Lambda Pricing Calculator - Savings Plans & Provisioned Concurrency

Compute Savings Plans discount Lambda duration charges by ~12% with a 1 or 3-year commitment. Request charges ($0.20/million) are not covered. The same Savings Plan also applies to EC2, Fargate, and other compute services, making it flexible for mixed workloads.

Provisioned Concurrency keeps Lambda functions initialized and ready, eliminating cold starts. You pay an hourly allocation cost plus reduced execution charges, and neither one draws from the AWS Free Tier, so costs start from the first second. This calculator applies the free tier to Provisioned Concurrency duration the same way it does for on-demand duration, which slightly understates the Provisioned Concurrency total versus AWS's own billing rules; treat that result as a lower bound. It's ideal for latency-sensitive APIs where consistent sub-100ms response times are required. See our complete guide on provisioned concurrency costs to decide if it's worth it for your workload.

Use Provisioned Concurrency when cold starts impact user experience, such as synchronous APIs, real-time applications, or payment processing. For async workloads like queues or scheduled jobs, standard concurrency is more cost-effective.

Lambda costs = (requests × $0.20 per 1M) + (GB-seconds × architecture rate). GB-seconds are calculated by memory (in GB) × duration (in seconds) × total invocations. ARM/Graviton2 is 20% cheaper than x86 per GB-second. See our complete Lambda pricing breakdown for the full rate tables, tiered discounts, and hidden costs like CloudWatch Logs and cold starts.

A Lambda function cost calculator estimates your monthly bill based on request charges, duration charges, and optional costs like Provisioned Concurrency and ephemeral storage. This calculator also factors in Compute Savings Plans discounts and the always-free tier. It uses a flat tier-1 duration rate and doesn't add CloudWatch Logs, SnapStart, Managed Instances, or Durable Functions costs, so treat the result as your requests-and-duration baseline, not a full bill.

AWS offers 1 million free requests and 400,000 GB-seconds of compute time per month, plus 100 GB of data transfer out. This free tier doesn't expire after 12 months - it's always free. One exception: the Provisioned Concurrency allocation charge and the duration it serves are never free-tier eligible, even though your request allowance still applies.