Amazon SageMaker is billed per use with no upfront fee for on-demand rates. In US East (N. Virginia), ml.t3.medium costs $0.05 an hour. Main rates in US East (N. Virginia): $0.04 per Credits ($0.04 per Credit for DataAgentCredit); $0.10 per GB-Month ($0.10 per GB of data stored per month). Prices from the AWS Price List published November 14, 2026.
On-demand rates in US East (N. Virginia), as AWS lists them in the Price List published November 14, 2026.
Source: AWS Price List file for Amazon SageMaker (version 20261114005740), read on November 14, 2026. Prices differ by Region and change without notice; check the AWS pricing page before you commit.
| Instance | Used for | vCPU | Memory | Per hour | Per month (730 h) |
|---|---|---|---|---|---|
| ml.t3.medium | Processing | 2 | 4 GiB | $0.05 | $36.50 |
| ml.m5.large | Processing | 2 | 8 GiB | $0.115 | $83.95 |
| ml.m5.xlarge | Processing | 4 | 16 GiB | $0.23 | $167.90 |
| ml.c5.xlarge | Processing | 4 | 8 GiB | $0.204 | $148.92 |
| ml.g5.xlarge | Processing | 4 | 16 GiB | $1.408 | $1,027.84 |
| ml.g5.2xlarge | Processing | 8 | 32 GiB | $1.515 | $1,105.95 |
| ml.g6.xlarge | Processing | 4 | 16 GiB | $1.12672 | $822.51 |
| ml.m5.large | Real-time inference | 2 | 8 GiB | $0.115 | $83.95 |
| ml.m5.xlarge | Real-time inference | 4 | 16 GiB | $0.23 | $167.90 |
| ml.c5.xlarge | Real-time inference | 4 | 8 GiB | $0.204 | $148.92 |
| ml.g5.xlarge | Real-time inference | 4 | 16 GiB | $1.408 | $1,027.84 |
| ml.g5.2xlarge | Real-time inference | 8 | 32 GiB | $1.515 | $1,105.95 |
| ml.g6.xlarge | Real-time inference | 4 | 16 GiB | $1.1267 | $822.49 |
| ml.g6e.xlarge | Real-time inference | 4 | 32 GiB | $2.6054 | $1,901.94 |
| ml.p4d.24xlarge | Real-time inference | 96 | 1152 GiB | $25.2513 | $18,433.4 |
| ml.p5.48xlarge | Real-time inference | 192 | 2 TiB | $63.296 | $46,206.1 |
| ml.inf2.xlarge | Real-time inference | 4 | 16 GiB | $0.99 | $722.70 |
| ml.trn1.2xlarge | Real-time inference | 8 | 32 GiB | $1.55 | $1,131.5 |
| ml.t3.medium | Studio notebook | 2 | 4 GiB | $0.05 | $36.50 |
| ml.m5.large | Studio notebook | 2 | 8 GiB | $0.115 | $83.95 |
| ml.m5.xlarge | Studio notebook | 4 | 16 GiB | $0.23 | $167.90 |
| ml.c5.xlarge | Studio notebook | 4 | 8 GiB | $0.204 | $148.92 |
| ml.g5.xlarge | Studio notebook | 4 | 16 GiB | $1.41 | $1,029.3 |
| ml.g5.2xlarge | Studio notebook | 8 | 32 GiB | $1.52 | $1,109.6 |
| ml.p4d.24xlarge | Studio notebook | 96 | 1152 GiB | $25.2513 | $18,433.4 |
| ml.trn1.2xlarge | Studio notebook | 8 | 32 GiB | $1.54 | $1,124.2 |
| ml.m5.large | Training | 2 | 8 GiB | $0.115 | $83.95 |
| ml.m5.xlarge | Training | 4 | 16 GiB | $0.23 | $167.90 |
| ml.c5.xlarge | Training | 4 | 8 GiB | $0.204 | $148.92 |
| ml.g5.xlarge | Training | 4 | 16 GiB | $1.408 | $1,027.84 |
| ml.g5.2xlarge | Training | 8 | 32 GiB | $1.515 | $1,105.95 |
| ml.g6.xlarge | Training | 4 | 16 GiB | $1.127 | $822.71 |
| ml.g6e.xlarge | Training | 4 | 32 GiB | $2.61 | $1,905.3 |
| ml.p4d.24xlarge | Training | 96 | 1152 GiB | $25.2513 | $18,433.4 |
| ml.p5.48xlarge | Training | 192 | 2 TiB | $63.296 | $46,206.1 |
| ml.trn1.2xlarge | Training | 8 | 32 GiB | $1.54 | $1,124.2 |
| Rate | Price (USD) | Billed |
|---|---|---|
| $0.04 per Credit for DataAgentCredit in US East (N. Virginia) | $0.04 | per Credits |
| $0.10 per GB of data stored per month in US East (N. Virginia) | $0.10 | per GB-Month |
