Before you deploy model services with EAS, understand the billing rules for different resource types to manage costs effectively. EAS supports pay-as-you-go, subscription, and Serverless billing for the EAS billable items and billing methods.
Billing overview
Deploying services with EAS may incur charges for compute resources, system disks, and dedicated gateways:
-
Compute resources: public resources, EAS dedicated resources, General Computing Resources 2.0, and Lingjun AI Computing resources.
General Computing Resources 1.0 is available only in DSW and DLC. General Computing Resources 2.0 is available in DSW, DLC, and EAS.
-
System disk (optional): A free quota is provided (30 GiB for public resources, 200 GiB for EAS dedicated resources and General Computing Resources 2.0). Usage beyond the free quota is billed separately.
-
Fully managed dedicated gateway(optional): Services use a shared gateway by default (free of charge). You can also purchase a fully managed dedicated gateway (manual setup required).
EAS supports two billing methods:
-
Pay-as-you-go: Billed based on resource usage duration, not service invocations. Suitable for workloads with unpredictable or fluctuating demand.
NoteFor SDWebUI, EAS offers a Serverless version. Service deployment is free of charge. You're billed only for the actual processing time when the service handles requests.
-
Subscription: Pay upfront at a discounted rate. Suitable for long-term, stable workloads.
If you use other Alibaba Cloud services such as Elastic IP Address (EIP), OSS, or NAS, the charges for those services are billed separately.
Billing period
Pay-as-you-go
For pay-as-you-go resources, note when billing starts and stops to avoid unexpected charges. Stop or delete model services that are no longer in use.
|
Resource type |
Billing start time |
Billing end time |
|
Public resources |
When the service instance provisions resources (billing applies even without service invocations). |
When the model service is stopped (resources are released). Important
You must stop or delete the service instance. |
|
EAS dedicated resources / General Computing Resources 2.0 |
When the machine is created and enters the Running state (billing applies even without deployed services). |
Important
The Stopped state is triggered only by overdue payments. There is currently no way to manually stop a machine. Delete the pay-as-you-go machine resource to stop billing. |
|
System disk |
When the system disk is purchased. |
|
|
Dedicated gateway |
When the dedicated gateway is created. |
When the dedicated gateway is deleted. |
Common misconceptions about stopping pay-as-you-go billing for public resources:
-
Billing continues during failed service startup: Billing is based on the duration of underlying Pod resource usage. Regardless of whether the service starts successfully, resources are billed for the entire occupation period. You can check whether billing is in progress by checking whether the service instance is marked as Billing in progress on the service overview page.
-
Service anomaly auto-retry may incur charges: When a deployment anomaly occurs (such as waiting or failure due to insufficient resources), the system continues to automatically retry for up to 48 hours. Monitor the service status and stop or delete abnormal services promptly to avoid continuous billing after a retry succeeds with no one maintaining the service.
-
Instances in the Stopped state: Services stopped by the system due to overdue payments don't automatically restart after you top up your account. Services that you manually stop are not restarted by scaling policies. To resume the service and billing, manually start the instance in the console or through an API call.
Scaling behavior
-
Scale-out: New resources are billed from the time the scale-out succeeds.
-
Scale-in: Released resources stop billing from the time the release succeeds. Remaining resources continue to be billed.
Subscription
|
Resource type |
Billing start time |
Billing end time |
|
EAS dedicated resources / General Computing Resources 2.0 |
00:00:00 the day after purchase. |
Expiration time. |
|
AI computing resources (Lingjun AI Computing resources) |
00:00:00 the day after purchase. |
Expiration time. |
|
System disk |
When the system disk is purchased. |
Expiration time. |
|
Dedicated gateway |
When the gateway is purchased. |
Expiration time. |
Serverless services
Serverless service deployment is free of charge. You're billed only for the actual runtime while the service processes requests.
-
Example: If you submit an image generation request in the WebUI and it takes 10 seconds to generate the image, the billing duration is 10 seconds.
