#How to Define a GPU Cost Model

This section provides configuration for metering and billing whole GPUs, whole NPUs, and HAMi virtual GPUs in a cluster.

Each resource configuration supports two billing methods:

Allocation prices can be differentiated by GPU model with the modelName label. Project quota uses a project-level base price because quota data has no GPU model dimension. A default price is used when no model-specific allocation price matches.

#Configure GPU billing

1. How to Define a Whole GPU Cost Model
2. How to Define a Whole NPU Cost Model
3. How to Define a HAMi Virtual GPU Cost Model
4. Configuration Conventions
5. Check the result and troubleshoot common issues