Step Functions cost: Standard vs Express and the 100x difference

Quick answer

Standard workflows bill $0.025 per 1,000 state transitions; Express workflows bill $1.00 per million requests plus a duration charge. For high-volume, short workflows, Express is often 50-100x cheaper. Use Standard for low-volume, long-running, audited orchestration; use Express for high-frequency event processing. The transition meter is what makes high-volume Standard workflows expensive.

Step Functions has two execution types that look similar in the console and bill on completely different meters. Choosing wrong — usually running a high-frequency workflow as Standard — is one of the more expensive defaults in AWS, and it's invisible until the transition count piles up.

Two different meters

Standard: $0.025 per 1,000 state transitions. Every state your workflow enters is one transition. A 10-state workflow is 10 transitions per execution.
Express: $1.00 per million requests plus ~$0.00001667 per GB-second of execution time. No per-transition charge — you pay for invocations and duration, like Lambda.

The 100x difference, concretely

Take a 10-state workflow invoked 1,000,000 times a month, each running ~1 second at 128 MB:

Standard:
  10 states × 1M execs = 10M transitions
  10M / 1,000 × $0.025                       = $250.00/month

Express:
  1M requests × $1.00/million                = $1.00
  1M × 1s × 0.125 GB × $0.00001667/GB-s      ≈ $2.08
                                       Total ≈ $3.08/month

Same logic, ~80x cost difference. The transition meter is linear in both state count and execution count, so high-frequency Standard workflows scale badly.

When Standard is the right call

Standard isn't a mistake — it's the right tool for durable orchestration:

Long-running workflows (up to a year) — order fulfillment, multi-day approvals.
Exactly-once execution semantics.
Full execution history and visual debugging for every run.
Human-approval and callback patterns.

For low-volume business processes, the per-transition cost is trivial and the durability is worth it.

When Express wins

Express suits high-throughput, short-lived workflows (up to 5 minutes): IoT and streaming ingestion, high-frequency microservice orchestration, and anything event-driven at scale. You give up per-execution history (Express logs to CloudWatch instead) in exchange for order-of-magnitude lower cost. If a workflow fronts a Lambda fleet processing events, it almost certainly should be Express — see estimating Lambda costs for the function side of that bill.

FAQ

How is AWS Step Functions priced?

Standard workflows bill $0.025 per 1,000 state transitions — every step in a workflow is a transition. Express workflows bill differently: $1.00 per million requests plus a duration charge based on memory and runtime (~$0.00001667 per GB-second). The two models can differ by 100x for the same logic at high volume.

Standard vs Express Step Functions — which is cheaper?

Express is dramatically cheaper for high-volume, short-duration workflows (event processing, streaming, high-frequency automation) because per-transition billing on Standard gets expensive fast. Standard is cheaper and more appropriate for low-volume, long-running, or audited workflows where you need exactly-once execution and full history.

Why is my Step Functions bill so high?

Almost always a high-frequency Standard workflow. At $0.025 per 1,000 transitions, a workflow with 10 states invoked a million times a month is 10M transitions = $250. The same workload on Express would cost a few dollars. High-volume Standard workflows are the classic Step Functions overspend.

When should I use Standard workflows?

When you need exactly-once execution, long-running workflows (up to a year), full execution history and visual debugging, or human-approval steps. Standard is built for durable, auditable orchestration — order fulfillment, ETL pipelines, approval flows — not for high-frequency event handling.

When should I use Express workflows?

For high-volume, short-lived workflows (up to 5 minutes) where throughput matters more than per-execution history: IoT ingestion, streaming data processing, high-frequency microservice orchestration. Express trades some durability guarantees for order-of-magnitude lower cost at scale.

How does C3X estimate Step Functions cost?

C3X prices an aws_sfn_state_machine by its type and treats transitions/requests as usage-driven — add expected monthly executions and state count in c3x-usage.yml to model the Standard transition bill or the Express request+duration bill.

What to do next

The decision hinges on execution volume, which you know before you deploy. C3X prices an aws_sfn_state_machine by type and lets you model expected executions and state count, so a high-frequency workflow accidentally left on Standard shows its real bill in review. The quickstart runs it on your Terraform in minutes.