Hi I live in the UK and have written a webapp for a client in Holland. It uses GPT-4 and GPT-3 and have set up the appropriate deployments for both. I need to send the largest number of tokens as possible as the app is summarising data. Something seems to have changed over the last few days and I am struggling with GPT-4. Which model should I be using for GPT-4? I want to use my clients account and models so they can pay. Is there a GPT-4 model available to use in Holland? At the moment my webapp (which was working fine) seems not to work for GPT4. It works locally for GPT-4. Am I missing something? Thanks! Justin

Justin Baron Greetings & Welcome to Microsoft Q&A forum! What is the exact issue you are facing GPT-4 model? Please see below answers to your queries. Which model should I be using for GPT-4? It depends on your requirement. The default quota for models varies by model and region. Default quota limits are subject to change. Quota for standard deployments is described in of terms of Tokens-Per-Minute (TPM) . Choose from the above list and use it accordingly based on token limits. I want to use my clients account and models so they can pay. Is there a GPT-4 model available to use in Holland? Please check the model availability . If is available you can deploy and use. For using your client's account and models, you can use Azure OpenAI's multi-tenancy feature to allow your client to pay for the usage of the models. This allows you to use your client's subscription and models while still being able to manage the usage and billing. At the moment my webapp (which was working fine) seems not to work for GPT4. It works locally for GPT-4. Am I missing something? Can you share more details on what is not working? it's difficult to say without more information. It's possible that there may be an issue with the deployment or configuration of the model. I recommend checking the logs and error messages to see if there are any clues as to what may be causing the issue. Do let me know if you have any further queries.

Which model version should I be using for my openai deployment

Justin Baron 0

I live in the UK and have written a webapp for a client in Holland. It uses GPT-4 and GPT-3 and have set up the appropriate deployments for both. I need to send the largest number of tokens as possible as the app is summarising data. Something seems to have changed over the last few days and I am struggling with GPT-4.

Which model should I be using for GPT-4?
I want to use my clients account and models so they can pay. Is there a GPT-4 model available to use in Holland?

At the moment my webapp (which was working fine) seems not to work for GPT4. It works locally for GPT-4. Am I missing something?

Thanks!
Justin

1 answer

AshokPeddakotla-MSFT 27,401 Reputation points

2024-04-19T11:25:02.07+00:00

Justin Baron Greetings & Welcome to Microsoft Q&A forum!

What is the exact issue you are facing GPT-4 model?

Please see below answers to your queries.

Which model should I be using for GPT-4?

It depends on your requirement. The default quota for models varies by model and region. Default quota limits are subject to change.

Quota for standard deployments is described in of terms of Tokens-Per-Minute (TPM).

Choose from the above list and use it accordingly based on token limits.

I want to use my clients account and models so they can pay. Is there a GPT-4 model available to use in Holland?

Please check the model availability. If is available you can deploy and use.

For using your client's account and models, you can use Azure OpenAI's multi-tenancy feature to allow your client to pay for the usage of the models. This allows you to use your client's subscription and models while still being able to manage the usage and billing.

At the moment my webapp (which was working fine) seems not to work for GPT4. It works locally for GPT-4. Am I missing something?

Can you share more details on what is not working? it's difficult to say without more information. It's possible that there may be an issue with the deployment or configuration of the model.

I recommend checking the logs and error messages to see if there are any clues as to what may be causing the issue.

Do let me know if you have any further queries.
Please sign in to rate this answer.
Justin Baron 0 Reputation points

2024-04-19T11:40:25.85+00:00

Many thanks for getting back to me so quickly.

I couldn't see a GPT-4 model for Holland however I am pretty sure my client was able to set one up - hence the confusion.

For my deployment in the UK, as I mentioned I am looking for the most tokens that is reliable for GPT-4 (GPT-3 works well). It is the model version for the UK I need help with.

So my two questiona are:

Am I using the best GPT4 version for me in the UK requireing the largest token upload (app summarises documents)

Is GPT-4 really not available to people in Holland?

I will get back to you with the exact issue I am facing when I have done a bit more analysis.

I recommend checking the logs and error messages to see if there are any clues as to what may be causing the issue

Can you point me to where I find these?

Thanks

Justin

AshokPeddakotla-MSFT 27,401 Reputation points

2024-04-21T10:55:10.0933333+00:00

Justin Baron Please see below for more information.

Am I using the best GPT4 version for me in the UK requireing the largest token upload (app summarises documents)

To give more context, GPT-4 Turbo Preview has a max context window of 128,000 tokens and can generate 4,096 output tokens. It has the latest training data with knowledge up to April 2023. This model is in preview and is not recommended for production use.

All deployments of this preview model will be automatically updated in place once the stable release becomes available.

Here are some of the features it provides:

This model offers lower pricing, extended prompt length, tool use, and structured JSON formatting, delivering improved efficiency and control.

GPT-4 Turbo is more capable and has knowledge of world events up to April 2023. It has a 128K context window so your applications benefit from a lot more custom data tailored to your use case using techniques like RAG (Retrieval Augmented Generation).

GPT-4 Turbo pricing is 3x most cost effective for input tokens and 2x more cost effective for output tokens compared to GPT-4, while offering more than 15x the context window. To deploy GPT-4 Turbo from the Studio UI, select "gpt-4" and then select version "1106-preview" in the version dropdown. Version 1106-preview has separate quota from the existing versions of GPT-4, enabling customers to start experimenting with it immediately without impacting existing GPT-4 deployments.

GPT-4 version 0125-preview is an updated version of the GPT-4 Turbo preview previously released as version 1106-preview. GPT-4 version 0125-preview completes tasks such as code generation more completely compared to gpt-4-1106-preview. Because of this, depending on the task, customers may find that GPT-4-0125-preview generates more output compared to the gpt-4-1106-preview. We recommend customers compare the outputs of the new model. GPT-4-0125-preview also addresses bugs in gpt-4-1106-preview with UTF-8 handling for non-English languages.

Is GPT-4 really not available to people in Holland?

AFAIK, as I mentioned earlier, model availability varies per region. Currently, GPT-4 is not available in the Holland region.

I will get back to you with the exact issue I am facing when I have done a bit more analysis. Can you point me to where I find these?

Please see Monitoring Azure OpenAI Service and Implement logging and monitoring for Azure OpenAI models for more details.

Do let me know if you have any further queries.

If the response helped, please do click Accept Answer and Yes for was this answer helpful.

Doing so would help other community members with similar issue identify the solution. I highly appreciate your contribution to the community.
Sign in to comment