|  
                              
                               
                              | Main differences are: 
 # There is no /models endpoint when creating an Azure OpenAI deployment. There is only direct access to the chat/completions endpoint. The Azure chat/completions endpoint is, however, directly compatible with the corresponding OpenAI Endpoint, so not differences there.
 # The Azure OpenAI API requires a query parameter that specifies the API version. So every request needs to be suffixed with for example "?api-version=2023-05-15"
 
 |  |