← AI
AI-authored analysis · Fictional correspondent, real sources
DeepSeek’s direct API price depends on when you use it
The Executive Brief, once a month, free. What it is

A developer pricing an application needs to know which service will send the bill. DeepSeek’s direct API currently lists V4.1 Flash output at US$0.60 per million tokens off-peak and US$1.20 at peak times, checked on 24 October. Those are provider prices, with a time-of-use condition attached. [1]

Peak hours are 01:00-04:00 and 06:00-10:00 UTC on weekdays, excluding Chinese public holidays. For a Hong Kong team, that is 09:00-12:00 and 14:00-18:00. Other hours, weekends and Chinese public holidays use the off-peak rate. [1]

Input is billed separately. Uncached input costs US$0.15 per million tokens off-peak and US$0.30 at peak times. Cache-hit input costs US$0.003 off-peak and US$0.006 at peak times per million tokens. Comparing output prices alone misses part of the application’s bill. [1]

A workload that can wait for the evening may benefit from the cheaper window. An interactive service has less freedom to move its demand. Before switching a working application, measure its input and output use, confirm the route through which it buys access, and test the replacement against the work it actually does.

Sources

Sources reviewed 24/10/2026, 20:57:22 HKT. Editorial source review.

The Wang Report uses OpenAI models through Codex under human editorial oversight. See our Editorial Standards.