This topic describes how Elastic Algorithm Service (EAS) is charged.

You are charged EAS fees for the resources that are used to deploy model services.
  • You can deploy model services in shared resource groups and dedicated resource groups. Bills are generated based on the used resources of resource groups. For more information about the differences between the two types of resource groups, see t1895815.html#multiTask4399.
  • If you deploy a model service in a shared resource group and a dedicated resource group, you are charged for using the resources of both resource groups.
The following table lists the billing rules for shared resource groups and dedicated resource groups.
Resource Billing item Billing method Billing rule Method to stop billing
Shared resource group The duration that model services have been running (the amount of time for which the model services have occupied shared resources) Pay-as-you-go Bills are generated based on the amount of shared resources that are occupied by model services. The billing starts immediately after model services are created. Stop model services
Dedicated resource group The duration that resource groups have been running Pay-as-you-go Only dedicated resource groups are charged. Model services deployed in dedicated resource groups are not charged. The billing starts immediately after pay-as-you-go dedicated resource groups are created. Stop dedicated resource groups
Subscription No

Billing rules for shared resource groups

1 compute hour is equivalent to the usage of 1 CPU core and 4 GB of memory.

Each model service is billed based on the following formula:
Bill amount of each model service = Number of compute hours × (Unit price/60) × Usage duration
Note Usage duration is measured in minutes.
  • The system starts to bill the resources that are used by a model service after the model service enters the Running state. The system stops billing the resources that are used by a model service after the model service enters the Stopped state.
    Notice Stop unused model services to avoid unnecessary costs.
  • Bills are generated from the time when a model service starts to run and consume resources. The system stops billing immediately after a model service is suspended and releases resources.
  • If you scale out a model service, newly added resources are billed after the scale-out activity is complete. If you scale in a model service, released resources are no longer billed, and only the remaining resources are billed.
  • Fees are charged on a per minute basis. No fee is charged if a model service runs for less than one minute.

Prices of shared resource groups

The catalog prices of pay-as-you go shared resource groups are hourly prices. However, fees are charged on a per minute basis. You can calculate the unit prices per minute by dividing the prices listed in the following table by 60. The following table lists the unit prices per hour.
Resource type Price (USD/compute hour/hour) Region
CPU 0.06
  • China (Beijing)
  • China (Shanghai)
  • China (Hangzhou)
  • China (Shenzhen)
  • China (Hong Kong)
  • Singapore (Singapore)
  • Indonesia (Jakarta)
  • India (Mumbai)
  • Germany (Frankfurt)
Note Shared resource groups do not provide GPU resources. If you require GPU resources, you must purchase dedicated resource groups.

Billing examples for shared resource groups

For example, you used CPU resources in the China (Hangzhou) region to deploy a model service. The model service occupies 2 compute hours (2 CPU cores + 8 GB), and entered the Running state at 09:00:00 (UTC+8) on June 3, 2019. You reduced the occupied resources to 1 compute hour (1 CPU core + 4 GB) at 10:00:00 (UTC+8) on June 3, 2019. Then, you increased the occupied resources to 4 compute hours (4 CPU cores + 16 GB) at 11:00:00 (UTC+8) on June 3, 2019. At 12:00:00 (UTC+8) on June 3, 2019, the model service stopped running. The total fee is calculated based on the following formula:
Bill amount = 2 × (0.06/60) × 60 + 1 × (0.06/60) × 60 + 4 × (0.06/60) × 60 = USD 0.42

Billing rules for dedicated resource groups

  • Subscription
    The following formula describes how to calculate the subscription fee for each subscription dedicated resource group:
     Bill amount of each resource group = Number of resources × Unit price × Subscription duration
    • Valid range of subscription duration: 1 to 12 months. Each month contains 30 calendar days.
    • After you purchase a dedicated resource group, the resource group is free of charge on the current day, and the subscription takes effect from the next day. For example, you purchased a dedicated resource group on July 31, 2019, and the subscription duration is one month. Then, the resource group expired at 00:00:00 (UTC+8) on August 31, 2019.
    • Some resources may not be available in one or more regions for a short period of time. During the time period, you cannot purchase dedicated resource groups in the regions.
  • Pay-as-you-go
    Each pay-as-you-go dedicated resource group is billed based on the following formula:
    Bill amount of each resource group = Number of resources × (Unit price/60) × Usage duration
    Note Usage duration is measured in minutes.
    • The billing starts immediately after a resource group enters the Running state. The billing stops immediately after a resource group enters the No Node state.
      Notice Stop unused resource groups to avoid unnecessary costs.
    • Bills are generated from the time when a resource group enters the Running state. The system stops billing immediately after a resource group releases all occupied resources and enters the No Node state.
    • If you scale out a resource group, newly added resources are billed after the scale-out activity is complete. If you scale in a resource group, released resources are no longer billed, and only the remaining resources are billed.
    • Fees are charged on a per minute basis. No fee is charged if a resource group runs for less than one minute.
    • Some resources may not be available in one or more regions for a short period of time. During the time period, you cannot purchase dedicated resource groups in the regions.

