Bedrock Adds Grok 4.7 in AWS GovCloud (US)

Amazon Web Services has made SpaceXAI’s Grok 4.7 available on Amazon Bedrock in AWS GovCloud (US), its isolated cloud regions for US government agencies and regulated customers. In a What’s New announcement on 19 October 2026, AWS described Grok 4.7 as “a frontier model built for coding, agentic tasks, and knowledge work”.

The move puts SpaceXAI’s latest model in front of public-sector and regulated buyers who usually need workloads to stay inside GovCloud. It follows the model’s arrival in Bedrock’s commercial regions earlier this autumn.

In brief

Grok 4.7 is now available on Amazon Bedrock in AWS GovCloud (US), according to AWS’s 19 October announcement.
AWS says it builds on Grok 4.6 with better mixed-document handling, more dependable repo-scale coding and enhanced browser-use agents.
Customers can start in the Bedrock console or through supported Bedrock APIs; the announcement gives no GovCloud pricing.

What AWS says is new

According to AWS, Grok 4.7 “builds on Grok 4.6 with better mixed-document handling, more dependable repo-scale coding with planning and error recovery, and enhanced browser-use agents for form fills and portal navigation.” That last point is relevant to government work, where so many tasks still run through web forms and online portals.

AWS says customers can get started in the Amazon Bedrock console or programmatically through supported Bedrock APIs. For regions, endpoints, features, inference profiles and pricing, the announcement points to the Bedrock documentation; it does not quote a price itself.

How the model is packaged

The Bedrock model card lists a context window of 524,288 tokens shared between input and output, text and image input, and configurable reasoning effort at four levels: low, medium, high and xhigh. It lists the GovCloud (US-East) and GovCloud (US-West) regions alongside Bedrock’s commercial regions.

Grok 4.7 first reached Amazon Bedrock on 28 September 2026, according to an AWS machine learning blog post. That post says the model supports the Responses, Chat Completions, InvokeModel and Converse APIs and is served through cross-region inference profiles. It also notes that independent benchmarker Artificial Analysis found the gains come with roughly double the output tokens per task compared with Grok 4.6, a cost trade-off buyers should plan for.

Our read

GovCloud availability is a procurement step more than a technical one, but it matters: it lets agencies and contractors trial SpaceXAI’s newest model without leaving their approved cloud boundary. It also widens the choice of frontier models on government-grade infrastructure. Watch for agency pilots, and for how its browser-agent features fare against strict public-sector security rules.