Amazon Bedrock expands IAM principal cost allocation to the bedrock-mantle endpoint
AWS has extended IAM principal cost allocation to the bedrock-mantle endpoint in Amazon Bedrock, enabling granular tracking of inference costs by user, team, or project.
Amazon Bedrock now allows cost allocation by AWS Identity and Access Management (IAM) principals for model inference requests made through the bedrock-mantle endpoint. This feature, previously available for the bedrock-runtime endpoint, enables customers to assign inference costs to specific users, teams, projects, or applications. By tagging IAM users and roles with attributes such as team, project, or cost center, organizations can track spending more precisely.
Customers can activate these IAM principal tags in the AWS Billing and Cost Management console to analyze costs in AWS Cost Explorer or at the line-item level in the AWS Cost and Usage Report 2.0 (CUR 2.0). The feature supports filtering or grouping costs by tags in Cost Explorer, or including caller identity (IAM principal) allocation data in CUR 2.0 exports. This provides greater visibility into how inference costs are distributed across different parts of an organization.
The capability is available in all AWS Regions where the bedrock-mantle endpoint is operational. To implement, users must first activate IAM principal tags in the billing console, then configure Cost Explorer or CUR 2.0 to reflect these allocations. This follows the same process as the existing bedrock-runtime endpoint cost allocation feature.
Amazon Bedrock remains a fully managed service offering secure access to foundation models from leading AI providers. The expanded cost allocation feature for the bedrock-mantle endpoint aligns with AWS’s broader effort to provide granular cost tracking for generative AI workloads. Additional details are available in the AWS documentation under Using IAM principal for cost allocation and IAM principal attribution in Amazon Bedrock.