-
Limitations: Currently, only SDWebUI supports Serverless deployment. Check the deployment page for pricing details.
Billable items
The following table describes the billing methods for different resources. Note:
-
Except for Serverless services, pay-as-you-go billing is measured in minutes.
-
For pay-as-you-go public resources, stop unused model services to avoid unnecessary charges.
-
Some instance types may be temporarily unavailable in certain regions.
|
Billable item |
Billing method |
Billing formula |
Unit price |
|
Public resources |
Pay-as-you-go (specified machine resources) |
|
For pricing details, see Table 1. Specified machine resources. The prices shown here are hourly for readability. Actual billing is per minute. Divide the hourly price by 60 to get the per-minute price. |
|
Pay-as-you-go (specified instance type) |
|
Pricing varies by region and instance type. Refer to the console for actual prices. For the list of supported instance types, see Appendix: Public resource group instance types. |
|
|
EAS dedicated resources / General Computing Resources 2.0 |
Pay-as-you-go |
|
For pricing details, visit the EAS dedicated machine pay-as-you-go purchase page. |
|
Subscription |
|
For pricing details, visit the EAS dedicated machine subscription purchase page. |
|
|
AI computing resources (Lingjun AI Computing resources) |
Subscription / Pay-as-you-go |
Purchase Lingjun AI Computing resources and use the corresponding resource quota to deploy EAS services. For AI computing resource billing details, see Billing of AI computing resources. |
|
|
System disk |
Pay-as-you-go |
|
For pricing details, visit the Block Storage (PL1 ESSD) page. |
|
Subscription |
|
For pricing details, visit the Block Storage (PL1 ESSD) page. |
|
|
Dedicated gateway |
Pay-as-you-go |
|
For pricing details, visit the EAS dedicated gateway pay-as-you-go page. |
|
Subscription |
|
For subscription pricing details, visit the EAS dedicated gateway subscription page. |
|
|
Service inference (Serverless) |
Billed by actual runtime while the service processes requests |
|
Currently, only SDWebUI supports Serverless deployment. Check the deployment page for pricing details. |
Table 1. Specified machine resources
The following pricing information is for reference only. Actual prices are shown on the purchase page.
|
Resource type |
Price |
|
CPU |
0.03 (USD/core/hour) |
|
Memory |
0.004 (USD/GB/hour) |
Usage notes
All prices in this topic are for reference only. Actual prices are shown on your invoice.
Billing examples
The following billing examples are for reference only. Actual prices are shown on the purchase page.
Public resource group billing examples
Pay-as-you-go example
-
Scenario: Deploy a model service in the public resource group using specified machine resources in the China (Hangzhou) region.
-
09:00:00 — The service enters the Running state with 2 CPU cores + 8 GB memory.
-
10:00:00 — Scale-in completes. Resources are reduced to 1 CPU core + 4 GB memory.
-
11:00:00 — Scale-out completes. Resources increase to 4 CPU cores + 16 GB memory.
-
12:00:00 — The service is stopped.
-
-
Cost calculation:
Bill amount = (2 × 0.03 + 8 × 0.004) + (1 × 0.03 + 4 × 0.004) + (4 × 0.03 + 16 × 0.004) = 0.322 USD
EAS dedicated resource billing examples
Subscription example
-
Scenario:
Purchase 2 machines, each with 4 CPU cores + 15 GB GPU T4, in the China (Hangzhou) region with a 3-month subscription at 570 USD/month (actual price may vary).
-
Cost calculation:
Total = 2 × 570 × 3 = 3420 USDPay-as-you-go example
-
Scenario:
Purchase 2 ecs.g6.6xlarge machines (24 CPU cores + 96 GB) in the China (Hangzhou) region using pay-as-you-go billing for 45 minutes at 1.02 USD/hour (actual price may vary).
-
Cost calculation:
Bill amount = 2 × (1.02 / 60) × 45 = 1.53 USDSystem disk billing examples
Subscription example
-
Scenario: Purchase 2 machine resources with EAS dedicated resources on subscription, each with a 300 GiB system disk, in the China (Hangzhou) region for 3 months.
