AWS Bedrock Token Cost Calculator
See how much you'll spend on LLM inference — pick a model, enter your token volume, get monthly costs in seconds.
Which Bedrock model are you using?
Select the model you're running inference on. Prices are per 1,000 tokens (input/output may differ).
What's your daily inference volume?
Enter your daily API calls and average token counts per request.
Prompt size: 100 tokens = ~75 words
Completion size: 100 tokens = ~75 words
Your Estimated Monthly Cost
Based on your token volume and selected model
Monthly Bedrock Cost
Cost Breakdown
Cost per 1M input tokens: $3.00
Cost per 1M output tokens: $15.00
Daily cost: ~$X.XX
Cost Optimization Tips
- Batch requests: Larger prompts = lower cost per output token
- Model selection matters: Nova Lite is 20–50× cheaper than Claude Opus
- Cache prompts: Bedrock prompt caching reduces input costs 90% for repeated context
- Monitor usage: Set up CloudWatch budgets to catch cost spikes early
See per-token costs for your model and strategies to reduce spend — free.
Full breakdown on its way!
Check your inbox for cost optimization strategies.
Detailed Cost Analysis
Ready to Scale Bedrock in Production?
FactualMinds helps enterprises optimize Bedrock costs, implement prompt caching, and ensure safe LLM guardrails in production.
Explore Bedrock Services →Who This Tool Is For
Engineering teams, data scientists, and product managers evaluating Bedrock for production workloads. If you're building with Claude, Nova, or Llama on Bedrock and need to forecast costs, this tool gives you quick, accurate estimates.
Why We Built This Tool
Bedrock pricing can be opaque when you're doing capacity planning. This calculator lets you plug in your token volume and get an instant monthly cost estimate. No algebra, no spreadsheets.
What Problem It Solves
- Cost forecasting. Know upfront what your Bedrock bill will be based on token volume and model choice.
- Model comparison. See how much more Nova costs vs. Claude, or how much cheaper Haiku is.
- Budget planning. Allocate costs accurately without over- or under-estimating LLM spend.
- Right-sizing decisions. Understand the trade-off between model capability and cost per token.
Learn more about our AWS Bedrock consulting services to optimize your inference architecture.
Frequently Asked Questions
Does Bedrock have a free tier?
No free tier, but you can use AWS free tier credits if you're a new account or have AWS Activate credits. After that, all inference is pay-per-token.
What about context window tokens?
This calculator assumes standard 200K context for Claude and Nova. Large context windows use the same per-token pricing but may require pre-processing for very large prompts.
Does prompt caching save money?
Yes. Cached input tokens cost 90% less than regular input tokens. If you repeat the same context (system prompts, documents), caching dramatically reduces costs.
What about batch processing?
Bedrock has a Batch API with 50% discount on input token costs. Use it for non-real-time workloads where latency isn't critical.