AWS announced on September 28, 2026, that xAI’s Grok 4.7 is available through Amazon Bedrock. Access uses cross-Region inference profiles, and AWS lists separate Standard rates for its US Geo and Global profiles. xAI had announced the model on September 21.
What Grok 4.7 supports on Bedrock
The Bedrock model card lists a 500,000-token context window. Grok 4.7 accepts text and image input and produces text output. Reasoning is active by default, with high as the default effort setting; you can also choose low, medium, or xhigh.
Geo and Global inference profiles
Amazon Bedrock requires a cross-Region inference profile for Grok 4.7; in-Region inference is not supported. The US Geo profile is us.xai.grok-4.7, and the Global profile is global.xai.grok-4.7.
AWS lists these on-demand Standard rates per 1 million tokens:
| Inference profile | Profile ID | Input rate per 1 million tokens | Output rate per 1 million tokens | Cache-read rate per 1 million tokens |
| US Geo cross-Region | us.xai.grok-4.7 | $2.20 | $6.60 | $0.55 |
| Global cross-Region | global.xai.grok-4.7 | $2.00 | $6.00 | $0.50 |
AWS prices the Priority tier at 1.75× the Standard rate and Flex at 0.5× the Standard rate. The Reserved tier is unsupported for this model.
APIs and Bedrock features
Grok 4.7 supports the Responses, Chat Completions, Converse, and Invoke APIs on bedrock-runtime. AWS also lists invocation logs, response streaming, implicit prompt caching, structured outputs, reasoning, and Guardrails as supported Bedrock features.
Application inference profiles work with Invoke and Converse, but not with Responses or Chat Completions. Bedrock does not support server-side tool use, intelligent prompt routing, or token counting for Grok 4.7.
US Regions listed for Grok 4.7
AWS lists both US Geo and Global cross-Region inference in four US Regions: N. Virginia (us-east-1), Ohio (us-east-2), N. California (us-west-1), and Oregon (us-west-2).
