AWS SageMaker Pricing (2026)
ML platform services provide end-to-end machine learning workflows, from data labeling and model training to deployment and monitoring. Managed infrastructure lets data scientists focus on models, not operations. SageMaker offers Studio IDE, built-in algorithms, and Inference endpoints with multi-model support. Starting from $0.0360/hr ($26.28/mo) for Machine Learning.
Key Features
- ✓Managed Jupyter notebooks and IDE environments
- ✓Built-in algorithms and framework support (TensorFlow, PyTorch)
- ✓Distributed training on GPU/TPU clusters
- ✓One-click model deployment and auto-scaling endpoints
- ✓MLOps: experiment tracking, model registry, and monitoring
Common Use Cases
Model Training
Train custom ML models on managed GPU/TPU infrastructure with automatic hyperparameter tuning.
Model Deployment
Deploy models as real-time endpoints or batch transform jobs with automatic scaling.
MLOps Pipelines
Build automated ML pipelines from data preparation through deployment and monitoring.
On-Demand Pricing
Pay-as-you-go pricing with no upfront commitment. You are billed per hour of usage and can start or stop at any time. Hourly rates start at $0.0360/hr ($26.28/mo) for Machine Learning.
| Instance | vCPU | Memory | Price/hr | Price/mo |
|---|---|---|---|---|
| ml.m4.16xlarge-Processing | - | 256 GB | - | - |
| ml.c5.4xlarge-Processing | - | 32 GB | - | - |
| ml.t3.xlarge-Notebook | - | 16 GB | - | - |
| ml.m5.12xlarge-RSession | - | 192 GB | - | - |
| Amazon SageMaker | - | - | - | - |
| ml.m6id.4xlarge | - | 64 GB | $1.1390 | $831.47 |
| ml.g6.48xlarge-training | - | 16 GB | - | - |
| ml.c5.4xlarge-Cluster | - | 32 GB | - | - |
| ml.r7i.2xlarge-BatchTransform | - | 64 GB | - | - |
| Amazon SageMaker | - | - | $0.0800 | $58.40 |
| Amazon SageMaker | - | 384 GB | - | - |
| ml.r5.24xlarge-AsyncInf | - | 768 GB | - | - |
| ml.r5d.12xlarge-Processing | - | 384 GB | - | - |
| ml.r6i.large-Notebook | - | 16 GB | - | - |
| ml.p6-b300.48xlarge-trainingplan | - | 4096 GB | - | - |
| ml.m5.xlarge-AsyncInf | - | 16 GB | - | - |
| Amazon SageMaker | - | - | - | - |
| ml.r7i.24xlarge-Cluster | - | 768 GB | - | - |
| ml.g5.24xlarge-Cluster | - | 384 GB | - | - |
| ml.m6id.large | - | 8 GB | - | - |
| ml.t3.2xlarge-Notebook | - | 32 GB | - | - |
| Machine Learning | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Storage | - | - | - | - |
| ML Instance | - | - | - | - |
| ml.m6i.32xlarge | - | 512 GB | - | - |
| ml.g6.24xlarge-Hosting | - | 384 GB | $8.3440 | $6,091.12 |
| FeatureStore PayPerRequestThroughput | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Machine Learning | - | - | - | - |
| ML Instance | - | - | - | - |
| Amazon SageMaker | - | 16 GB | $0.1510 | $110.23 |
| ml.c6g.xlarge-AsyncInf | - | 8 GB | - | - |
| ml.m6gd.large-Hosting | - | 8 GB | - | - |
| Machine Learning | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| ml.r5.4xlarge-SpotTraining | - | 128 GB | - | - |
| ml.r5.24xlarge-Notebook | - | 768 GB | - | - |
| ml.g4dn.xlarge-Hosting | - | 16 GB | - | - |
| ml.r7i.16xlarge | - | 512 GB | - | - |
| Amazon SageMaker | - | - | - | - |
| Machine Learning | - | - | - | - |
