FAQ
How is usage billed?
Cost per request = input tokens × input price + output tokens × output price. Token counts come from the usage returned by the model service, and every request is itemized under Request logs in the dashboard. Researcher has thinking mode on, and its reasoning also counts as output tokens.
Are failed requests charged?
No. Requests that return an error (4xx/5xx) are free. If a streaming request disconnects midway, you pay for what was already generated.
What happens when my balance runs out?
The API returns 402 insufficient_balance. In the dashboard you can set a low-balance alert and a monthly spend cap to avoid surprises.
How are CNY prices converted?
Balances are kept in US dollars. CNY prices are converted at the rate shown on the page and are for reference only.
How can I add credit?
WeChat Pay (in CNY, by scanning a QR code) and redeem codes. WeChat payments are checked by hand and usually credited within minutes.