Amazon SageMaker pricing

SageMaker Unified Studio is a data and AI development environment that provides an integrated experience to use all your data and tools for analytics and AI. SageMaker Unified Studio uses Amazon SageMaker Catalog, built on Amazon DataZone, for end-to-end governance and access control through entities such as domains, projects, and assets. You will be charged for your usage of SageMaker Catalog. In addition, each AWS service that you use through the SageMaker Unified Studio is subject to its own individual pricing. This includes AWS storage and compute services, as well as any third-party services, like Git providers. There is no separate cost for using SageMaker Unified Studio itself.

To accurately estimate your costs, review the individual pricing for the various AWS services available within SageMaker Unified Studio below. For detailed pricing information, consult the official AWS pricing pages for each AWS service you plan to use.

The SageMaker Unified Studio Free Tier helps you quickly get started innovating with data and AI and at no cost by offering a selection of always-free features and honoring your current AWS Free Tier allocations or pay-per-use agreements (PPAs) for AWS services that you use through the SageMaker Unified Studio.

Always-free features

Core requests in the SageMaker Unified Studio, including initial setup through manual creation of domains, project management, user management, and managing policy grants, are provided at no cost to help you set up and navigate your SageMaker Unified Studio environment. The full list of free requests can be found at the bottom of this pricing page. 

Other AWS Free Tier benefits

AWS honors and carries over your current AWS Free Tier allocations or pay-per-use agreements (PPAs) for AWS services that you use through SageMaker Unified Studio, such as Amazon SageMaker AI and Amazon Redshift. This includes a monthly AWS Free Tier for SageMaker Catalog, the JupyterLab notebooks, Amazon Q, metadata storage, and API requests. Please note that AWS Free Tier benefits are associated with your AWS account, and any usage through service consoles or other interfaces also consumes your available allocation. You can continue to use remaining AWS Free Tier benefits through SageMaker Unified Studio or other AWS interfaces until the full allocation is consumed or the AWS Free Tier period expires. To learn more, see the pricing details below.

SageMaker Lakehouse unifies all your data across Amazon Simple Storage Service (Amazon S3) data lakes and Amazon Redshift data warehouses, helping you build powerful analytics and AI/ML applications on a single copy of data. With SageMaker Lakehouse, you only pay for the resources you use, allowing for cost-effective scaling of your data lake infrastructure. Metadata storage and API requests follow AWS Glue Data Catalog pricing, including the AWS Free Tier. Data storage and access costs depend on your choice of Amazon S3 or Amazon Redshift Managed Storage (RMS), with associated storage and compute charges. Automated statistics collection and Apache Iceberg table maintenance incur additional fees based on usage. Fine-grained permissions, powered by AWS Lake Formation, are provided at no extra cost. For the most accurate and detailed pricing information, consult SageMaker Lakehouse pricing.

SageMaker AI follows a pay-as-you-go pricing model with no upfront commitments or minimum fees. The key pricing dimensions for SageMaker AI include instance usage (compute resources used in training, hosting, and notebook instances), storage (Amazon SageMaker notebooks, Amazon Elastic Block Store (Amazon EBS) volumes, and Amazon S3), data processing jobs, model deployment, and MLOps (Amazon SageMaker Pipelines and Model Monitor). Additional components like SageMaker AI Feature Store and Data Wrangler tools might have their own pricing considerations. The actual costs can vary based on factors such as AWS Region, instance types, storage volume, and specific usage patterns. For the most accurate and detailed pricing information, consult SageMaker AI pricing.

Amazon Bedrock IDE offers a flexible, consumption-based pricing model. You pay only for the Amazon Bedrock resources you use, without any minimum fees or upfront commitments. Pricing is based on your usage of Amazon Bedrock, such as the Agents, Flows, Knowledge Bases, Functions, and Guardrails features and the models available through Amazon Bedrock IDE. For the most accurate and detailed pricing information, consult Amazon Bedrock pricing.