-
Cost calculation:
Bill amount = 2 × (300 - 200) × 0.153 × 3 = 91.8 USDPay-as-you-go example
-
EAS dedicated resources
-
Scenario: Purchase 2 machine resources with EAS dedicated resources on pay-as-you-go, each with a 300 GiB system disk, in the China (Hangzhou) region for 5 hours.
-
Cost calculation:
Bill amount = 2 × (300 - 200) × 0.000319 × 5 = 0.319 USD -
Public resource group
-
Scenario: Purchase 2 instance nodes in the public resource group on pay-as-you-go, each with a 300 GiB system disk, in the China (Hangzhou) region for 5 hours.
-
Cost calculation:
Bill amount = 2 × (300 - 30) × 0.000319 × 5 = 0.8613 USDFAQ
Q: How do I view deduction details?
On the Billing Details page, filter by PAI to view the detailed bill. For details, see View PAI billing details.
Q: How do I view pay-as-you-go billing details and the corresponding resource ID for EAS public resources?
-
Get the resource instance ID:
The resource instance ID in EAS public resource pay-as-you-go bills uses the format
<service instance ID>;<service ID>. To find it:-
Log on to the PAI console and go to the Elastic Algorithm Service (EAS) page. On the Inference Service tab, find the service deployed on public resources and get the service ID.
-
Click the service name to go to the details page. At the bottom of the Overview tab, get the service instance ID from the Service Instance section.
If the service is stopped, the instance ID isn't available here.
-
-
View billing details: Go to the Bill Details page. Filter by product name EAS Service Pay-As-You-Go and resource instance ID
<service instance ID>;<service ID>.
If you receive an overdue payment notification but can't find the corresponding instance in the console, check the service lists in all regions (especially China (Shanghai) and China (Hangzhou)) for any undeleted residual instances.
Q: Why did PAI-EAS generate a 0 CNY order after activation when no instances were created?
Activating EAS pay-as-you-go billing is free. The 0 CNY order simply confirms that the service module has been activated. Billing occurs only when EAS instances (Pods) are actually created. If no running EAS instances are found in the console, no charges are incurred.
Q: How is the SLA compensation calculated for PAI-EAS service issues?
EAS SLA compensation is calculated per individual service (task), based on the downtime of the affected service and all resources used by that service, not on the total resources across your account.
Q: Why am I still being charged after stopping an EAS service?
-
Bill delay: For DSW/EAS pay-as-you-go billing, there's a delay in bill generation and deduction. The charge notification you received may reflect actual usage before you stopped the instance, not charges incurred after stopping.
-
EAS service running in the background: A service that you don't use is still running. As long as the service isn't manually stopped or deleted, it continues to incur charges for resource usage. Before deleting, make sure all necessary data is backed up.
Q: Why am I still being charged when using EAS trial resources?
-
Cause: When deploying
stable-diffusion-webuiorcomfyuiin EAS, if the selected image version includes the-apior-clustersuffix, an async queue resource is provisioned during deployment. This resource is billed and can't be offset by EAS trial resources. -
Solution: When deploying
stable-diffusion-webuiorcomfyui, select the standard version — an image version with-standardor no suffix.
Q: Do I need to manually release the PAI savings plan after it expires?
No. The PAI savings plan is a pre-paid product that expires automatically. No manual release or deletion is needed, and no additional charges are incurred. After receiving the expiration notification, you don't need to take any action if you no longer need the service. If you have running EAS service instances in the PAI console that you no longer use, stop or delete them to avoid pay-as-you-go charges.
Q: How do I check whether my savings plan or free trial credits are exhausted?
Log on to the Billing and Cost Management console, go to the My Subscriptions page to check the remaining balance and validity period of the savings plan. If the credits are exhausted or expired, continued use of EAS public resource group services is billed at standard pay-as-you-go rates.