lambda cost calculator

How Lambda pricing works

Lambda pricing is usually driven by two things: how many times your function runs and how much compute time it consumes. Unlike fixed servers, you pay as usage happens. This makes Lambda great for bursty or event-driven workloads, but it also means cost can climb quickly when request volume or execution duration spikes.

1) Request charges

Each invocation counts as a request. AWS commonly prices this at $0.20 per 1 million requests (region-specific). If your workload sends millions of events from API Gateway, SQS, EventBridge, or cron jobs, this line item becomes meaningful.

2) Compute charges (GB-seconds)

Compute cost depends on allocated memory and runtime duration. Lambda measures this as GB-seconds:

• Duration in seconds = milliseconds / 1000
• Memory in GB = MB / 1024
• GB-seconds = Requests × Duration × Memory

The calculator multiplies total GB-seconds by a compute rate based on architecture. Arm/Graviton is usually cheaper than x86 for many workloads.

Formula used in this calculator

Without free tier

• Request Cost = (Requests / 1,000,000) × $0.20
• Compute Cost = GB-seconds × Compute Rate
• Total Monthly Cost = Request Cost + Compute Cost

With free tier enabled

• Billable Requests = max(0, Requests - 1,000,000)
• Billable GB-seconds = max(0, GB-seconds - 400,000)
• Costs are calculated only on the billable portion

Practical ways to lower Lambda cost

Right-size memory

More memory increases cost per millisecond, but it can reduce runtime significantly. Test multiple memory levels. Often, a moderate increase in memory decreases duration enough to lower total cost.

Reduce cold-start impact

Use lightweight dependencies, optimize initialization code, and avoid unnecessary package bloat. Faster startup reduces billed duration for intermittent traffic patterns.

Optimize external calls

Slow database queries and network calls are common cost drivers. Add caching, improve query plans, and minimize retries. Every unnecessary millisecond gets billed.

Use Arm where compatible

If your runtime and dependencies support it, Arm often provides lower GB-second pricing. The calculator lets you compare cost quickly by changing architecture.

Common estimation mistakes

• Ignoring duration variance (p95/p99 can be much higher than average).
• Forgetting retries from upstream services.
• Assuming dev traffic patterns match production peaks.
• Not accounting for batch size in stream or queue triggers.
• Treating one region’s rate as universal for all regions.

Final note

This tool is a strong planning estimator, especially during architecture and budget discussions. For exact billing, verify your AWS region pricing, workload behavior, and any additional features you use (such as provisioned concurrency, ephemeral storage, or data transfer). Use this calculator to get quick directional clarity before deeper cost modeling.