Amazon Bedrock Guardrails now offers a Standard tier for contextual grounding checks. The new tier is more robust than the Classic tier and can filter hallucinations in conversational responses that reason over a source. It scores differently worded answers with the same meaning more consistently and supports more languages. Pricing is unchanged and the tier can be selected per guardrail in the console or API. It is available in twelve AWS regions, including US East (N. Virginia) and Europe (Frankfurt).
Why it matters
Developers can reduce AI hallucinations in chat‑based apps without extra cost, improving trust for end users.