| ml.m7i.48xlarge-BatchTransform | - | 768 GB | - | - |
| ml.r6id.12xlarge-Notebook | - | 384 GB | - | - |
| ml.c6id.2xlarge-Notebook | - | 16 GB | - | - |
| ml.m5.4xlarge-Processing | - | 64 GB | - | - |
| ml.m5.2xlarge-Notebook | - | 32 GB | $0.4610 | $336.53 |
| ml.g5.12xlarge | - | 192 GB | - | - |
| ml.r7i.48xlarge-BatchTransform | - | 1536 GB | - | - |
Pricing by Instance Family
ML Instance
71 instance types available
| Instance | vCPU | Memory | Price/hr | Price/mo |
|---|---|---|---|---|
| ml.m4.16xlarge-Processing | - | 256 GB | - | - |
| ml.c5.4xlarge-Processing | - | 32 GB | - | - |
| ml.t3.xlarge-Notebook | - | 16 GB | - | - |
| ml.m5.12xlarge-RSession | - | 192 GB | - | - |
| ml.m6id.4xlarge | - | 64 GB | $1.1390 | $831.47 |
| ml.g6.48xlarge-training | - | 16 GB | - | - |
| ml.c5.4xlarge-Cluster | - | 32 GB | - | - |
| ml.r7i.2xlarge-BatchTransform | - | 64 GB | - | - |
| ml.r5.24xlarge-AsyncInf | - | 768 GB | - | - |
| ml.r5d.12xlarge-Processing | - | 384 GB | - | - |
| ml.r6i.large-Notebook | - | 16 GB | - | - |
| ml.p6-b300.48xlarge-trainingplan | - | 4096 GB | - | - |
| ml.m5.xlarge-AsyncInf | - | 16 GB | - | - |
| ml.r7i.24xlarge-Cluster | - | 768 GB | - | - |
| ml.g5.24xlarge-Cluster | - | 384 GB | - | - |
| ml.m6id.large | - | 8 GB | - | - |
| ml.t3.2xlarge-Notebook | - | 32 GB | - | - |
| ML Instance | - | - | - | - |
| ml.m6i.32xlarge | - | 512 GB | - | - |
| ml.g6.24xlarge-Hosting | - | 384 GB | $8.3440 | $6,091.12 |
| ML Instance | - | - | - | - |
| ml.c6g.xlarge-AsyncInf | - | 8 GB | - | - |
| ml.m6gd.large-Hosting | - | 8 GB | - | - |
| ml.r5.4xlarge-SpotTraining | - | 128 GB | - | - |
| ml.r5.24xlarge-Notebook | - | 768 GB | - | - |
| ml.g4dn.xlarge-Hosting | - | 16 GB | - | - |
| ml.r7i.16xlarge | - | 512 GB | - | - |
| ml.m7i.48xlarge-BatchTransform | - | 768 GB | - | - |
| ml.r6id.12xlarge-Notebook | - | 384 GB | - | - |
| ml.c6id.2xlarge-Notebook | - | 16 GB | - | - |
| ml.m5.4xlarge-Processing | - | 64 GB | - | - |
| ml.m5.2xlarge-Notebook | - | 32 GB | $0.4610 | $336.53 |
| ml.g5.12xlarge | - | 192 GB | - | - |
| ml.r7i.48xlarge-BatchTransform | - | 1536 GB | - | - |
| ml.g6.48xlarge-AsyncInf | - | 768 GB | - | - |
| ml.m5.2xlarge-RSessionGateway | - | 32 GB | - | - |
| ml.m5d.xlarge-RSession | - | 16 GB | - | - |
| ml.c5d.2xlarge-Notebook | - | 16 GB | - | - |
| ml.m4.10xlarge-Processing | - | 160 GB | - | - |
| ml.r7i.12xlarge-Cluster | - | 384 GB | - | - |
| ml.p3.16xlarge-Notebook | - | 488 GB | - | - |
| ml.c7i.xlarge | - | 8 GB | - | - |
| ml.m4.2xlarge-Processing | - | 32 GB | - | - |
| ml.c5.9xlarge-AsyncInf | - | 72 GB | - | - |
| ML Instance | - | - | - | - |
| ml.c6i.16xlarge | - | 128 GB | - | - |
| ml.r5.16xlarge-Notebook | - | 512 GB | - | - |
| ml.r5.4xlarge | - | 128 GB | - | - |
| ml.m5d.2xlarge-RSessionGateway | - | 32 GB | - | - |
| ml.g5.4xlarge-Cluster | - | 64 GB | - | - |
Amazon SageMaker
20 instance types available
| Instance | vCPU | Memory | Price/hr | Price/mo |
|---|---|---|---|---|
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | $0.0800 | $58.40 |
| Amazon SageMaker | - | 384 GB | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | 16 GB | $0.1510 | $110.23 |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | 384 GB | - | - |
