Tts-1 & tts-1-hd API RPM and RPD based on chosen tier

Dear OpenAI Community,

I hope this message finds you well. I am reaching out to inquire about the rate limits and scaling options for the Text-to-Speech (TTS) API provided by OpenAI.

Firstly, I would like to understand the daily rate limits (RPD) and requests per minute (RPM) for both the TTS-1 and TTS-1-HD APIs across all tiers. Could you please provide clarification on these limits?

Secondly, in the event that I exceed the rate limits for a specific tier, such as tier 1, I am curious to know if there is an automatic scaling mechanism in place to upgrade to the next tier (e.g., tier 2) to avoid rate limit errors. Alternatively, would I need to manually adjust the tier settings on the OpenAI platform to accommodate increased demand?

Thank you for your time and assistance. I look forward to your response.

Best regards from Croatia,
Dominik Kajba

Tier is based on how much your organization has previously paid.

The limit for tier 1, having prepaid the minimum, is only by requests per minute. Make more requests than the limit (or call them all the same second) you will get API errors.

Here’s a link to the limits, with more documentation linked about the rates:

You can use your own strategy for managing the number of API calls you make to stay below what you are allowed.

1 Like

Dear OpenAI Community,

I appreciate the information provided regarding the TTS API rate limits and scaling options. I have a few follow-up questions to ensure I fully understand how to manage our usage effectively:

  1. Does the tier system require us to pay upfront for the tier that best suits our business needs?

  2. I received information from chatGPT (as you can see in attachment below) that the daily rate limit (RPD) is calculated by multiplying the requests per minute (RPM) by the number of minutes in a day (1440). For instance, with a Tier 5 RPM limit of 500, the RPD would be 720,000 requests per day. Is this calculation correct?

  3. What strategies or best practices would you recommend for organizing our OpenAI environment to avoid exceeding the rate limits and encountering errors?

Thank you for your assistance. I look forward to your response.

Best regards from Croatia,
Dominik Kajba