quick test with https://github.com/quarkiverse/quarkus-langchain4j/tree/main/samples/secure-poem-multiple-models, it works, for example, using the old turbo model:

quarkus.langchain4j.azure-openai.openai.resource-name=quarkus-langchain4j
quarkus.langchain4j.azure-openai.openai.deployment-name=gpt-35-turbo
quarkus.langchain4j.azure-openai.openai.api-version=2024-12-01-preview

where quarkus-langchain4j is Azure OpenAI resource name, where model deployments are created

You should change your configuration as follows:

quarkus:
  langchain4j:
    azure-openai:
      log-responses: true
      chat-model:
        resource-name: "your_azure_openai_resource"
        deployment-name: "o1"
        api-key: "key"
        api-version: "2025-01-01"

[sberyozkin]

@GodierLuna Hmm, never mind, this is what you already do, it looks fine, assuming your deployment is called o1. I actually can't deploy due the quota limitation so I can't verify, I guess the other thing to check is where this model is supported in the resource's location, West Europe vs US East etc