Amazon Bedrock now supports SpaceXAI Grok 4.6 with Cross Region Inferencing
AWS now offers SpaceXAI’s Grok 4.6 on Amazon Bedrock with cross-region inference, enabling higher throughput and lower costs while supporting US data residency requirements.
Amazon Bedrock has added support for SpaceXAI’s Grok 4.6, a frontier model optimized for coding, agentic tasks, and knowledge work. Customers can deploy the model with cross-region inference, which automatically routes requests across multiple AWS Regions to improve throughput and reduce per-token costs. Two inference profiles are available: US Geo restricts processing to US-based regions for compliance, while Global allows broader access for maximum capacity during demand spikes.
The model operates via the bedrock-runtime endpoint and integrates with Responses, Chat Completions, and Converse APIs. Existing Bedrock account controls, including model invocation logging to Amazon S3 or CloudWatch Logs, CloudWatch metrics, and cost tracking in AWS Cost Explorer and the Cost and Usage Report, remain applicable. No additional configuration is required beyond standard Bedrock usage.
Cross-region inference for Grok 4.6 is available in all AWS Regions where Amazon Bedrock is offered. Customers can select the US Geo or Global profile based on their data residency and performance needs. The feature aims to balance scalability with compliance and cost efficiency.
To begin using Grok 4.6, customers should review the model card in the Amazon Bedrock User Guide. The model’s integration follows existing Bedrock workflows, ensuring consistency with other supported models.