b stack radar
A

Amazon SageMaker Pricing

Every Amazon SageMaker plan with its monthly and annual price, what it costs for common team sizes, and a calculator to estimate your own monthly cost before you buy.

Compare Amazon SageMaker alternatives by priceRead our Amazon SageMaker reviewCompare AI Software pricing

Plans and prices

PlanMonthly billingAnnual billing (per month)Per user or flatUsersFree tier
Free Tier (first 2 months)FreeFree—No limit listedYes
On-Demand (pay as you go)Contact salesContact sales—No limit listedNo
SageMaker Savings PlansContact salesContact sales—No limit listedNo

Prices last updated November 2026.

What each Amazon SageMaker plan includes

Free Tier (first 2 months)

  • 160 hours per month of Canvas session time
  • 50 hours per month of HyperPod on m5.xlarge
  • 250 hours per month of ml.t3.medium notebooks
  • 150,000 seconds per month of serverless inference
  • 50 hours per month of m4.xlarge or m5.xlarge training
  • Feature Store: 10M write units, 10M read units, 25 GB storage
  • 125 hours per month of m4.xlarge or m5.xlarge real-time inference

On-Demand (pay as you go)

  • No minimums or upfront commitment
  • HyperPod ml.g5.24xlarge at $10.18 per hour
  • Billed per instance-hour by ML instance type and Region
  • Studio itself has no extra charge beyond compute and storage
  • Examples: ml.c5.xlarge hosting at $0.204 per hour, ml.m4.4xlarge training at $0.96 per hour, ml.g4dn.xlarge notebook at $0.7364 per hour

SageMaker Savings Plans

  • Commitment to consistent usage
  • Up to 64% lower cost than on-demand
  • Applies across instance families, sizes, and Regions

What Amazon SageMaker costs by team size

Team size1 user5 users10 users25 users50 users
Monthly costFreeFree Tier (first 2 months)FreeFree Tier (first 2 months)FreeFree Tier (first 2 months)FreeFree Tier (first 2 months)FreeFree Tier (first 2 months)

The plan that fits each team size, billed monthly, before add-ons and usage charges.

Add-ons

Add-onMonthly billingAnnual billing (per month)Charged
Human-based model evaluationCharged per completed human evaluation task, plus instance charges.Not listedNot listedPer completed human task
Model MonitorMonitoring jobs for deployed models, billed per instance-hour (for example ml.m5.4xlarge at $0.922 per hour). Built-in rules include up to 30 hours free on ml.m5.xlarge.Not listedNot listedPer instance-hour

Usage pricing

MetricPlanIncluded per monthPrice after that
Inference duration on provisioned concurrencyOn-Demand (pay as you go)None$0.00 per second
Real-time hosting on ml.c5.xlargeOn-Demand (pay as you go)None$0.204 per instance-hour
Serverless inference computeOn-Demand (pay as you go)None$0.00 per second
Serverless inference data processingOn-Demand (pay as you go)None$0.016 per GB
Serverless provisioned concurrency capacityOn-Demand (pay as you go)None$0.00 per second

Amazon SageMaker Pricing Calculator

Estimate your team's real monthly and annual cost based on seats, billing cycle, and add-ons.

Billing cycle
Add-ons

Cost breakdown

Total monthly costFree
Per year on annual billingFree
Cost per userFree/mo

Estimated from the vendor's published prices, last updated November 2026. Actual costs may vary based on negotiated contracts. How we collect prices

Ready to try Amazon SageMaker?

Use this estimate to align with your team, then start a trial or subscription.

Visit Amazon SageMaker

Better value alternatives

See what similar tools would cost for the same team size.

Amazon SageMaker pricing FAQs

Yes, Amazon SageMaker has a free plan: Free Tier (first 2 months).

Amazon SageMaker's Free Tier (first 2 months) plan includes 160 hours per month of Canvas session time, 50 hours per month of HyperPod on m5.xlarge, 250 hours per month of ml.t3.medium notebooks, 150,000 seconds per month of serverless inference, 50 hours per month of m4.xlarge or m5.xlarge training, Feature Store: 10M write units, 10M read units, 25 GB storage and 1 more.

Amazon SageMaker is free for 10 users on its Free Tier (first 2 months) plan.

Yes. Amazon SageMaker sells add-ons on top of its plans: Human-based model evaluation and Model Monitor. It also charges by usage for Inference duration on provisioned concurrency, Real-time hosting on ml.c5.xlarge, Serverless inference compute and 2 more, past what the plan includes.