Amazon Bedrock now offers Global cross-Region inference (CRIS) for Anthropic Claude models (Opus 4.6, Sonnet 4.6, Haiku 4.5, and earlier variants) to customers in India using the Mumbai (ap-south-1) and Hyderabad (ap-south-2) regions. Global CRIS routes inference requests across all commercial AWS regions to handle traffic bursts and peak demand seasons common in India (Diwali, tax season, cricket tournaments). The post covers how to use global inference profile IDs with InvokeModel, Converse, and streaming APIs, configure IAM permissions, and monitor requests via CloudWatch Gen AI Observability and CloudTrail event data stores to track which destination region processed each request.
Table of contents
Core functionality of Global cross-Region inferenceUnderstanding inference profilesImplementing global cross-Region inferenceGlobal cross-Region inference for India’s peak demand seasonsStep by step guidance for getting started with inferencing with Global cross-Region inference on Amazon Bedrock:Code samples to invoke the Anthropic Claude model with global cross-Region inference with different API typesMonitoring and logging with Global cross-Region inferenceTake your AI applications globalAbout the authorsSort: