How billing works
Text models are usually billed by input and output tokens. Image models are usually billed per generation. Some models may have cache-read or cache-write pricing when upstream cache is supported.
| Item | Explanation |
|---|---|
| Balance | You add balance to your account and API calls deduct from that balance. |
| Input tokens | Tokens sent to the model, including your prompt and context. |
| Output tokens | Tokens generated by the model. |
| Usage logs | You can review usage and billing details in the console. |
| Currency | The console may primarily display CNY. USD values, when shown, are for easier international reading and may be approximate. |
Start with a small recharge for testing. After your client can call the API successfully, recharge based on your actual usage.
JBBToken Pricing