Billing Items

ModelArts billing items during AI development include storage resources and compute resources.

Storage resources: Fees are generated for the used OBS storage and EVS storage (applicable only to notebook instances).
Compute resources: Fees are generated for the used ModelArts compute resources.

Storage Resources

**Table 1** Billing for storage resources
Billing Item	Description
OBS	ModelArts uses OBS to store data and models, which incurs fees. For details, see OBS Pricing Details.

Compute Resources

ModelArts allows you to select proper compute resources for AI development based on your region and required service type. For details, see ModelArts Pricing Details.

For details about how to use compute resources for ModelArts, see Pay-Per-Use.

**Table 2** Billing for compute resources
Billing Item	Description
AI development lifecycle	Machine learning and deep learning algorithm development and deployment for developers with AI development experience, including data processing, model development, model training, as well as AI application management and deployment. The following items are billed: Development environment (notebook instances) Model training (training jobs) Service deployment (real-time services)
ExeML	Automatic design, tuning, and training of models, as well as service deployment based on labeled data for developers with few AI development experience, enabling code-free AI development. Resources are billed for ExeML-powered job training and AI application deployment. The following items are billed: ExeML-powered training jobs ExeML-powered AI application deployment NOTE: Only pay-per-use billing is supported.