Difference between Tokens per Minute Rate Limit (thousands) & Rate limit (Tokens per minute)

Question

In a GPT-Vision model version deployment in Switzerland North region, I noticed the following specification:

• Tokens per Minute Rate Limit (thousands): 10
• Rate limit (Tokens per minute): 30000

While I understand from this matrix that the default quota limit for gpt-4 (vision-preview)
GPT-4 Turbo with Vision is 30K TPM, I don't understand the specification of 'Tokens per Minute Rate Limit (thousands): 10'.

Please help me understand these specifications better.

Answer 1

Hey Parimesh Panda

The specs you mentioned to rate limiting refers for API usage in the GPT model deployment. and lemme clear that up

Rate Limit (Tokens per minute): 30,000

• This represents the maximum number of tokens that the model can process per minute. In your case, it's set to 30,000 tokens per minute for the GPT-4 Turbo with Vision.

Tokens per Minute Rate Limit (thousands): 10

• This value, expressed in thousands, is an alternative representation of the rate limit. In this case, it's set to 10,000 tokens per minute (10 * 1,000).

So both specs convey the same information but are presented in different units . It's common to express large numbers in terms of thousands or millions for readability. So answering to your question for both the out come is the same and only the units is differently calculated

If this helps kindly accept the answer thanks much;