| Amazon SageMaker | - | 1536 GB | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | 64 GB | - | - |
| Amazon SageMaker | - | 16 GB | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | - | - | - |
| Amazon SageMaker | - | 16 GB | - | - |
| Amazon SageMaker | - | - | - | - |
Machine Learning
7 instance types available
| Instance | Price/hr | Price/mo |
|---|---|---|
| Machine Learning | - | - |
| Machine Learning | - | - |
| Machine Learning | - | - |
| Machine Learning | - | - |
| Machine Learning | $0.0360 | $26.28 |
| Machine Learning | - | - |
| Machine Learning | - | - |
Storage
1 instance type available
| Instance | Price/hr | Price/mo |
|---|---|---|
| Storage | - | - |
FeatureStore PayPerRequestThroughput
1 instance type available
| Instance | Price/hr | Price/mo |
|---|---|---|
| FeatureStore PayPerRequestThroughput | - | - |
Reserved Instance & Savings Plans Pricing
Commit to 1 or 3 years for lower hourly rates.
| Instance | vCPU | Memory | Price/hr | Price/mo | 1yr RI/hr | 3yr RI/hr |
|---|---|---|---|---|---|---|
| ml.m4.16xlarge-Processing | - | 256 GB | - | - | - | - |
| ml.c5.4xlarge-Processing | - | 32 GB | - | - | - | - |
| ml.t3.xlarge-Notebook | - | 16 GB | - | - | - | - |
| ml.m5.12xlarge-RSession | - | 192 GB | - | - | - | - |
| Amazon SageMaker | - | - | - | - | - | - |
| ml.m6id.4xlarge | - | 64 GB | $1.1390 | $831.47 | - | - |
| ml.g6.48xlarge-training | - | 16 GB | - | - | - | - |
| ml.c5.4xlarge-Cluster | - | 32 GB | - | - | - | - |
| ml.r7i.2xlarge-BatchTransform | - | 64 GB | - | - | - | - |
| Amazon SageMaker | - | - | $0.0800 | $58.40 | - | - |
| Amazon SageMaker | - | 384 GB | - | - | - | - |
| ml.r5.24xlarge-AsyncInf | - | 768 GB | - | - | - | - |
| ml.r5d.12xlarge-Processing | - | 384 GB | - | - | - | - |
| ml.r6i.large-Notebook | - | 16 GB | - | - | - | - |
| ml.p6-b300.48xlarge-trainingplan | - | 4096 GB | - | - | - | - |
| ml.m5.xlarge-AsyncInf | - | 16 GB | - | - | - | - |
| Amazon SageMaker | - | - | - | - | - | - |
| ml.r7i.24xlarge-Cluster | - | 768 GB | - | - | - | - |
| ml.g5.24xlarge-Cluster | - | 384 GB | - | - | - | - |
| ml.m6id.large | - | 8 GB | - | - | - | - |
| ml.t3.2xlarge-Notebook | - | 32 GB | - | - | - | - |
| Machine Learning | - | - | - | - | - | - |
| Amazon SageMaker | - | - | - | - | - | - |
| Storage | - | - | - | - | - | - |
| ML Instance | - | - | - | - | - | - |
| ml.m6i.32xlarge | - | 512 GB | - | - | - | - |
| ml.g6.24xlarge-Hosting | - | 384 GB | $8.3440 | $6,091.12 | - | - |
| FeatureStore PayPerRequestThroughput | - | - | - | - | - | - |
| Amazon SageMaker | - | - | - | - | - | - |
| Amazon SageMaker | - | - | - | - | - | - |
| Machine Learning | - | - | - | - | - | - |
| ML Instance | - | - | - | - | - | - |
| Amazon SageMaker | - | 16 GB | $0.1510 | $110.23 | - | - |
| ml.c6g.xlarge-AsyncInf | - | 8 GB | - | - | - | - |
| ml.m6gd.large-Hosting | - | 8 GB | - | - | - | - |
| Machine Learning | - | - | - | - | - | - |
| Amazon SageMaker | - | - | - | - | - | - |
| ml.r5.4xlarge-SpotTraining | - | 128 GB | - | - | - | - |
| ml.r5.24xlarge-Notebook | - | 768 GB | - | - | - | - |
| ml.g4dn.xlarge-Hosting | - | 16 GB | - | - | - | - |
