Announced 13 August 2026
The DeepSeek API price increase
DeepSeek’s new rate card takes effect 16 August 2026, 16:00 UTC. Peak output on DeepSeek V4-Pro moves from $0.87 to $3.96 per million tokens, 4.6× the old rate.
Two things change at once. Prices go up, and time-of-day billing returns after eleven months of a single flat rate — so the hour you send a request starts mattering again. The discount is measured against the new peak rate, though, not the old card: off-peak output still costs 2.3× what it did before.
Change per meter
Input, cache hit
$0.003625 $0.044
12× at peak · 6.1× at off-peak
Input, cache miss
$0.435 $1.32
3.0× at peak · 1.5× at off-peak
Output
$0.87 $3.96
4.6× at peak · 2.3× at off-peak
DeepSeek V4-Pro, per million tokens. Multiples are calculated from the published rates on both cards; the old card had a single rate for all hours.
What happened, and when
New rates in —
-
26 February 2025 background
Off-peak discounts introduced
DeepSeek first applies time-of-day pricing, discounting V3 by 50% and R1 by 75% during its quieter hours.
-
5 September 2025 background
Off-peak discounts withdrawn
The time-of-day mechanism is retired in favour of a permanent price cut, leaving a single flat rate around the clock.
-
13 August 2026
New rates announced
DeepSeek publishes a substantially higher rate card and brings time-of-day pricing back — this time as a peak surcharge rather than an off-peak discount on the old level.
Official source ↗ -
16 August 2026, 16:00 UTC
New rates take effect
Requests are billed against the new card from this instant, and the peak/off-peak split starts applying on the UTC windows.
Official source ↗
Effective date as published: 16 August 2026, 16:00 UTC.
Full rate card, before and after
DeepSeek V4-Pro
Flagship model. Higher capability, lower concurrency allowance.
deepseek-v4-pro 1M context · 384K max output · 500 concurrency
| Per 1M tokens | Before the change | From 16 Aug 2026 | Increase | ||
|---|---|---|---|---|---|
| Flat, all hours | Peak | Off-peak | at peak | at off-peak | |
| Input — cache hit Prompt prefix served from DeepSeek’s cache | $0.003625 | $0.044 | $0.022 | 12× | 6.1× |
| Input — cache miss Prompt processed fresh, no cache reuse | $0.435 | $1.32 | $0.66 | 3.0× | 1.5× |
| Output Generated tokens, including reasoning tokens | $0.87 | $3.96 | $1.98 | 4.6× | 2.3× |
The old card had no time-of-day pricing — one rate applied around the clock. Both increase columns therefore compare against that single flat rate, and even the new off-peak rate sits above it.
DeepSeek V4-Flash
Cheaper, faster tier with a far higher concurrency allowance.
deepseek-v4-flash 1M context · 384K max output · 2,500 concurrency
| Per 1M tokens | Before the change | From 16 Aug 2026 | Increase | ||
|---|---|---|---|---|---|
| Flat, all hours | Peak | Off-peak | at peak | at off-peak | |
| Input — cache hit Prompt prefix served from DeepSeek’s cache | $0.0028 | $0.014 | $0.007 | 5.0× | 2.5× |
| Input — cache miss Prompt processed fresh, no cache reuse | $0.14 | $0.44 | $0.22 | 3.1× | 1.6× |
| Output Generated tokens, including reasoning tokens | $0.28 | $1.32 | $0.66 | 4.7× | 2.4× |
The old card had no time-of-day pricing — one rate applied around the clock. Both increase columns therefore compare against that single flat rate, and even the new off-peak rate sits above it.
What it changes in practice
Caching stops being optional
The gap between the cache-hit and cache-miss rates is what now decides most bills. Prompt layouts that keep a stable prefix are worth revisiting before the change lands.
Scheduling matters again
Under a flat card the hour a job ran was irrelevant to its cost. It is not any more, and a fixed-percentage discount on a higher base is a larger absolute saving than the equivalent discount was in early 2025.
The comparison reopens
Workloads that chose DeepSeek purely on price were decided against the old card. Whether the answer holds depends on your own input/output mix, not on headline rates.
Questions about the increase
Why did DeepSeek raise its API prices?
DeepSeek published the new rate card on 2026-08-13. Peak output on DeepSeek V4-Pro goes from $0.87 to $3.96 per million tokens — 4.6× the old rate.
The company has not published a cost breakdown, so any explanation beyond the announcement itself is speculation. What the numbers show is a move away from the aggressive undercutting that defined DeepSeek’s earlier pricing, toward rates closer to the rest of the market.
When exactly do the new DeepSeek prices take effect?
The new rates apply from 16 August 2026, 16:00 UTC. Requests billed before that instant use the old card; requests after it use the new one.
Because billing runs on UTC rather than your local clock, the switch lands mid-working-day in some regions. The countdown at the top of this site tracks it in your own timezone.
How much cheaper is the DeepSeek off-peak rate?
Off-peak rates are 50% lower than peak, applied to input and output alike. On DeepSeek V4-Pro that is $1.98 per million output tokens instead of $3.96.
It is worth being clear about what the discount is measured against: it is half of the new peak rate, not a return to the old one. Off-peak output still costs 2.3× what the same tokens cost under the previous flat card.
The rate is decided by when the request reaches DeepSeek, not when your job was queued locally. Batch work that can tolerate delay is the clearest way to capture it.
Is DeepSeek still cheaper than Claude, OpenAI and Gemini after the increase?
For most workloads, yes — but by a smaller margin than before, and the gap narrows further against the cheaper tiers of each provider once you account for cache hits.
The right comparison depends on your own mix of input, output and cache rate, which is what the calculator is for: it applies your volumes to every rate card at once instead of comparing headline numbers that assume a workload you may not have.
Does the price change affect the DeepSeek chat app or only the API?
The rate card covers API usage, billed per token against your API account balance.
Consumer access through the DeepSeek app and website is a separate product on separate terms, and is not what the prices on this site describe.
Do credits bought at the old price keep their value?
Balance is held in currency, not in tokens. Credit already on your account stays valid, but from the change-over each request draws down that balance at the new per-token rates, so a given balance buys proportionally fewer tokens than it did before.
Check the official pricing page for the terms that actually govern your account: https://api-docs.deepseek.com/quick_start/pricing/