Predict your AWS AI bill before you build it

Pick a reference architecture, set your workload in plain terms, and get a monthly estimate broken down by service — Bedrock model tokens, AgentCore Runtime and Memory, OpenSearch, Lambda and more. Every rate shows its math and verification date.

Reference architectures

Amazon Bedrock model pricing

Estimate the token cost of any Amazon Nova model on Bedrock, before you build.

Example: $288.00/month

Bedrock AgentCore agent cost

Predict a Bedrock AgentCore agent bill: model tokens + Runtime + Memory + optional tools.

Example: $778.00/month

RAG chatbot on Bedrock

Estimate a retrieval-augmented chatbot: OpenSearch vectors + Bedrock generation + Lambda.

Example: $593.32/month

Batch document pipeline

Estimate a high-volume document extraction/summarisation pipeline with batch pricing.

Example: $738.55/month

Voice agent on AWS

Estimate a voice bot: Transcribe (STT) + Bedrock + Polly (TTS) + Lambda.

Example: $3,989.11/month

Minimal serverless Q&A

The cheapest starting point: API Gateway + Lambda + Bedrock, no vector store.

Example: $137.01/month

What this covers (and what it doesn't)

Covered: Amazon Bedrock inference, Bedrock AgentCore, and the managed services these architectures use — OpenSearch Serverless, Lambda, S3, DynamoDB, API Gateway, Textract, Transcribe and Polly.

Deferred (on purpose): raw EC2/GPU compute, per-region compute pricing, and committed-use or savings-plan discounts. Those are too volatile to keep honest, so we leave them out rather than mislead you.

AgentCore and managed-service rates last verified 14 July 2026 (us-east-1). Bedrock model prices refresh daily. Estimates are for planning, not a quote.