DeepSeek moved V4 Flash and V4 Pro to peak and off-peak API pricing
Changed: Aug 16, 2026
Published
From 16:00 UTC on August 16, 2026, DeepSeek charges peak and off-peak rates. At peak, Flash cache-miss input rose from $0.14 to $0.44 and output from $0.28 to $1.32 per 1M tokens. Off-peak is half the peak rate.
| What changed | Before | Now |
|---|---|---|
| V4 Flash input (cache miss), peak, per 1M tokens | $0.14 (flat) | $0.44 |
| V4 Flash output, peak, per 1M tokens | $0.28 (flat) | $1.32 |
| V4 Flash input (cache miss), off-peak, per 1M tokens | $0.14 (flat) | $0.22 |
| V4 Flash output, off-peak, per 1M tokens | $0.28 (flat) | $0.66 |
| V4 Pro input (cache miss), peak, per 1M tokens | $0.435 (flat) | $1.32 |
| V4 Pro output, peak, per 1M tokens | $0.87 (flat) | $3.96 |
| V4 Pro input (cache miss), off-peak, per 1M tokens | $0.435 (flat) | $0.66 |
| V4 Pro output, off-peak, per 1M tokens | $0.87 (flat) | $1.98 |
What changed
On July 25, 2026, DeepSeek’s pricing page showed one flat rate per model. Flash cost $0.0028 per 1M input tokens on a cache hit, $0.14 on a miss, and $0.28 for output. Pro cost $0.003625, $0.435 and $0.87.
The copy archived on August 16 shows two rates. DeepSeek’s announcement says the new pricing takes effect at 16:00 UTC on August 16, 2026, and that off-peak rates are 50% lower than peak. Peak Flash is $0.44 for cache-miss input and $1.32 for output. Peak Pro is $1.32 and $3.96. Off-peak, Flash is $0.22 and $0.66, and Pro is $0.66 and $1.98.
Even off-peak, the new rates sit above the old flat ones. Peak input costs about three times as much (Flash $0.44 against $0.14, Pro $1.32 against $0.435), and peak output about 4.5 to 4.7 times as much.
One caveat: the July page doesn’t say whether the flat rates were a launch discount. It shows them without a label.
The live page now defines peak as 01:00 to 04:00 and 06:00 to 10:00 UTC, Monday through Friday, excluding Chinese public holidays. Everything else is off-peak, weekends included.
Who feels it
Teams that run API traffic during European and Asian working hours pay the peak rate. Chat products and assistants with live users feel it most.
Batch jobs can move to the off-peak window. They still pay more than before, but about half the peak price.
What to do
- Re-run your cost estimate with both rates. A flat $0.14 / $0.28 Flash figure no longer applies.
- Schedule batch work for evenings UTC and weekends.
- Articles quoting $0.14 input and $0.28 output, or $0.435 and $0.87 for Pro, describe the pre-August 16 rates.
- Flash was repriced again on September 10, 2026. Check the live page before you budget.
Compare current API prices: check current API token costs for GPT, Claude, Gemini, Grok, DeepSeek and Mistral.
Sources
- DeepSeek API pricing page, archived July 25, 2026, flat rates (Wayback Machine)
- DeepSeek API pricing page, archived August 16, 2026, peak and off-peak table (Wayback Machine)
- DeepSeek announcement: API pricing update, effective 16:00 UTC Aug 16, 2026
- DeepSeek API pricing page, live
How we check prices: our method. Spotted a price we got wrong? Tell us at the contact page and we correct it in a dated note like this one.
Pricing notes
Get the monthly roundup of price changes
One short email a month with every price change we confirmed. No promos in between.