My hate for the usage limits

API
1

I want the usage limit to be removed, but keep the roles like pro, they could add more models, while also giving you money, but please get rid of usage limits

1 Like
2

Agreed, I have a production outage window that continues to grow because I have exceeded usage limits without the ability to pay them back down to 0. No word from support for 5 days at this point.

3

Hey - how long did it end up taking you, im in this predicament now

4

Alright it looks like Openai seems as if they aren’t fussing with upping anyone’s usage limit in platform.Openai.com anymore (since MS acquisition).

What seems to the new process here is:

  • create an azure cloud acct/ use existing one
  • create a new subscription for your open ai service
  • turn on the azure openai service
  • apply via their form in azure
  • wait for application to process

Once that process goes through you can just adjust a couple of params to your Openai client/ lib function calls or api calls and you should be all set.

1 Like
5

What does this solution do exactly? Higher rate limits via Azure?

6

From what I could find online, yes. This should, following an acceptance from Openai enable you to have higher usage limits.

I would like to iterate here that I really haven’t found and great documentation around solving this and is my current approach.

Still awaiting a response from azure on my open ai activation

1 Like
7

Have you requested a limit increase? Their form has explicit Tiers now…

How rate limit increases work

Rate limits are segmented into distinct tiers, each corresponding to a specific spend on the API platform. Requests for rate limit increases are processed weekly, and the applicable rate limit tier is determined by your API spend/usage in the last month. The limits are categorized under two types, Chat (gpt-3.5-turbo) and Embedding (text-embedding-ada-002). For DALL-E and Whisper, please submit the rate limit increase form, and we’ll assess the need dependent on usage. Below are the specifics on each tier, with spend referring to the completed payment on the API platform:

Tier 1:

To qualify for this tier, a minimum spend of $50 is necessary. For chat models, your rate limit will increase to 350K TPM and 3K RPM. For embedding models, the rate limit will increase to 500K TPM and 3K RPM.

Tier 2:

To qualify for this tier, a minimum spend of $150 is necessary. For chat models, your rate limit will increase to 600K TPM and 6K RPM. For embedding models, the rate limit will increase to 1M TPM and 12K RPM.

Tier 3:

To qualify for this tier, a minimum spend of $500 is necessary. For chat models, your rate limit will increase to 800K TPM and 10K RPM. For embedding models, the rate limit will increase to 2M TPM and 12K RPM.

Tier 4:

To qualify for this tier, a minimum spend of $1500 is necessary. For chat models, your rate limit will increase to 1M TPM and 12K RPM. For embedding models, the rate limit will increase to 5M TPM and 12K RPM.

We hope this breakdown will help you understand the requirements and benefits at each tier. If you have any further queries, feel free to contact us. If you need a further rate limit increase, please contact sales.

*Tiers subject to change.

Are you Tier 4?