| $1.91 per Large Tracking Server hour in US East (N. Virginia) | $1.91 | per hour |
| $1.04 per Medium Tracking Server hour in US East (N. Virginia) | $1.04 | per hour |
| $0.60 per Small Tracking Server hour in US East (N. Virginia) | $0.60 | per hour |
| $15.00 per 1M cells - for next 90M cells used in Autopilot per job in US East (N. Virginia) | $15.00 | per MillionCells |
| $1.9 per Hrs for Canvas:Workspace Instance (Session-Hrs):RunInstance in US East (N. Virginia) | $1.90 | per hour |
| $8.63 per TrainingHour for Autopilot pricing estimation (not for actual billing) | $8.63 | per TrainingHour |
| $15.00 per million cells for over 10 million cells and up to 90 million cells per model | $15.00 | per MillionCells |
| $7.00 per million cells for over 100 million cells per CREATE MODEL request with RedshiftML | $7.00 | per MillionCells |
| $7.00 per million cells for over 100M cells per model | $7.00 | per MillionCells |
| $20.00 per million cells for the first 10 million cells per CREATE MODEL request with RedshiftML | $20.00 | per MillionCells |
| $30.00 per million cells for the first 10 million cells per model | $30.00 | per MillionCells |
| $20.00 per million cells, per CREATE MODEL request for RedshiftML, after exceeding free tier count of 2 requests | $20.00 | per CreateModelRequest |
| $30.00 per million cells, per model after exceeding free tier count of 10 requests | $30.00 | per CreateModelRequest |
| $0.016 per GB for Endpoint Data IN | $0.016 | per GB |
| $0.016 per GB for Endpoint Data OUT | $0.016 | per GB |
| $1.00 per registered SageMaker Edge Manager device for 1 to 100,000 devices in US East (N. Virginia) | $1.00 | per Units |
| $0.2 per registered SageMaker Edge Manager device for 100,001 to 750,000 devices in US East (N. Virginia) | $0.20 | per Units |
| $0.1 per registered SageMaker Edge Manager device for more than 750,000 devices in US East (N. Virginia) | $0.10 | per Units |
| $10.00 per SageMaker Edge Manager managed model for 1 to 10,000 models in US East (N. Virginia) | $10.00 | per Units |
| $5.00 per SageMaker Edge Manager managed model for 10,001 to 50,000 models in US East (N. Virginia) | $5.00 | per Units |
| $2.00 per SageMaker Edge Manager managed model for 50,001 to 100,000 models in US East (N. Virginia) | $2.00 | per Units |
| $0.00013 per SageMaker Feature Store RCU per hour in US East (N. Virginia) | $130.00 | per 1M ReadCapacityUnit-hour |
| $0.00065 per SageMaker Feature Store WCU per hour in US East (N. Virginia) | $650.00 | per 1M WriteCapacityUnit-hour |
| $0.0224 per million SageMaker Feature Store in-memory read request units in US East (N. Virginia) | $0.0224 | per 1M ReadRequestUnits |
| $0.0249 per million SageMaker Feature Store in-memory write request units in US East (N. Virginia) | $0.0249 | per 1M WriteRequestUnits |
| $0.25 per million SageMaker Feature Store read request units in US East (N. Virginia) | $0.25 | per 1M ReadRequestUnits |
| $1.25 per million SageMaker Feature Store write request units in US East (N. Virginia) | $1.25 | per 1M WriteRequestUnits |
| $0.233 per GB-Hour of SageMaker Feature Store in-memory storage in US East (N. Virginia) | $0.233061 | per GB-hour |
| $0.45 per GB-Month of SageMaker Feature Store Storage in US East (N. Virginia) | $0.45 | per GB-month |
