Track, allocate, and manage your generative AI cost and usage with Amazon Bedrock

This article discusses Amazon Bedrock's new capability to tag and track the costs of using its on-demand foundation models. Application inference profiles allow organizations to tag and monitor costs based on organizational taxonomies like cost centers, departments, and applications.

Specifically, the article covers:

• Challenges of managing generative AI costs as usage scales across projects and business units
• Introducing Amazon Bedrock application inference profiles for tagging on-demand models
• Differences between system-defined and application inference profiles
• Creating and managing application inference profiles using Bedrock APIs
• Using tags with cost management tools like AWS Budgets, Cost Explorer, and CloudWatch
• Methods for dynamically retrieving inference profile ARNs based on tags
• Overview of Bedrock's multi-account resource tagging capabilities