Can OpenAI Models be deployed in private networks

David Colwell Currently, the Azure Open AI based models/endpoints cannot be deployed in a specific Vnet, only the scale settings can be set at this point while deploying the endpoint through REST API. Please see the privacy and security page on how Azure OpenAI is currently securing the requests and data.

Update: As an update on above, if you enable networking settings of your Open AI resource to use private endpoints it should effectively apply the restrictions based on the settings selected for your endpoints too. My original response was based on the ability to select a VNET during endpoint deployment. If you however use a different network setting on your resource then it should behave the same way as any other cognitive service resource while configuring the virtual networks.