- 最新
- 最多得票
- 最多評論
Hi,
With no special action, you will be charged with the "on-demand" model: "provisioned throughput" mode needs some very specific actions on your side describe in service documentation.
Note: to avoid billing surprise, you can quite precisely estimate your costs with on-demand mode by leveraging the input and output token counts present in the response http header metadata of your Bedrock inference.
You can access them this way (Python)
response = await client.invoke_model(
body=body,
modelId=model_id,
accept="application/json",
contentType="application/json"
)
response['ResponseMetadata']['HTTPHeaders']['x-amzn-bedrock-input-token-count']
response['ResponseMetadata']['HTTPHeaders']['x-amzn-bedrock-output-token-count']
Best,
Didier
Hello.
There is no charge if you simply enable model access in the Amazon Bedrock console.
You will only be charged for actually using that model.
If you only want to enable model access, you will be charged an on-demand fee.
This means that once you use the model, you only pay for what you use.
I think what you are talking about as a monthly fee is probably the fee for a monthly contract for provisioned throughput.
This will not be a monthly fee unless you set it up using the steps in the document below.
https://docs.aws.amazon.com/bedrock/latest/userguide/prov-throughput.html
Hi, The pricing model on the marketplace indicates different pricing with the same label, starting from 0€ to more. There seems to be missing a dimension in the pricing table, does the pricing vary based on total use for instance ? "Output Images 1024 x 1024, 50 steps or less $0 / unit" "Output Images 1024 x 1024, 50 steps or less $0.04 / unit" "Output Images 1024 x 1024, more than 50 steps $0.08 / unit" => which one will apply ? Link: https://aws.amazon.com/marketplace/pp/prodview-a54dufkbweqgs?ref_=aws-mp-console-subscription-detail-payg#pricing-information
相關內容
- 已提問 7 個月前
- 已提問 2 年前
