This topic describes how you are billed for Machine Learning Platform for AI (PAI).

Billing rules

PAI supports two activation methods that are subject to different billing rules.
  • Standalone activation

    You perform a standalone activation for PAI. Standalone activations are free-of-charge. When you activate PAI, a default workspace is created. However, you are charged when you use PAI modules. You can view the billing details of these modules in PAI bills. For more information about the billing details of each PAI module, see the "Billing description" section of this topic.

  • Activate PAI together with other cloud services
    You can activate PAI together with other cloud services, such as Object Storage Service (OSS), MaxCompute, and DataWorks. Each cloud service is billed separately. For more information about the billing details of these services, see the following topics:

Billing description

The following table describes the billing details of each PAI module.
ModuleDescriptionBillable itemBilling methodReferences
iTAGiTAG is an intelligent data labeling platform. You can create data labeling tasks for the following data formats: image, text, video, and audio. You can also create multimodal data labeling tasks. None. iTAG is available for free trial.N/AN/A
Machine Learning Designer (previously known as Machine Learning Studio)Machine Learning Designer is a cloud-native visual modeling tool that you can use to develop machine learning algorithms.
  • You are charged when you run machine learning components.
  • Pay-as-you-go
Billing of Machine Learning Designer
Data Science Workshop (DSW)DSW is a cloud-based integrated development environment (IDE) that integrates open source JupyterLab and provides a variety of plug-ins to support custom development.
  • You are charged when you use Individual Edition resources.
  • You are charged when you use GPU On-sale Edition resources.
  • Pay-as-you-go
  • Subscription
Billing of DSW
Deep Learning Containers (DLC)DLC is a cloud-native platform for basic AI computing jobs. DLC provides an elastic, stable, user-friendly, and high-performance environment for model training.
  • You are charged when you use the public resource group.
  • If you use self-managed resource groups, you are charged for the Container Service for Kubernetes (ACK) clusters, including the network and storage resources. The charges for these resources are billed in ACK instead of DLC.
  • Pay-as-you-go
  • Subscription
Billing of DLC
Elastic Algorithm Service (EAS)You can deploy a trained model as a RESTful API in EAS to perform real-time inference.
  • You are charged when you use the public resource group.
  • You are charged when you use dedicated resource groups.
  • Pay-as-you-go
  • Subscription
Billing of EAS
Scenario-based solutionThis module provides features such as multimedia analysis, and smart retail.
  • You are charged when you use smart retail.
  • You are charged when you use multimedia analysis.
  • Pay-as-you-go
  • Subscription
Billing of scenario-based solution

References