Moderation Agent: Introduction and billing-AI Guardrails(AI Guardrails)-阿里云帮助中心

This topic describes the features and billing of the Moderation Agent.

Product features

Feature overview

Content Moderation allows you to create and configure custom Moderation Agents. This feature is built on large language models (LLMs). By flexibly configuring the interaction content, you can quickly detect and filter content based on the custom detection categories of your business. You can currently configure Moderation Agents of the following modalities:

Text modality: The content to be detected is plain text.
Image modality: The content to be detected is a pure image.
Image-text multi-modal: The content to be detected is a mix of image and text.

Billing

The Moderation Agent currently supports only the pay-as-you-go billing method.

Metering method

After you activate the Content Moderation Enhanced Edition service, the pay-as-you-go billing method is used by default, and no fee is charged if you do not call the service. The billing logic of the Moderation Agent detection API is as follows:

Text: The billing is related to the text length and the total number of characters in the custom prompt. Usage is metered once every 2,000 characters, and the API currently supports a maximum text length of 2,000 characters. Custom prompts within 3,000 characters are free of charge. For the portion that exceeds 3,000 characters, one additional metering is added for every 3,000 characters, and a portion of less than 3,000 characters is counted as 3,000 characters.
Image: The billing is related to the number of images and the total number of characters in the custom prompt. Usage is metered once per image, and the API currently supports only one image. Custom prompts within 3,000 characters are free of charge. For the portion that exceeds 3,000 characters, one additional metering is added for every 3,000 characters, and a portion of less than 3,000 characters is counted as 3,000 characters.
Image-Text Multimodal: The billing is related to the text length, the number of images, and the total number of characters in the custom prompt. Text is metered once every 2,000 characters, and the API currently supports a maximum text length of 2,000 characters. Images are metered once per image, and the API currently supports only one image. Custom prompts within 3,000 characters are free of charge. For the portion that exceeds 3,000 characters, one additional metering is added for every 3,000 characters, and a portion of less than 3,000 characters is counted as 3,000 characters.

For example, after pay-as-you-go billing is enabled, the following examples show how usage is metered and billed when you configure and call a Moderation Agent:

Scenario 1: The custom prompt for the text content is configured with 2,000 characters, and the usage is metered once. In this call, the text is metered once, and the custom prompt does not exceed 3,000 characters and is free of charge.
Scenario 2: The custom prompt for the image content is configured with 4,000 characters, and the usage is metered twice. In this call, the image is metered once, and the 1,000 characters of the prompt that exceed the free quota add one additional metering.
Scenario 3: The content contains both text and images, and the custom prompt is configured with 4,000 characters. The usage is metered three times. In this call, the text is billed once, the image is billed once, and the 1,000 characters of the prompt that exceed the free quota add one additional metering.

Unit prices

China regions (Shanghai, cn-Shanghai)

Modality	Review type	Unit price	Description
Text	Moderation Agent - Text - Per Call - Basic text_llm_standard	CNY 20 per 10,000 calls	This item is metered when you select the text moderation large model Qwen3.7-Flash or Qwen3-Flash (to be deprecated) as the large model.
	Moderation Agent - Text - Per Call - Advanced text_llm_advanced	CNY 40 per 10,000 calls	This item is metered when you select Qwen3.8-Flash or Qwen3-Plus (to be deprecated) as the large model.
	Moderation Agent - 3.6 - Per Call - Basic text_llm36_tier_1	CNY 60 per 10,000 calls	This item is metered when you select Qwen3.6-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Advanced text_llm36_tier_2	CNY 100 per 10,000 calls	This item is metered when you select Qwen3.7-Plus or Qwen3.6-Plus as the large model.
Image	Moderation Agent - Text - Per Call - Advanced text_llm_advanced	CNY 40 per 10,000 calls	This item is metered when you select Qwen3.8-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Basic text_llm36_tier_1	CNY 60 per 10,000 calls	This item is metered when you select Qwen3.6-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Advanced text_llm36_tier_2	CNY 100 per 10,000 calls	This item is metered when you select Qwen3.7-Plus or Qwen3.6-Plus as the large model.
	Moderation Agent - Per Call - Entry image_vl_basic	CNY 15 per 10,000 calls	This item is metered when you select Qwen3.7-Flash as the large model.
	Moderation Agent - Per Call - Basic image_vl_standard	CNY 45 per 10,000 calls	This item is metered when you select Qwen3-VL-Plus (to be deprecated) as the large model.
Image-text multi-modality	Moderation Agent - Text - Per Call - Advanced text_llm_advanced	CNY 40 per 10,000 calls	This item is metered when you select Qwen3.8-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Basic text_llm36_tier_1	CNY 60 per 10,000 calls	This item is metered when you select Qwen3.6-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Advanced text_llm36_tier_2	CNY 100 per 10,000 calls	This item is metered when you select Qwen3.7-Plus or Qwen3.6-Plus as the large model.
	Moderation Agent - Per Call - Entry image_vl_basic	CNY 15 per 10,000 calls	This item is metered when you select Qwen3.7-Flash as the large model.
	Moderation Agent - Per Call - Basic image_vl_standard	CNY 45 per 10,000 calls	This item is metered when you select Qwen3-VL-Plus (to be deprecated) as the large model.

