AWS Security ChangesHomeSearch

AWS bedrock: Document safeguard tier support for contextual grounding checks

Service: bedrock · 2026-10-21 · Documentation low

File: bedrock/latest/userguide/guardrails-tiers.md

Summary

Adds contextual grounding checks to the list of tier-aware policies, describes Standard tier behavior (reasoning/conversational responses, more consistent scores), adds a comparison-table row, adds guidance on when to choose Standard vs Classic tier, and lists the Regions where Standard tier for contextual grounding checks is available (other Regions fail with a validation error).

Security assessment

The page documents Amazon Bedrock Guardrails safeguard tiers, a security feature that includes prompt-attack protection and harmful-content filtering, and the edit clarifies tier capabilities and Regional availability. It is descriptive guidance about an existing security feature rather than a fix for a specific weakness, so it is security documentation of low-to-moderate impact.

Evidence

+Provides more robust performance compared to Classic tier and has more comprehensive language and code-related prompt support. For example, protection against prompt attacks performs more consistently and reliably with Standard tier. Guardrails with Standard tier also use [cross-Region inference](./guardrails-cross-region.html). Standard tier supported with content filters and denied topics provides protection against harmful content introduced within code elements including comments, variable and function names, and string literals. For contextual grounding checks, Standard tier supports responses that include reasoning or conversational content and returns more consistent scores for differently worded responses with the same meaning.

Diff

diff --git a/bedrock/latest/userguide/guardrails-tiers.md b/bedrock/latest/userguide/guardrails-tiers.md
index 2c73504..8333454 100644
--- a//bedrock/latest/userguide/guardrails-tiers.md
+++ b//bedrock/latest/userguide/guardrails-tiers.md
@@ -20,0 +21,2 @@ The following guardrails policies support safeguard tiers:
+  * [Contextual grounding checks](./guardrails-contextual-grounding-check.html)
+
@@ -31 +33 @@ Amazon Bedrock Guardrails provides the following safeguard tiers:
-Provides more robust performance compared to Classic tier and has more comprehensive language and code-related prompt support. For example, protection against prompt attacks performs more consistently and reliably with Standard tier. Guardrails with Standard tier also use [cross-Region inference](./guardrails-cross-region.html). Standard tier supported with content filters and denied topics provides protection against harmful content introduced within code elements including comments, variable and function names, and string literals.
+Provides more robust performance compared to Classic tier and has more comprehensive language and code-related prompt support. For example, protection against prompt attacks performs more consistently and reliably with Standard tier. Guardrails with Standard tier also use [cross-Region inference](./guardrails-cross-region.html). Standard tier supported with content filters and denied topics provides protection against harmful content introduced within code elements including comments, variable and function names, and string literals. For contextual grounding checks, Standard tier supports responses that include reasoning or conversational content and returns more consistent scores for differently worded responses with the same meaning.
@@ -45,0 +48 @@ Denied topics | Maximum of 1,000 characters per definition | Maximum of 200 char
+Contextual grounding checks | More robust than Classic tier. Supports responses that include reasoning or conversational content. | Established performance. Strict grounding only.  
@@ -59 +62,3 @@ For example, consider Standard tier when:
-  * You need higher accuracy and performance for content filters, prompt attacks, and denied topics
+  * You need higher accuracy and performance for content filters, prompt attacks, denied topics, and contextual grounding checks
+
+  * Strict grounding blocks too many valid responses because your application's responses summarize, reason over, or converse about the grounding source rather than restate it
@@ -69,0 +75,2 @@ Alternatively, you might still use Classic tier when:
+  * Your contextual grounding thresholds are tuned for Classic tier and have not been re-tested against Standard tier
+
@@ -140,0 +148,29 @@ Safeguard tiers are supported in the following [AWS Regions](./guardrails-suppor
+Standard tier for contextual grounding checks is available in the following Regions:
+
+  * US East (N. Virginia)
+
+  * US East (Ohio)
+
+  * US West (N. California)
+
+  * US West (Oregon)
+
+  * Europe (Frankfurt)
+
+  * Europe (Ireland)
+
+  * Europe (Milan)
+
+  * Europe (Paris)
+
+  * Europe (Spain)
+
+  * Europe (Stockholm)
+
+  * Asia Pacific (Tokyo)
+
+
+
+
+Requests that specify Standard tier for contextual grounding checks in any other Region fail with a validation error.
+