Flash Cards · Implementation

Flash Card: Provisioned Throughput vs On-Demand

August 02, 2026 · 3 min read

Exam-style

A media company's Bedrock summarisation service handles roughly 2 million tokens an hour, all day and every day, and the product team has committed to a p99 latency target. The per-token bill keeps climbing. What should they move to?

Reveal the answer

D. Purchase Provisioned Throughput model units for the model, sized to the sustained hourly volume

Steady, high, round-the-clock volume with a latency commitment is what Provisioned Throughput is priced for: reserved model units give predictable throughput and consistent latency for a committed hourly rate. Backoff manages throttling but does nothing for the latency target or the bill. Savings Plans cover compute services and do not apply to Bedrock model inference. Batch inference is cheaper but asynchronous, so it cannot serve a p99 target, and cross-Region inference profiles spread load without reserving any capacity.

Generative AI Development · part of The Exam Room

Q. Steady high-volume Bedrock traffic with latency guarantees. Provisioned Throughput or on-demand?

A. Provisioned Throughput: reserved Model unitThe billing block Provisioned Throughput is sold in – one unit delivers a fixed tokens-per-minute rate for a specific model. for predictable, high, steady load and consistent latency (and required for some custom or imported models). On-demand suits spiky or low volume, paying per token with no commitment.

Why? Match the pricing model to the traffic shape: steady-and-high favours provisioned.

These posts are LLM-aided. Backbone, original writing, and structure by Craig. Research and editing by Craig + LLM. Proof-reading by Craig.