'Unknown model: chatgpt' error in Azure AI Foundry's chat playground

Question

A Deepseek R1 model has been deployed in the chat playground, and custom data was added in Azure AI Search. When attempting to add the Azure AI search index to the model via the "Add your data" option in the chat playground, the following error occurs for any prompt:

An error occurred when calling Azure OpenAI: Server responded with status 400. Error message: {'error': {'code': 'unknown_model', 'message': 'Unknown model: chatgpt', 'details': 'Unknown model: chatgpt'}}

The Data and Indexes were created without error, and the integration works when selecting the gpt-4o-mini model in the chat playground. What could be the reasons for the error message encountered? It works correctly when the my own data is removed.

Answer 1

Based on this old thread :

You will need to verify model availability and check model configuration, especially your data configuration to know the exact name and mode of Llama. If you are sure everything is okay or using a custom model or a specific version, ensure that all references to the model are updated and consistent across your configuration. Because the data index setup is referencing an invalid model name (chatgpt).

By explicitly specifying the correct Llama deployment name and ensuring compatibility with search types, you bypass the incorrect default reference. If Keyword search is unsupported for Llama, switching to Vector search. Also, refer to the Azure AI Foundry documentation for any specific requirements or limitations related to the Llama-3.3-70B-Instruct-2 model and data indexing - https://learn.microsoft.com/en-us/azure/ai-studio/how-to/deploy-models-llama

Answer 2

Hi @Samir Pathak,

The issue likely occurs because the "Add your data" feature is designed to work with OpenAI models like GPT-4o-mini and may not be fully compatible with custom models such as Deepseek R1. Since Deepseek R1 is a different LLM, it may not support the same data augmentation capabilities that OpenAI models do.

Additionally, the "Add your data" option may not be available for Deepseek R1 in Azure AI Foundry, meaning that when attempting to integrate Azure AI Search, the system may not be available to index data with the model. This could explain why the integration works correctly with GPT-4o-mini but fails when using Deepseek R1.

I hope you understand. And, if you have any further query do let us know.

---

If this answers your query, do click Accept Answer and Yes for was this answer helpful.