Hi everyone,
I’m working with Azure OpenAI and trying to build a Logic Apps workflow that requires text completion/chat (for example, GPT-3.5 Turbo or GPT-4). My Azure OpenAI resource (dev-autodailer-openai, resource group arkserv-dev-rg) only lists old, deprecated models such as text-davinci-003 and several embedding models. I have run the following CLI commands:
text
az cognitiveservices account list-models --name dev-autodailer-openai --resource-group arkserv-dev-rg
The only available models are either legacy/completion (which are deprecated and disabled) or embedder models (text-embedding-3-small, text-embedding-3-large). Attempts to deploy text-davinci-003 and similar fail with "ServiceModelDeprecated". Trying to deploy GPT-4/GPT-3.5-turbo results in "DeploymentModelNotSupported" or the model does not appear at all.
My questions:
- How can I enable access to current completion/chat models (like GPT-3.5 Turbo, GPT-4, GPT-4o) for my Azure OpenAI resource?
Is there a process or special approval required for my subscription?
Are there regional or subscription limitations I should be aware of?
Has anyone successfully transitioned from deprecated legacy models to the newer/generally available chat/completion models, and how?
Relevant context:
Use case: Automating communications in Logic Apps, requires high-quality text completions.
Resource Name: dev-autodailer-openai
Resource Group: arkserv-dev-rg
I have read through the deprecation and update docs and model availability guidance.
Any insights or help with next steps would be appreciated! Thank you.Hi everyone,
I’m working with Azure OpenAI and trying to build a Logic Apps workflow that requires text completion/chat (for example, GPT-3.5 Turbo or GPT-4).
My Azure OpenAI resource (dev-autodailer-openai, resource group arkserv-dev-rg) only lists old, deprecated models such as text-davinci-003 and several embedding models.
I have run the following CLI commands:
text
az cognitiveservices account list-models --name dev-autodailer-openai --resource-group arkserv-dev-rg
The only available models are either legacy/completion (which are deprecated and disabled) or embedder models (text-embedding-3-small, text-embedding-3-large).
Attempts to deploy text-davinci-003 and similar fail with "ServiceModelDeprecated". Trying to deploy GPT-4/GPT-3.5-turbo results in "DeploymentModelNotSupported" or the model does not appear at all.
My questions:
How can I enable access to current completion/chat models (like GPT-3.5 Turbo, GPT-4, GPT-4o) for my Azure OpenAI resource?
Is there a process or special approval required for my subscription?
Are there regional or subscription limitations I should be aware of?
Has anyone successfully transitioned from deprecated legacy models to the newer/generally available chat/completion models, and how?
Relevant context:
Any insights or help with next steps would be appreciated!
Thank you.