International regions (Singapore, ap-southeast-1)

Modality	Review type	Unit price	Description
Text	Moderation Agent - Text - Per Call - Basic - Global Edition text_llm_standard_cb	CNY 50 per 10,000 calls	This item is metered when you select the text moderation large model Qwen3.7-Flash or Qwen3-Flash (to be deprecated) as the large model.
	Moderation Agent - Text - Per Call - Advanced - Global Edition text_llm_advanced_cb	CNY 100 per 10,000 calls	This item is metered when you select Qwen3.8-Flash or Qwen3-Plus (to be deprecated) as the large model.
	Moderation Agent - 3.6 - Per Call - Basic - Global Edition text_llm36_tier_1_cb	CNY 150 per 10,000 calls	This item is metered when you select Qwen3.6-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Advanced - Global Edition text_llm36_tier_2_cb	CNY 250 per 10,000 calls	This item is metered when you select Qwen3.7-Plus or Qwen3.6-Plus as the large model.
Image	Moderation Agent - Text - Per Call - Advanced - Global Edition text_llm_advanced_cb	CNY 100 per 10,000 calls	This item is metered when you select Qwen3.8-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Basic - Global Edition text_llm36_tier_1_cb	CNY 150 per 10,000 calls	This item is metered when you select Qwen3.6-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Advanced - Global Edition text_llm36_tier_2_cb	CNY 250 per 10,000 calls	This item is metered when you select Qwen3.7-Plus or Qwen3.6-Plus as the large model.
	Moderation Agent - Per Call - Entry - Global Edition image_vl_basic_cb	CNY 38 per 10,000 calls	This item is metered when you select Qwen3.7-Flash as the large model.
	Moderation Agent - Per Call - Basic - Global Edition image_vl_standard_cb	CNY 113 per 10,000 calls	This item is metered when you select Qwen3-VL-Plus (to be deprecated) as the large model.
Image-text multi-modality	Moderation Agent - Text - Per Call - Advanced - Global Edition text_llm_advanced_cb	CNY 100 per 10,000 calls	This item is metered when you select Qwen3.8-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Basic - Global Edition text_llm36_tier_1_cb	CNY 150 per 10,000 calls	This item is metered when you select Qwen3.6-Flash as the large model.
	Moderation Agent - 3.6 - Per Call - Advanced - Global Edition text_llm36_tier_2_cb	CNY 250 per 10,000 calls	This item is metered when you select Qwen3.7-Plus or Qwen3.6-Plus as the large model.
	Moderation Agent - Per Call - Entry - Global Edition image_vl_basic_cb	CNY 38 per 10,000 calls	This item is metered when you select Qwen3.7-Flash as the large model.
	Moderation Agent - Per Call - Basic - Global Edition image_vl_standard_cb	CNY 113 per 10,000 calls	This item is metered when you select Qwen3-VL-Plus (to be deprecated) as the large model.

Usage

Integration

The Moderation Agent supports integration by using SDKs and native HTTPS calls:

To integrate by using an SDK, see Moderation Agent service SDK integration.
To integrate by using native API calls, see Moderation Agent synchronous detection API.

Console operations

You can configure a custom Moderation Agent in the Content Moderation console. For more information about the operations, see Build custom detection with Moderation Agent.