← Back to AI News

DeepSeek V4 Pricing Is Now Peak/Off-Peak — and Peak Can Be 10x the Old Flat Rate
Chris Harper
1 min read
Aug 18, 2026 · 04:09 UTC
AI
News
LLM
Developer Tools
DeepSeek's new peak/off-peak pricing, live since August 16, replaces the flat V4 rates — peak can be up to 10x the old promotional price; off-peak is half of peak.
The August 14 post here flagged that DeepSeek had announced a price increase with no details. The numbers are now live:
| Model | Old flat | Off-peak | Peak |
|---|---|---|---|
| deepseek-v4-flash | $0.14 / $0.28 | $0.22 / $0.66 | $0.44 / $1.32 |
| deepseek-v4-pro | $0.435 / $0.87 | $0.66 / $1.98 | $1.32 / $3.96 |
Per 1M input / output tokens.
Peak hours run 01:00–04:00 UTC and 06:00–10:00 UTC daily. Everything else is off-peak.
Why it matters: If your pipeline is schedulable or batched, routing calls outside peak hours halves the cost. Budget projections built before August 16 need revision — V4 Flash at peak is now 3x the old rate, V4 Pro is 3x–4x.
Sources: DeepSeek API Pricing · qz.com