OpenAI: Can't deploy model despite having quota

Question

Hi,

I currently have an OpenAI resource that has GPT-4-vision deployed. On the OpenAI playground I have the following deployments.

Region: West US

Deployed model name: gpt-4

Deployed model version: vision-preview

However, if I want to deploy any other GPT-4 model, I am facing this issue where it says "no quota is available.

But when I check under the Quotas tab, it seems like I have not reached the usage limit.

This problem also occurs with deploying GPT-3.5.

Is there a hidden dependency that I am not aware of?

Thank you.

Answer 1

Hello @johananmahendran,

Based on the information provided, the current deployment type you are utilizing is Provision-managed, which requires the purchase of PTUs prior to use. For more details, please refer to this Microsoft Learn document.

The document lists all the regions where the models are available. In your case, in West US, you can use gpt-4-0125-preview. You simply need to adjust the deployment type to Standard.

Best regards,
Charlie

---

If you find my response helpful, please consider accepting this answer and voting 'yes' to support the community. Thank you!