Abstract:
On September 9, according to the DeepSeek open platform announcement, DeepSeek will
reduce the price of Flash series model calls from 12:00 on September 10, Beijing time
. Among them, the highest decreases in cache hits, cache misses and output prices are60%, 33.33% and 11.11% respectively
.
After adjustment, the input cache hit price during the idle period is calculated by per million Tokens The price dropped from 0.05 yuan to
0.02 yuan
, the input cache miss price dropped from 1.5 yuan to1 yuan
, and the output price dropped from 4.5 yuan to4 yuan
.Prices during peak hours are still double those during idle hours. Among them, the input cache hit price dropped from 0.1 yuan per million Tokens to
0.04 yuan
, the input cache miss price dropped from 3 yuan to2 yuan
, and the output price dropped from 9 yuan to8 yuan
.Public information shows that the DeepSeek Flash series currently includes two models:
deepseek-v4-flash
anddeepseek-v4-flash-vision-exp
. The two models are subject to the same price adjustment standards this time.
Comments