ACorrect
The number of input and output tokens processed
Correct. On-demand use of Bedrock is fundamentally pay-as-you-go pricing based on the number of input tokens and output tokens processed. You pay only for what you use.
A company is preparing a cost estimate for using Amazon Bedrock. For on-demand use of Bedrock, what is pricing mainly based on? Which option is MOST appropriate?
1 / 1
Select an answer
CorrectA
Explanation
Choose the billing basis for Bedrock on-demand use.
- 1「For on-demand use of Bedrock, what is pricing mainly based on」Asks for the basis of billing = input/output token count
BIncorrect
The running time of a provisioned endpoint
Time-based pricing reflects the idea of a reserved usage model such as Provisioned Throughput.
On-demand use is billed by the number of tokens for what you use, so this is incorrect.
CIncorrect
The GPU hours used to train the model
GPU hours reflect the cost idea of training a model yourself.
On-demand inference on Bedrock does not involve training, and billing is by input/output token count, so this is incorrect.
DIncorrect
The number of stored prompts
The number of stored prompts is not the billing basis for on-demand inference.
Billing is based on the number of input and output tokens processed, so this is incorrect.
Key Takeaway
On-demand use of Amazon Bedrock is fundamentally pay-as-you-go pricing based on the number of input tokens and output tokens processed. Longer prompts and longer responses use more tokens and cost more. To control cost, be mindful of the token volume of prompts and outputs.