Console

Pricing

Pay only for the tokens you use, deducted from your balance. No complicated plan tiers, just transparent usage-based pricing.

Beijing time 00:16 · Off-peak (base rate) · Peak starts Today at 09:00

Model pricing

deepseek-flash
CNY / 1M tokens

Input (cache miss)

Standard context input tokens

¥1.00/1MPeak ¥2.00

Input (cache hit)

Repeated prefixes hit the cache automatically, at 1/50 of the base input rate.

¥0.02/1MPeak ¥0.04

Model output

Generated response and reasoning tokens

¥4.00/1MPeak ¥8.00

Peak and off-peak

Beijing time / UTC+8

09:00 - 12:0014:00 - 18:00

The periods above on Weekdays (Mon–Fri) are peak periods; all three token rates are ×2 the base price. Other hours and all of Weekends use the base rate.

All periods use server time in Beijing (UTC+8).

Billing notes

Settlement rules

  1. 01

    Actual token usage is deducted from your balance. No usage, no charge.

  2. 02

    Account credits do not expire or reset at an expiry date.

  3. 03

    Context caching reduces input costs; hits cost 1/50 of the off-peak input rate.

  4. 04

    Charges follow server-side measurements. Check itemized usage in the console.