E-MapReduce:Model calls (pay-as-you-go)

Feature

Description

Billing rules

Billing is based on the token usage of built-in model services in a workspace within a one-hour billing cycle. The billing rules vary for different built-in models:

• qwen-plus: Billed based on input tokens and output tokens.

• text-embedding-v4: Billed based on input tokens. Output tokens are not billed.

The cost for model calls is calculated as follows: Input token usage × Unit price for input + Output token usage × Unit price for output

For example, if you call `ai_query()` 10,000 times in the China (Hangzhou) region, with each call having 260 input tokens and 50 output tokens, the cost is calculated as follows: 0.00096 × 260 × 10000 ÷ 1000 + 0.0024 × 50 × 10000 ÷ 1000 = 3.696 CNY.

Billing cycle

Fees are calculated on the hour, every hour (UTC+8). A new billing cycle begins after each calculation. At the end of each billing cycle, the system generates a bill and deducts the fees from your account. There may be a delay between the actual resource consumption and the data that appears on your bill. For more information, see Bill usage instructions.