| ml.r7i.16xlarge | - | 512 GB | - | - | - | - |
| Amazon SageMaker | - | - | - | - | - | - |
| Machine Learning | - | - | - | - | - | - |
| ml.m7i.48xlarge-BatchTransform | - | 768 GB | - | - | - | - |
| ml.r6id.12xlarge-Notebook | - | 384 GB | - | - | - | - |
| ml.c6id.2xlarge-Notebook | - | 16 GB | - | - | - | - |
| ml.m5.4xlarge-Processing | - | 64 GB | - | - | - | - |
| ml.m5.2xlarge-Notebook | - | 32 GB | $0.4610 | $336.53 | - | - |
| ml.g5.12xlarge | - | 192 GB | - | - | - | - |
| ml.r7i.48xlarge-BatchTransform | - | 1536 GB | - | - | - | - |
How SageMaker Pricing Works
On-Demand
Pay per hour with no long-term commitment. Ideal for variable workloads and development environments.
Reserved / Committed Use
Commit to 1 or 3 years for significant discounts.
Spot / Preemptible
Use spare capacity at steep discounts. Best for fault-tolerant, batch, and stateless workloads.
Pricing Insights
Amazon SageMaker offers 11 pricing tiers from $0.0360/hr to $14.69/hr in the US East region.
Compare tiers by specs and price to find the best fit for your workload.
By Instance Family
- ●ML Instance: 8 tiers, $0.3020/hr – $14.69/hr
- ●Amazon SageMaker: 2 tiers, $0.0800/hr – $0.1510/hr
- ●Machine Learning: 1 tier at $0.0360/hr
Cost Optimization Tips
- ✓With 11 available tiers, right-sizing your instance to match actual usage is key to cost optimization.
Monthly Cost Examples
AWS SageMaker Free Tier
Free limit: 250 hrs/mo (ml.t3.medium). Available for the first 12 months after account creation, then standard on-demand pricing applies.
Free Tier Comparison
| Provider | Service | Free Offering | Duration | Limit |
|---|---|---|---|---|
| AWS | SageMaker | Studio notebooks | 12 months | 250 hrs/mo (ml.t3.medium) |
| Azure | Machine Learning | None | - | - |
| GCP | Vertex AI | None | - | - |
Frequently Asked Questions
What is AWS SageMaker?
AWS SageMaker is a cloud service offered by Amazon Web Services. It provides various configurations (100 pricing tiers available) with pay-as-you-go and committed-use pricing options.
How much does AWS SageMaker cost per month?
Prices range from $26.28/month for Machine Learning to $10,722.24/month for ml.p3.8xlarge-Notebook on On-Demand pricing in us-east-1.
Does AWS SageMaker have a free tier?
AWS offers various free tier options. Check the official AWS pricing page for the most current free tier details for SageMaker.
How many AWS SageMaker pricing tiers are available?
There are 100 pricing tiers available for AWS SageMaker. These range from entry-level configurations to high-performance options for enterprise workloads.
What pricing models does AWS SageMaker offer?
AWS SageMaker offers On-Demand (pay-per-hour, no commitment), Reserved/Committed Use (1-3 year commitments for significant discounts), and in some cases Spot/Preemptible pricing for interruptible workloads at the lowest cost.
How is GPU instance pricing structured?
GPU instances are priced per hour and vary significantly by GPU model (T4, A10G, A100, H100). On-demand rates are highest; spot/preemptible instances can cut costs 60-90% for fault-tolerant training jobs. Some providers also offer per-second billing and committed-use discounts for GPUs.