A quilting-shop app team wants to call several foundation models through an API and does not want to stand up training clusters. Which AWS service matches this job?
Select an answer to reveal the explanation.
Short Explanation
Think of calling several foundation models through an API with no training cluster. That is Amazon Bedrock. SageMaker is the custom kitchen, Quick is the wall chart, and invented services are off the exam.
Full Explanation
Amazon Bedrock is the managed way to consume foundation models through an API without standing up training clusters. A quilting-shop app team that wants to call several FMs and avoid cluster operations is describing Bedrock’s job: select models, invoke them, and leave host lifecycle to the service. That is the exam’s standard match for managed FM consumption.
Amazon SageMaker AI is the custom train-and-host path; claiming every prompt requires a custom training cluster invents a requirement this team rejected. Amazon Quick is AI-assisted analytics and dashboards on business data, not the FM invocation layer for an application. An invented AWS training service that does not appear on the exam list cannot be correct; stay with named offerings the blueprint expects.
Exam caveat: Bedrock still requires you to handle prompts, data shared with the model, and application IAM—managed consumption is not unmanaged content risk. Operational check: confirm the app’s IAM roles allow only intended model access and regions, verify no unnecessary training cluster was opened for a pure prompt/API workload, and document which models the shop is allowed to call.
If the team later needs custom training, that is a separate decision and a different service path; do not collapse prompt-time FM calls into an assumed training cluster. For this item, success is invoking managed models quickly while the quilting-shop app stays free of host babysitting.