Amazon Bedrock has released an open source Model Profiler to help users navigate its extensive catalog of foundation models. The tool aggregates metadata from multiple AWS APIs and external sources into a single interface, enabling teams to compare models based on capabilities, pricing, and regional availability. This simplifies the process of selecting the right model for specific use cases, reducing the time spent on manual research and decision-making.
The Model Profiler is a web application that allows users to browse, filter, and compare all foundation models available on Amazon Bedrock. It provides detailed model cards, regional availability maps, and pricing breakdowns updated daily. The tool sources data from seven distinct inputs, including five AWS APIs and two public URLs, ensuring the catalog remains accurate without manual intervention. This automated pipeline collects and processes data to provide users with a comprehensive view of model specifications and constraints.
The Model Profiler operates through a fully automated serverless pipeline orchestrated by AWS Step Functions. It uses seventeen AWS Lambda functions to process data across four phases, with inter-Lambda S3 caching reducing API calls from approximately 480 to 29 per execution. The pipeline completes in 8–12 minutes and runs daily at 6 AM UTC, with a configurable schedule through the Amazon EventBridge rule in the AWS CloudFormation template.
Source: awsml
