How is AWS Lambda billed exactly?

Lambda bills in GB-seconds plus a per-request charge. The classic surprise: 10M invocations at 200ms with 256MB costs about $8.50 in compute plus $2.00 in requests — but 100 functions like that is $1,050/month. The calculator reproduces the GB-second formula with current pricing so the real number is visible.

Why does each Lambda invocation look free but add up?

A single invocation at 256MB × 200ms is fractions of a cent, which is why serverless feels free in a demo. Multiply by millions of invocations across many functions and the compute + request charges compound into real money. The calculator projects monthly cost from your invocation volume and duration, so the architecture decision uses the annual number.

What are the best Lambda cost optimization strategies?

Reduce execution duration (cold-start tuning, larger memory can paradoxically cut cost by finishing faster), minimize request count with batching, and consolidate cold-start-prone functions. The calculator's optimization section quantifies each lever — duration, memory, invocation volume — so you prioritize the one that moves your bill.

Tool Hub

📰📡

Home · JSON · Calculators · Articles · About

λ Serverless Lambda Compute Cost Estimator

Lambda bills in GB-seconds. Each invocation looks free — $0.0000166667 per GB-second is too small to feel. At 10M invocations/month with 200ms avg duration and 256MB, you're at $10.50/month per function. 100 functions. The decimal place shifts fast. Model it before your CFO asks why serverless costs more than EC2.

📐 Lambda Workload Profile

Configure invocation volume, avg duration (ms), and memory (MB). The engine computes GB-seconds at $0.0000166667/GB-s plus $0.20/million requests. Small unit cost, big multiplier effect.

Monthly Executions (Millions)

Average Duration (ms)

Memory Allocated (MB)

Pricing Region

AWS Lambda Pricing: GB-Seconds + Request Charges

AWS Lambda employs a two-part pricing model that charges separately for compute duration and invocation count. The compute charge is based on GB-seconds — the product of memory allocated (in GB) and execution duration (in seconds). The request charge is a flat $0.20 per million invocations, regardless of function size or duration. This decoupled model means that short, memory-light functions can be extremely cheap, while long-running, high-memory functions can accrue significant costs even at modest invocation volumes.

How This Calculator Works

Component	Formula	Rate (US East)
Compute Cost	`Invocations × Duration(s) × Memory(GB) × $0.0000166667`	$0.0000166667 per GB-second
Request Cost	`Invocations × $0.20 / 1,000,000`	$0.20 per million requests
Total Cost	`Compute Cost + Request Cost`	—

The GB-Second Formula in Practice

A GB-second is the fundamental unit of Lambda compute billing. A function configured with 1,024 MB (1 GB) of memory running for 1 second consumes exactly 1 GB-second. For a function with 512 MB running for 200 ms, each invocation consumes 0.5 GB × 0.2 s = 0.1 GB-seconds. At 5 million invocations per month, that is 500,000 GB-seconds. Multiplied by the $0.0000166667 rate, the pure compute cost is approximately $8.33. The request charge adds another $1.00 (5 × $0.20), totaling roughly $9.33 per month for a moderate-volume API backend.

Cost Optimization Strategies

Lambda costs scale linearly with duration and memory, but not independently. Increasing memory also increases vCPU allocation proportionally (1,769 MB = 1 vCPU), which can reduce execution duration for CPU-bound workloads. This creates a non-obvious optimization surface: a 256 MB function running for 400 ms costs the same in GB-seconds as a 512 MB function running for 200 ms, but the latter may provide lower latency. Use AWS Compute Optimizer or Lambda Power Tuning to find the cost-latency Pareto frontier for your specific function payloads.

Pricing Basis, Sources & Assumptions

Every rate on this page is a published vendor list price — no negotiated discounts, private pricing, or credit offsets. Totals are in USD and exclude tax. Rates were last checked against the sources below on 2026-09-23.

Pricing input	Basis used on this page
Region	us-east-1 (N. Virginia) unless a pricing region is selected — the selector applies published regional multipliers, which are highest in ap-northeast-1 and lowest in us-east-1.
Currency	USD — on-demand list price, tax excluded
Last checked	2026-09-23 — The compute ($0.0000166667 per GB-second) and request ($0.20 per million) rates are the us-east-1 list prices cited on the linked page.

Modelling assumptions

Compute: $0.0000166667 per GB-second. Duration is billed rounded up to the nearest millisecond, and the billed memory is the memory you configure, not the memory used.
Requests: $0.20 per million invocations.
A full month is 730 hours; the free tier (400,000 GB-seconds and 1 million requests per month) is not applied.
Architecture (x86 vs Graviton) is not a variable here — the x86 rate is used.

What this model excludes

Provisioned concurrency, which bills for pre-warmed capacity whether it serves traffic or not.
Ephemeral storage beyond the free 512 MB, and CloudWatch Logs ingest at $0.50/GB plus $0.03 per GB-month of storage — usually the second-largest line on a Lambda bill.
Data transfer, Step Functions transitions, and Lambda Extensions.
Tax, credits, compute Savings Plans, and private pricing.

Sources

AWS (2026). "AWS Lambda Pricing." Per-GB-second compute, per-million-request, and provisioned concurrency rates. aws.amazon.com
AWS (2026). "Lambda Execution Environment Lifecycle." Cold-start mechanics and execution-environment reuse. docs.aws.amazon.com
AWS (2026). "Amazon CloudWatch Logs Pricing." Log ingest and storage rates — usually the second line on a Lambda bill. aws.amazon.com

Vendor list prices change without notice — re-check the linked pages before committing spend. jslet takes no vendor sponsorship and carries no affiliate links; see about.