# Amazon Bedrock cost tips: cache TTL, Flex tier, thinking effort, and maxTokens

> AWS explains how four request‑level settings—prompt cache duration, service tier, reasoning effort, and maxTokens—change token usage and pricing on Bedrock.

Oossa · 2026-11-12 · https://oossa.com/en/amazon-bedrock-cost-tips-cache-ttl-flex-tier-thinking-effort-and-maxtokens

AWS’s blog shows how four Amazon Bedrock inference options affect your bill. Shorter cache TTLs, the Flex service tier, low‑effort reasoning, and a tighter maxTokens ceiling can cut token usage and rates. The post gives sample code, test results, and links to pricing tables so developers can measure the impact before rolling changes live.

## The facts

- The post was published on 2026-11-12.
- A 1‑hour cache becomes cheaper after the second request within an hour, and Flex tier on Nova 2 Lite costs half the Standard rate.

## Why it matters

Developers can lower their generative‑AI spend on Bedrock by tuning these four settings instead of switching models.

## Sources & references

1. [Four Amazon Bedrock inference choices that affect cost](https://aws.amazon.com/blogs/machine-learning/four-amazon-bedrock-inference-choices-that-affect-cost/) – AWS, 2026-11-12

Last updated: 2026-11-12