Model usage

View usage for Alibaba Cloud Model Studio models.

For more information about free quotas, see New user free quota. Manage free quota usage on the Free Quota page.

Availability

• Models: All models in the Model list, including fine-tuned models that are based on them.

View free quota usage

In the console

1. Go to the Free Quota page. Select a model type tab to view the free quota usage for each model.
2. In the All Models table, find specific models by searching, sorting, and filtering. Manage free quota settings by toggling the Free Quota Only switch or using batch operations.

View model usage

View usage on the Model Usage page. Data is aggregated by workspace, not by Alibaba Cloud account. For more information, see How do I check the total token usage for my Alibaba Cloud account?. The data latency is approximately 1 hour.

1. On this page, select a model type tab, such as Large Language Model. Then, select a time range. The page summarizes the usage of all invoked models for that inference type within the selected time period.

   Statistical Time Range: You cannot view statistics that are older than 30 days. To query older usage information, go to the Expenses and Costs page.

   Large Language Model tab supports filtering by inference type, which can be real-time inference or batch inference. If no batch inference usage data exists for the workspace, the inference type drop-down list displays only real-time inference.

2. To query usage for a specific model, enter the model name (such as qwen-plus) in the search box on the right side of the page to filter the data.

   Find model names in the Model List.

View billing overview

On the Billing Overview page, view the cost summary for Model Studio services, including:

• Cost cards: View total expenses, subscription costs, and bill costs for the current billing period, with links to detailed breakdowns.
• Bill trends: View cost trends in chart or list format. Filter by monthly or daily statistics, product category, API Key ID, or model.
• Cost alerts: Set up cost alerts to receive notifications for abnormal expenses.

Usage units

In Model Studio, usage statistics for different models are:

Type	Subcategory	Unit	Billing (invocation)
Large language model	Text generation	Token	Billed by the number of tokens for input and output.
	Deep thinking
	Visual understanding
Visual model	Image generation	Images	Billed by the number of images successfully generated.
	Video generation	Seconds	Billed by the number of video seconds successfully generated.
Speech model	Speech synthesis	Second, character, or token	May be billed by audio duration (seconds), corresponding text characters, or tokens, depending on the model.
	Real-time speech synthesis
	Audio file recognition
	Real-time speech recognition
	Audio and video translation
Omni-modal model	Omni-modal	Token	Text is billed by token count. Other modalities (audio, image, video) are billed by their corresponding token counts.
	Real-time multimodal
Embedding model	Multimodal embedding	Token	Billed by the number of tokens in the input text.
	Text embedding

Going live

Recommendations for managing model usage:

• Control model output length: Limit the maximum length of content generated per invocation, and thus control costs, by reasonably limiting the thinking length and setting the max_tokens parameter.
• Select models based on task type: For simple jobs such as categorization and summary, prioritize lower-cost, lightweight models over powerful but more expensive models.
• Monitoring and alerting: Monitor usage trends. Set usage alerts to receive timely notifications for abnormal usage.
• Optimize prompts: Concise and clear prompts improve model output quality and reduce unnecessary input tokens.
• Use batch inference: For non-real-time, large-batch processing tasks, batch inference is often more cost-effective than real-time invocation.

Glossary

Noun	Explanation
Token	Large language models process input and output in tokens. A token can be: Single character: such as A, I Complete word: such as large, Model Part of a long word: A long word is often split into multiple tokens. This splitting process is called tokenization. Based on experience, on average, 1 Token corresponds to approximately 1.5-2 Chinese characters; 1 English letter corresponds to approximately 0.25 Tokens; 1 English word corresponds to approximately 1.3 Tokens: Alibaba Cloud Model Studio: approximately 4-5 Tokens Hello World: approximately 2 Tokens Each model has a maximum input and output token count (see Model list). Exceeding this limit causes failure.
Real-time	Refers to all direct and indirect invocations of a model, primarily covering the following scenarios: API invocation Model experience Model Studio applications (agent/workflow/agent orchestration applications, and nodes involving model invocation, such as LLM nodes, intent categorization nodes, and agent group nodes) in test and published states Assistant API (Deprecated) invocation Application Invocation Prompt feedback optimization Model evaluation
Batches	Large-scale data processing performed offline for scenarios that do not require real-time responses, using the OpenAI-compatible - Batch (file input) API.

FAQ

Q: How do I view the total token usage for my Alibaba Cloud account?

A: Use your Alibaba Cloud account to access the Bill Details page and export the bill.

On the Bill Details page, filter and export the bill:

1. Set the Billing Month to the target month (such as 2025-05).
2. Under Product Name, select Model Studio.
3. Click the download icon in the upper-right corner of the table to export billing data.