Is a Copilot Plus license required to use custom models with Azure OpenAI endpoints?

[lefweh]

Hi everyone,

I'm encountering issues similar to those reported in issues #1247 and #1267. When trying to use custom models with Azure OpenAI endpoints, I receive errors such as "request aborted," "401," or "404." I've tried both Azure OpenAI and OpenAI formats in the model option "Provider."

Is a Copilot Plus license required to use custom models with Azure OpenAI endpoints? Any guidance on resolving these errors would be greatly appreciated.

Also, I am using the latest version of Copilot, 2.8.6 . Models tested were gpt-4o & Deepseek-R1

[logancyang]

No, adding custom model is a free feature.

Azure now has a confusing array of model deployment formats. Some like R1 can be accessed via an API key only, some require you to breakdown the deploy name, instance name etc. You need to know how to translate a working curl command into an openai format request. gpt-4o and R1 from Azure are both working in 2.8.6.

[lefweh]

Thanks for the quick reply. I marked your answer as correct since that is the case.

Please let me give you two examples to show where I am stuck. Both examples work with Postman & curl commands. Both are deployed in an Azure AI project as separate deployments. The Copilot plugin has been deleted and reinstalled, and all other plugins are disabled.

1. OpenAI Inference SDK API
Working curl command:
curl --location 'https://xx-ai-service12345.services.ai.azure.com/models/chat/completions?api-version=2024-05-01-preview'
--header 'Content-Type: application/json'
--header 'api-key: CORRECT API KEY, COPIED FROM AZURE'
--data '{
"messages": [{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "What are 3 things to visit in Seattle?"}],
"max_tokens": 1000,
"model": "DeepSeek-R1"
}'
Challenge: Options: "Add Custom Chat Model" --> Provider: "OpenAI Format". I am not able to add the body parameter "model": "DeepSeek-R1". Without the parameter, the request will not work.

Image of Copilot Config:

2. Azure OpenAI SDK
Working curl command:
curl --location 'https://xx-m7exkjln-southcentralus.openai.azure.com/openai/deployments/gpt-4o/chat/completions?api-version=2024-08-01-preview'
--header 'Content-Type: application/json'
--header 'api-key: CORRECT API KEY, COPIED FROM AZURE'
--data '{
"messages": [{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "What are 3 things to visit in Seattle?"}],
"max_tokens": 1000
}'

Challenge: "Add Custom Chat Model" --> Provider: "Azure OpenAI" The request gets aborted.

Image:

Any idea what my mistake is?

[lefweh]

Resolved, gpt-4o now working with exact same settings.