Amazon Bedrock Custom Model Import introduces real-time cost transparency
Amazon Bedrock Custom Model Import enables customers to import and run their customized foundation models on-demand without managing the underlying infrastructure. Customers can now get full transparency into the compute resources being used and calculate inference costs in real-time.
With this launch, customers are able to see the minimum compute resources, custom model units (CMUs), required to run their model prior to model invocation in the Bedrock console and through Bedrock APIs. As the model scales to handle more traffic, CloudWatch metrics provide real-time visibility into the inference costs by showing the total number of CMUs used. This enables customers to better control spend through near-instant cost visibility and take actions like making on-the fly model configuration changes to optimize for costs.
This feature is available in all regions where Amazon Bedrock Custom Model Import is supported. To learn more and get started, visit the Amazon Bedrock Custom Model Import page and see the documentation page for more details.