How DeepSeek V4 API billing works on cloudgpu.app

DeepSeek V4 Flash and Pro calls on cloudgpu.app are billed per token from one prepaid balance, with no subscription. The current rate for each model is on the API page; this post explains what you are billed for and how to work out the cost of a job, and does not repeat the rates, because they change.

Updated 2026-10-24.

What you are billed for

• Input tokens: the prompt you send, including system messages and earlier turns.
• Cached input tokens: when the upstream prompt cache fires, cached input is charged at the cached rate shown on the API page.
• Output tokens: the tokens the model returns.

Model IDs: deepseek-v4-flash and deepseek-v4-pro. deepseek-v4-flash is the April open-weights checkpoint, not DeepSeek’s newer V4.1 Flash. Off-peak DeepSeek pricing is passed through automatically.

Work out the cost of a job

Cost = input tokens ÷ 1,000,000 × input rate + cached input tokens ÷ 1,000,000 × cached rate + output tokens ÷ 1,000,000 × output rate.

Take the three rates from the API page on the day you run the job. The usage block in each API response gives the token counts.

How you pay

All calls are paid from one prepaid balance. Top up by card or with USDT on TRC-20; businesses can also pay by bank transfer. The balance does not expire, and the same balance pays for GPU time. Methods and minimums: /payments.

Limits

Public-beta defaults: 100 requests per minute per API key and 10,000 requests per day per account; both can be raised on request.

Set it up

Create a key in the API console and point any OpenAI-compatible client at the base URL https://cloudgpu.app/v1. Full setup with curl, Python, Cursor and Cline: DeepSeek V4 API With USDT or Card: Use the OpenAI SDK.

Frequently asked questions

How is the DeepSeek V4 API billed?

Per token, from one prepaid balance, with no subscription. Input, cached input and output tokens each have their own rate, listed for every model at https://cloudgpu.app/api.

What are the model IDs?

deepseek-v4-flash and deepseek-v4-pro. deepseek-v4-flash is the April open-weights checkpoint, not the newer V4.1 Flash.

How do I work out the cost of a job?

Multiply the input, cached input and output token counts from the usage block by the matching per-token rates on the API page, then add the three results.

How can I pay?

Top up one prepaid balance by card or with USDT on TRC-20. Businesses can also pay by bank transfer. Methods and minimums are listed at https://cloudgpu.app/payments.