Updated endpoint guidance to recommend the bedrock-runtime endpoint: Updated the endpoint and API guidance across the Amazon Bedrock User Guide to recommend the bedrock-runtime endpoint for new applications, including for OpenAI- and Anthropic-compatible w

Amazon Bedrock Endpoints

Amazon Bedrock supports two endpoints for inference operations: bedrock-runtime and bedrock-mantle.

bedrock-runtime

• Recommended for: New applications, Bedrock-native APIs, Guardrails, intelligent prompt routing, cross-Region inference.
• APIs: InvokeModel, Converse, Chat Completions, Responses, Messages.
• Features: Structured outputs, synchronous requests, default project support.
• Quotas: Fixed throughput limits (RPM and TPM), request increases available.

bedrock-mantle

• Recommended for: Existing applications, server-side tool use, asynchronous inference, Projects, Workspaces.
• APIs: Responses, Chat Completions, Messages.
• Features: Server-side tool use, pre-configured tools, Projects, Workspaces.
• Quotas: Advanced scheduling, higher initial throughput limits, queueing for in-flight workloads.

What to do

• Choose bedrock-runtime for new applications and Bedrock-native features.
• Use bedrock-mantle for existing applications and advanced capabilities.
• Consider VPC interface endpoints to reduce data egress costs.

Source: AWS release notes

If you need further guidance on AWS, our experts are available at AWS@westloop.io. You may also reach us by submitting the Contact Us form.