Bedrock Guardrails grounding checks add Standard tier
Amazon Bedrock's document history records that on 19 Oct 2026 Guardrails contextual grounding checks gained support for the Standard safeguard tier. AWS says Standard handles responses that include reasoning or conversational content, returns more consistent results, and supports more languages than the earlier tier.
Facts read on .
On 19 Oct 2026, the Amazon Bedrock User Guide document history added an entry titled "Contextual grounding checks support safeguard tiers." Contextual grounding checks are the part of Bedrock Guardrails that filters hallucinations in model responses.
AWS says the Standard tier supports responses that include reasoning or conversational content, returns more consistent results and supports additional languages. The entry points to the contextual grounding documentation and to the page on safeguard tiers for guardrails policies.
For teams that use the built-in guardrail on Bedrock, the change affects output filtering at the model layer. AWS does not publish a decision latency or a fail-open behavior in this entry, so a team moving to the Standard tier should measure both in its own workload.
Sources
- Amazon Bedrock User Guide: Document history (docs.aws.amazon.com), read