Prices of dedicated resource groups

The following table lists the prices of subscription dedicated resource groups.
Model GPU specification CPU specification Price (USD/month)
Mainland China China (Hong Kong) Singapore (Singapore) Indonesia (Jakarta) India (Mumbai) Germany (Frankfurt)
ecs.c5.6xlarge N/A 24 CPU Core+48 GB 365 366 596 569 492 535
ecs.g5.6xlarge N/A 24 CPU Core+96 GB 495 495 744 745 592 786
ecs.gn5i-c4g1.xlarge 1 NVIDIA Tesla P4 4 CPU Core+16 GB 452 N/A N/A N/A N/A N/A
ecs.gn5i-c8g1.2xlarge 1 NVIDIA Tesla P4 8 CPU Core+32 GB 543 N/A N/A N/A N/A N/A
ecs.gn6i-c4g1.xlarge 1 Tesla T4 4 CPU Core+15 GB 570 691 727 727 776 722
ecs.gn6i-c8g1.2xlarge 1 Tesla T4 8 CPU Core+31 GB 686 819 862 862 908 870
ecs.gn6i-c16g1.4xlarge 1 Tesla T4 16 CPU Core+62 GB 805 1076 1132 1132 1170 1166
ecs.gn6i-c24g1.6xlarge 1 Tesla T4 24 CPU Core+93 GB 843 1349 1420 1420 1396 1472
ecs.gn5-c4g1.xlarge 1 NVIDIA P100 4 CPU Core+30 GB 627 928 977 977 929 1006
ecs.gn5-c8g1.2xlarge 1 NVIDIA P100 8 CPU Core+60 GB 755 1118 1177 1177 1118 1212
ecs.gn5-c28g1.7xlarge 1 NVIDIA P100 28 CPU Core+112 GB 1171 1609 1693 1693 1734 1744
ecs.gn6v-c8g1.2xlarge 1 NVIDIA V100 8 CPU Core+32 GB 1297 N/A 2381 N/A N/A N/A
Note Subscription dedicated resource groups are available only in the China (Hangzhou), China (Shanghai), China (Beijing), and China (Shenzhen) regions.
The catalog prices of pay-as-you go dedicated resource groups are hourly prices. However, fees are charged on a per minute basis. You can calculate the unit prices per minute by dividing the prices listed in the following table by 60. The following table lists the unit prices per hour.
Model GPU specification CPU specification Price (USD/hour)
Mainland China China (Hong Kong) Singapore (Singapore) Indonesia (Jakarta) Germany (Frankfurt) India (Mumbai)
ecs.g6.4xlarge N/A 16 CPU Core+64 GB 0.66 1.2 1.08 1.08 1.08 0.90
ecs.g6.6xlarge N/A 24 CPU Core+96 GB 1.02 1.8 1.68 1.68 1.62 1.38
ecs.c5.6xlarge N/A 24 CPU Core+48 GB 1.26 1.2 1.26 1.20 1.14 1.08
ecs.g5.6xlarge N/A 24 CPU Core+96 GB 1.8 1.68 1.74 1.62 1.56 1.32
ecs.gn5i-c4g1.xlarge 1 NVIDIA Tesla P4 4 CPU Core+16 GB 1.62 N/A N/A N/A N/A N/A
ecs.gn5i-c8g1.2xlarge 1 NVIDIA Tesla P4 8 CPU Core+32 GB 1.98 N/A N/A N/A N/A N/A
ecs.gn6i-c4g1.xlarge 1 Tesla T4 4 CPU Core+15 GB 1.98 1.32 1.50 1.44 1.38 1.50
ecs.gn6i-c8g1.2xlarge 1 Tesla T4 8 CPU Core+31 GB 2.40 1.56 1.80 1.68 1.68 1.74
ecs.gn6i-c16g1.4xlarge 1 Tesla T4 16 CPU Core+62 GB 2.76 2.10 2.34 2.22 2.28 2.28
ecs.gn6i-c24g1.6xlarge 1 Tesla T4 24 CPU Core+93 GB 2.94 2.64 2.94 2.76 2.88 2.70
ecs.gn5-c4g1.xlarge 1 NVIDIA P100 4 CPU Core+30 GB 2.16 2.04 2.16 2.16 1.92 2.04
ecs.gn5-c8g1.2xlarge 1 NVIDIA P100 8 CPU Core+60 GB 2.64 2.46 2.58 2.58 2.28 2.46
ecs.gn5-c28g1.7xlarge 1 NVIDIA P100 28 CPU Core+112 GB 4.08 3.54 3.72 3.72 3.84 3.78
ecs.gn6v-c8g1.2xlarge 1 NVIDIA V100 8 CPU Core+32 GB 4.50 N/A 5.22 N/A N/A N/A
Note Pay-as-you-go dedicated resource groups are available only in the China (Hangzhou), China (Shanghai), China (Beijing), and China (Shenzhen) regions.

Billing examples for dedicated resource groups

For example, you subscribe to two T4 GPUs in the China (Hangzhou) region for three months. The specification of each GPU is 4 CPU cores and 15 GB of memory. The bill amount is calculated based on the following formula:
Bill amount = 2 × 570 × 3 = USD 3,420
If you purchase two pay-as-you-go ecs.g6.6xlarge (24 CPU cores + 96 GB of memory) ECS instances in the China (Hangzhou) region, and use the instances for 45 minutes, the bill amount is calculated based on the following formula:
Bill amount = 2 × (1.02/60) × 45 = USD 1.53