For a month I cannot connect anymore using AZURE.AI.OPENAI
I updated my package to the latest 2.0

And either I have an error that it can not find a model, either it can not connect.

Setup is quite very simple :

builder.Services.AddSingleton<OpenAIClient>(_ => new OpenAIClient("OpenAI"));

Config is

  "ConnectionStrings": {
    "OpenAI": "Key=sk-proj-Mykry"
  }

And here an example of a ChatCompletion method

        [HttpPost("translate-text")]
        public ChatCompletion TranslateText(TranslationRequest request)
        {
            var client = _openAiClient.GetChatClient(DefaultModel);
            var messages = new List<ChatMessage>();

            messages.Add($"Describe this image. Language should be {request.TargetLanguage}.");

            if (request.Prompts != null)
            {
                messages.Add($"Translate the following text from {request.SourceLanguage} to {request.TargetLanguage} =>");
                messages.Add(new OpenAI.Chat.UserChatMessage(string.Join("n", request.Prompts)));
            }

            var options = BuildBaseOptions((OpenAiBaseRequest)request);
            messages.AddRange(GetSystemMessages("You are a helpful creative translator.", request.SystemMessages, request.ResponseFormat));

            var results = client.CompleteChatAsync(messages, options);
            // Incorrect API key provided: OpenAI. You can find your API key at 
            // https://platform.openai.com/account/api-keys.)'
            return results.Result.Value;
        }

The returns stuck as API Key provided is Incorrect, I already tried to instantiate openclient directly with the API key, I deleted and recreated 10 times my keys, the last ones were project keys. But no way… And it was working perfectly before the launch of GPT4-o.

Did any one had the same issue ? or have an idea what is going wrong ?