| $1.24 per 1M token for MultiTurnRFT:Session-LORA-Prefill-huggingface-vlm-gemma-4-31b-it in US East (N. Virginia) | $1.24 | per 1M tokens |
| $3.73 per 1M token for MultiTurnRFT:Session-LORA-Sample-huggingface-vlm-gemma-4-31b-it in US East (N. Virginia) | $3.73 | per 1M tokens |
| $3.73 per 1M token for MultiTurnRFT:Session-LORA-Train-huggingface-vlm-gemma-4-31b-it in US East (N. Virginia) | $3.73 | per 1M tokens |
| $0.50 per 1M token for Serverless Training:Data Generation-Contextual-Input in US East (N. Virginia) | $0.50 | per 1M tokens |
| $1.50 per 1M token for Serverless Training:Data Generation-Contextual-Output in US East (N. Virginia) | $1.50 | per 1M tokens |
| $0.50 per 1M token for Serverless Training:Data Generation-NonContextual-Input in US East (N. Virginia) | $0.50 | per 1M tokens |
| $1.50 per 1M token for Serverless Training:Data Generation-NonContextual-Output in US East (N. Virginia) | $1.50 | per 1M tokens |
| $1.45 per 1M token for Serverless Training:Evaluation-Input-deepseek-llm-r1-distill-llama-70b in US East (N. Virginia) | $1.45 | per 1M tokens |
| $0.20 per 1M token for Serverless Training:Evaluation-Input-deepseek-llm-r1-distill-llama-8b in US East (N. Virginia) | $0.20 | per 1M tokens |
AWS Activate credits pay for eligible AWS services, and Amazon SageMaker usage is billed on your AWS invoice like any other service. Activate Founders gives $1,000 to start, up to $5,000; you apply for it yourself. AWS promotional credits do not pay Savings Plans or Reserved Instance upfront fees.
For a larger project on SageMaker, a vetted AWS partner can request funding only partners can file: MAP for a migration, POC to test something, incremental workload funding for new work. For startups that is usually $50K to $150K. Depends on the workload; AWS approves.
| Funding | Who files it | Amount | When it fits |
|---|---|---|---|
| Migration Acceleration Program (MAP) | Your matched AWS partner | $50K to $150K for startups. Depends on the workload; AWS approves. | You are not on AWS yet and want to migrate. |
| Proof of concept (POC) funding | Your matched AWS partner | $50K to $150K for startups. Depends on the workload; AWS approves. | You want to test something, whether or not you are on AWS today. |
| Incremental workload (IW) funding | Your matched AWS partner | $50K to $150K for startups. Depends on the workload; AWS approves. | You are already on AWS and are launching new work: a new AI feature, a new customer or a new contract. |
| AWS Activate Founders | Founder applies | $1,000 to start, up to $5,000 | Self-funded startups with no investor or accelerator behind them. |
| AWS Activate Portfolio | Founder applies | up to $200,000 | Startups backed by a VC firm, accelerator or angel investor that is an AWS Activate Provider. |
| AWS Credits for AI Startups | Invite only (AWS account manager) | $200,000 or more, by invitation | AI startups ready to scale after Activate Portfolio, invited by AWS. |
Enter how much you expect to use in a month. Uses the on-demand rates above (US East (N. Virginia), AWS Price List November 14, 2026). Tiered discounts, free tier allowances and taxes are not included.
A vetted AWS partner checks which AWS funding fits your workload: MAP, POC or incremental workload funding. We reply within 24 hours; your matched partner reaches out within 5 business days.