Automated Updates for GPT Models API

Question

Hi All,

We are using multiple GPT models, now when new version come and old one retire it's lot of change management , is there automated way to update all models and any best practices to avoid manual interventions in GPT models upgrades.

Thanks

Answer 1

Yep - managing multiple GPT model versions in Azure OpenAI can be streamlined through automation and best practices to minimize manual intervention during upgrades.

For starters, consider setting your deployments to Auto-update to default or Upgrade when expired to automatically update to use the new default version.

More at https://learn.microsoft.com/en-us/azure/ai-services/openai/how-to/working-with-models?tabs=powershell

In addition, instead of directly referencing specific model versions (e.g., gpt-4-0613), use consistent deployment names in your API calls. For example, deploy gpt-4-turbo with an alias like my-default-gpt4 and reference this alias in your applications.

If you prefer a more controlled approach, you can automate model version updates via Azure CLI & APIs. For example, you can run a scheduled Azure Function or Logic App to query available models and update deployments automatically. This will allow you deploy the new version alongside the old one for testing (A/B testing) and use a traffic-splitting strategy where a small percentage of requests go to the new model before full migration.

---

If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.

hth

Marcin