All Products
Search
Document Center

AnalyticDB:Pricing

Last Updated:Dec 12, 2025

This topic describes the pricing details of AnalyticDB for PostgreSQL.

Billing methods

AnalyticDB for PostgreSQL provides two billing methods:

  • Pay-as-you-go: With this billing method, you are charged on an hourly basis. This method is suitable for short-term use because you can release instances at any time to save costs.

  • Subscription: With this billing method, you must pay upfront when you create an instance. This method is suitable for long-term use and is more cost-effective than the pay-as-you-go method.

    For AnalyticDB for PostgreSQL instances in elastic storage mode and Serverless Pro mode, you can receive discounts for subscription durations of one year or longer. A one-year subscription receives a 15% discount, a two-year subscription receives a 30% discount, and a three-year subscription receives a 50% discount.

Elastic storage mode

AnalyticDB for PostgreSQL in elastic storage mode was commercially released in 2020. This product uses a dedicated architecture that integrates storage and compute resources. It supports elastic scaling and provides a wide range of specifications. This mode is available in High-availability and High-performance (Basic Edition) instance editions. The editions differ in compute specifications and the number of storage replicas. The fees for elastic storage mode consist of master node resource fees, compute resource fees, storage resource fees, and Retrieval-Augmented Generation (RAG) Service fees.

Master node resources

In AnalyticDB for PostgreSQL, master nodes are responsible for accepting service requests, optimizing queries, distributing workloads, and managing metadata and transactions. AnalyticDB for PostgreSQL master nodes are measured in Master Compute Units (CUs) and are available in multiple specifications.

Note
  • Before January 1, 2024, master node resources are free of charge.

  • Starting from January 1, 2024, master node resources with specifications of 8 CU or lower are free of charge. You are charged for master node resources with specifications of 16 CU or higher.

  • The prices of master node resources are subject to the prices on the buy page and in your bills.

Compute resources

The prices of compute resources are subject to the prices on the buy page and in your bills.

Storage resources

AnalyticDB for PostgreSQL in elastic storage mode provides two disk storage types :

  • Enterprise-class Solid-State Drive (ESSD) (Recommended): This type is applicable to High-availability and High-performance (Basic Edition) instance editions.

  • Ultra disk: This type is applicable to High-availability instance editions.

Note
  • The High-availability edition uses a dual-replica architecture. Therefore, the unit price of ESSDs for this edition is twice that of the High-performance (Basic Edition) edition.

  • The High-performance (Basic Edition) edition uses a single-replica architecture. The underlying storage medium provides triplicate replicas to ensure high reliability.

  • You can no longer purchase new instances that use ultra disks. Existing instances are not affected.

The prices vary based on the disk storage type. The prices on the buy page and in your bills prevail.

Important

On the buy page for a pay-as-you-go AnalyticDB for PostgreSQL instance in elastic storage mode, the displayed Configuration Fee is for the disk storage of a single compute node. The actual disk storage fee is calculated based on the total number of compute nodes.

RAG Service

The RAG Service for AnalyticDB for PostgreSQL is an innovative AI service that combines retrieval and generation technologies. It provides a more precise, flexible, and high-quality end-to-end RAG experience. You are charged for calls made to the RAG Service. The following tables show the pricing details.

Model type

Model name

USD/1,000 tokens

Embedding

text2vec

0.00248128189231723

text-embedding-v1

0.000994837476099427

text-embedding-v2

0.000994837476099427

text-embedding-v3

0.000994837476099427

m3e-small

0.000239493670886076

m3e-base

0.001253987

Rerank

bge-reranker-v2-minicpm-layerwise

0.00814631831023507

bge-reranker-v2-m3

0.00131577727293801

Model type

Model name

USD/1,000 images

Embedding (images)

multimodal-embedding-one-peace-v1

0

clip-vit-b-16

0.8447965828

clip-vit-b-32

0.8447965828

clip-vit-l-14

0.8447965828

clip-vit-l-14-336px

0.8447965828

clip-rn50

0.8447965828

clip-rn50x4

0.8447965828

clip-rn50x16

0.8447965828

clip-rn50x64

0.8447965828

Serverless

AnalyticDB for PostgreSQL in Serverless mode was commercially released on February 1, 2022. This product is fully cloud-native. It supports pay-as-you-go storage, requires no provisioning, and supports scaling in seconds. You can adjust compute resources based on your business needs, providing high business flexibility. It also offers hardware-accelerated caching. The fees for Serverless mode consist of master node resource fees, compute resource fees, and storage resource fees.