SQL with Amazon Redshift offers a flexible, consumption-based pricing model. You pay only for the compute resources you use, without minimum fees or upfront commitments. Pricing is based on the number of nodes and instance hours used, with costs varying by node type. The model includes options for on-demand and reserved instances, with the latter providing cost savings for long-term commitments. For the most accurate and detailed pricing information, consult Amazon Redshift pricing.

SageMaker Data Processing, which brings together capabilities from Amazon Athena, Amazon EMR, AWS Glue, and Amazon Managed Workflows for Apache Airflow (Amazon MWAA), offers a flexible, pay-as-you-go pricing model without upfront commitments or a minimum fee. The pricing structure is based on the usage of various components, including compute resources, storage, and specific tools within the service. You pay for the actual resources consumed, which might vary depending on factors such as AWS Region, instance types, and usage patterns. For the most accurate and detailed pricing information, consult the Athena, Amazon EMR, AWS Glue, and Amazon MWAA pricing pages.

Amazon Q Developer is available within SageMaker Unified Studio to support development tasks like coding, testing, security scanning, resource optimization, and data integration. Amazon Q charges a fixed cost per user per month. For the most accurate and detailed pricing information, consult Amazon Q Developer pricing.

The next generation of SageMaker simplifies discovery, governance, and collaboration for data and AI across your lakehouse, AI models, and applications. With SageMaker Catalog, built on Amazon DataZone, you can consistently define and enforce access policies using a single permission model with fine-grained access controls.

SageMaker Catalog follows a pay-as-you-go pricing model with no upfront commitments or minimum fees. Pricing is based on four key dimensions: requests, metadata storage, compute, and AI recommendations for generating business descriptions. Each account receives some free usage per billing month, including a set amount of metadata storage, API requests, and compute units. Charges only apply if usage exceeds these free limits. The service offers core APIs that are always free and don't count towards the request limit. The rates for each of the dimensions and allowed free usage are described below.

SageMaker Catalog free usage

SageMaker Catalog, built on Amazon DataZone, provides free usage up to a limit so that you can explore its capabilities without having to worry about incurring charges. Each account where you create an Amazon DataZone domain gets 20 MB of free metadata storage, 4,000 free API requests, and 0.2 free compute units. You will only be charged the SageMaker Catalog standard pay-as-you-go rates if your account exceeds these free usage limits within a billing month. In addition, certain core APIs that power the key user experiences such as CreateDomain, CreateProject, and Search are provided free of charge and do not count towards the 4,000 free API requests.

Metadata storage

You will be charged at standard rate of $0.4 per GB in a billing month for metadata storage if you exceed your free usage 20 MB for your account. Metadata storage is the size of your domains, glossaries and terms, metadata forms, projects, environments, users, and data assets. The table below provides a few scenarios to explain how metadata storage-based pricing works.

Requests

SageMaker Catalog provides certain core APIs such as CreateDomain, CreateProject, and Search that are never charged. All the other APIs will be charged a standard rate of $10 per 100,000 requests if you exceed your free usage of 4,000 requests for your account in a billing month. A request is an interaction with SageMaker Catalog, whether through the SageMaker Unified Studio, AWS Management Console, Amazon DataZone portal, the Amazon DataZone SDK, an integrated service, or an API. Some of the common requests are invoking a search in the SageMaker Unified Studio or Amazon DataZone portal, viewing data asset details, or creating a project.

Compute

You will be charged at standard rate of $1.776 per compute unit if you exceed your included 0.2 compute units billing month for your account. SageMaker Catalog compute resources are used to ingest asset and schema metadata from your data sources, including AWS Glue databases and Amazon Redshift, and optionally to provide you with business name recommendations when publishing assets to your catalog. You are billed for successfully completed ingestion or name-generation compute jobs. The table below provides a few scenarios to explain how compute-based pricing works.