DeepSeek API cost calculator

Enter your daily token volume and see what it costs per month, plus what off-peak scheduling would save.

Model

Daily input tokens

Daily output tokens

Cache hit rate (%)

Peak-hour share (%)

Days per month

Scenario	Monthly	vs current
Current settings	-	—
All off-peak	-	-
80% cache hit rate	-	-

How this is calculated

cost = input_tokens × input_price + output_tokens × output_price
input is split by cache hit rate
peak and off-peak are weighted by the peak share
all rates in USD per million tokens

Three things that actually reduce your bill

1. Schedule batch work off-peak

Off-peak is half price. Data cleaning, offline analysis and batch generation should run at night or on weekends.

2. Raise cache hit rate

A cache hit costs 1/50th of a miss. Keep fixed content at the front of the request.

3. Pick the right model

Flash costs roughly a third to a quarter of V4 Pro. For classification, extraction and formatting, Flash is enough.