Master node resources

In AnalyticDB for PostgreSQL, master nodes are responsible for accepting service requests, optimizing queries, distributing workloads, and managing metadata and transactions. AnalyticDB for PostgreSQL master nodes are measured in Master Compute Units (CUs) and are available in multiple specifications.

Note
  • Before January 1, 2024, master node resources are free of charge.

  • Starting from January 1, 2024, master node resources with specifications of 8 CU or lower are free of charge. You are charged for master node resources with specifications of 16 CU or higher.

  • The prices of master node resources are subject to the prices on the buy page and in your bills.

Compute resources

Manual scheduling mode

The prices for subscription and pay-as-you-go instances are subject to the prices on the buy page and in your bills.

Storage resources

The disk storage for Serverless mode instances only supports the pay-as-you-go billing method. You do not need to purchase storage space in advance. You are billed for the storage you actually use.

The prices are subject to the prices on the buy page and in your bills.

Serverless Pro

AnalyticDB for PostgreSQL in Serverless Pro mode will be commercially released in June 2025 to replace Serverless mode. This product is fully cloud-native. It supports pay-as-you-go storage, requires no provisioning, and supports scaling in seconds. You can adjust compute resources based on your business needs, providing high business flexibility. It also offers hardware-accelerated caching. The fees for Serverless Pro mode consist of master node resource fees, compute resource fees, and storage resource fees.

Master node resources

In AnalyticDB for PostgreSQL, master nodes are responsible for accepting service requests, optimizing queries, distributing workloads, and managing metadata and transactions. AnalyticDB for PostgreSQL master nodes are measured in Master Compute Units (CUs) and are available in multiple specifications.

Note
  • Before January 1, 2024, master node resources are free of charge.

  • Starting from January 1, 2024, master node resources with specifications of 8 CU or lower are free of charge. You are charged for master node resources with specifications of 16 CU or higher.

  • The prices of master node resources are subject to the prices on the buy page and in your bills.

Compute resources

The compute resources for Serverless Pro instances only support the pay-as-you-go billing method. You do not need to purchase compute resources in advance. You are billed for the resources you actually use.

The prices of compute resources are subject to the prices on the buy page and in your bills.

Storage resources

The cache storage and disk storage for Serverless Pro instances only support the pay-as-you-go billing method. You do not need to purchase storage space in advance. You are billed for the storage you actually use.

The prices are subject to the prices on the buy page and in your bills.

Supabase

Compute node specifications

Region

Subscription (USD/month)

Pay-as-you-go (USD/hour)

2C2G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

6.838

0.0142458333333333

China (Hong Kong)

15.054

0.0313624999999999

  • Indonesia (Jakarta)

  • Singapore

13.585

0.0283020833333333

2C4G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

25.779

0.0537062500000001

China (Hong Kong)

45.786

0.0953875

Indonesia (Jakarta)

50.518

0.105245833333333

Singapore

48.971

0.102022916666667

2C8G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

39.228

0.081725

  • China (Hong Kong)

  • Indonesia (Jakarta)

  • Singapore

57.852

0.120525

4C16G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

78.456

0.163450000000001

  • China (Hong Kong)

  • Indonesia (Jakarta)

  • Singapore

115.704

0.24105

8C32G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

156.912

0.3269

  • China (Hong Kong)

  • Indonesia (Jakarta)

  • Singapore

231.408

0.4821

16C64G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

313.824

0.653800000000001

  • China (Hong Kong)

  • Indonesia (Jakarta)

  • Singapore

462.816

0.9642

32C128G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

627.648

1.3076

  • China (Hong Kong)

  • Indonesia (Jakarta)

  • Singapore

925.632

1.9284

64C256G

  • China (Beijing)

  • China (Hangzhou)

  • China (Shanghai)

  • China (Shenzhen)

  • China (Hohhot)

  • China (Zhangjiakou)

  • China (Chengdu)

1255.296

2.61520000000001

  • China (Hong Kong)

  • Indonesia (Jakarta)

  • Singapore

1851.264

3.8568