Zhiqian Large Model
ZQAgent comes with a built-in large model — just register and you can use it; no installation required, and no need to apply for a Key.
The models you use in the app are all billed against this account balance. This page covers just two things: how billing works and how to top up.
Want to call it programmatically?
Developers who need an API Key, endpoint URL, or SDK examples should see the API Integration docs. Regular users don't need this.
How billing works
You're charged by actual usage — pay for what you use, and nothing when you don't.
| Model | Input (cache miss) | Input (cache hit) | Output |
|---|---|---|---|
deepseek-flash | ¥3 | ¥1 | ¥9 |
CNY / 1M tokens · Prices are per million tokens. A token is the billing unit for LLMs; for Chinese text, roughly 1 character ≈ 1 token.
What that costs in practice：A typical exchange (say, tidying up 20 file names) uses a few thousand tokens and costs less than a cent. The 5 CNY starter credit covers several hundred such tasks.
Billing rules:
- Prepaid: top up before use. When the balance is insufficient, the task stops and prompts you to top up — no overdraft and no automatic charges.
- Usage-based billing: charged only for the tokens actually consumed; no monthly fee and no minimum spend.
- No charge for failures: failed requests and invalid parameters are not billed.
- Unit prices always available: the amount and details of every charge are in the user center, so you can check them line by line.
How to top up
There is only one place to top up, and the funds you add are shared between large model calls and remote device management:
- Open the ZQAgent User Center
- Go to the Zhiqian Large Model API page
- Click Top Up and choose an amount
- Pay by scanning the code with WeChat Pay or Alipay — the funds take effect immediately
New users get ¥5 in credit on registration, which you can use right away to try it out.
Balance and reminders
- Your balance, usage details, and top-up history are all available in the User Center → Settings → Usage, and can be viewed by day or by model.
- You'll receive a reminder when your balance covers fewer than 15 / 7 / 3 days, so you can top up in advance.
- When the balance runs out: large model calls will report insufficient balance; remote device management will suspend the excess portion based on connection time and resume automatically after you top up (see Pricing for details).
- Need an invoice or corporate bank transfer? Register and contact us in the User Center.
FAQ
Do I need to install or configure anything first?
No. It's activated automatically after registration — just open the desktop app and sign in. No parameters to fill in.
How is this different from an API I buy on another platform?
You don't need to apply for a Key yourself or top up separately with each provider. Here you have one account and one balance, billed by usage, with all usage details in one place.
Can I use my own model Key instead?
Yes. Both the desktop and CLI versions support entering your own model Key; charges go directly to the model provider and don't consume your platform balance.
Why are some tasks more expensive than others?
Billing is by token: longer input and longer answers cost more. Cached portions have a lower unit price. A typical everyday task usually costs less than ¥0.01.
Can I get a refund after topping up?
Top-up amounts are used to offset actual usage and are prepaid funds. If you have special circumstances, please contact us in the User Center.