This topic describes the billing rules for each module in Machine Learning Studio.
Billing
One compute hour is equivalent to the usage of one CPU core and 4 GB of memory in one hour.
The following formula describes how to calculate the number of compute hours.
The following formula shows how to calculate the number of compute hours when you
use 2 CPU cores and 5 GB of memory in one hour.
Number of compute hours = Max (Number of CPU cores × Usage duration, Memory size × Usage duration/4)
Note The memory size is measured in GB and the usage duration is measured in hours.
Number of compute hours = Max (2 × 1, 5 × 1/4) = 2
The following formula describes how to calculate the bill amount.
Bill amount= Number of compute hours × Unit price
Pricing
Machine Learning Studio is billed on the payasyougo basis. The following table
shows the prices of the commonly used modules in Machine Learning Studio.
Module  Description  Price (USD/compute hour)  Region 

Data processing  Includes components such as data preprocessing and feature engineering.  0.16 

Data analysis  Includes components such as statistical analysis, machine learning, time series, and financials.  0.21  
Text analysis  The algorithm for text analysis.  0.27 
Note MaxCompute SQL statements are the basic units for executing SQL scripts, JOIN, UNION,
and filtering and mapping components. Therefore, bills may be generated for MaxCompute
when you use Machine Learning Studio. For more information, see Billing method.
The following table shows the prices of deep learning components.
Module  Description  Price (USD/GPU/hour)  Region 

Deep learning (M40)  Includes deep learning frameworks such as TensorFlow, Caffe, and MXNet.  1.308  China (Shanghai) 
Deep learning (P100)  1.872  China (Beijing) 
Examples
For more information about billing examples, see Billing examples of Machine Learning Studio.