AI Token Price
DeepSeek

DeepSeek · 2 current models

DeepSeek API pricing

Estimate your costCompare with…

DeepSeek's API models. V4.1 Flash costs $0.30 / $1.20 per 1M tokens at peak times and half that off-peak; V4 Pro costs $1.32 / $3.96 at peak.

DeepSeek V4 Pro

Repriced 16 Aug
Input
$1.32
per 1M tokens
Output
$3.96
per 1M tokens
Cached input
$0.044
97% off input
Off-peak
50% off
outside peak hours
Context window
1M
tokens
Verified 27 Sep 2026DeepSeek pricing Repriced on 16 Aug 2026: input down 24%, output up 14% (was $1.74 / $3.48)

What it costs in practice

DeepSeek V4 Pro at list prices, no batch discount.

1,000 chat replies1,500 tokens in and 300 out each$3.17
Summarising 100 long reports20,000 tokens in and 800 out each$2.96
A month of a coding agent400 tasks, 85% of input cached$208

The coding-agent example assumes 85% of input is served from cache, typical for agents that resend the same context each step. Model your own workload

Details

API model IDdeepseek-v4-pro
Released24 Apr 2026
Off-peak price$0.66 / $1.98 per 1M tokens
WeightsOpen, 1600B total and 49B active parameters
StatusCurrent
  • Peak price shown. Off-peak requests cost $0.66 / $1.98.
  • Updated model (0813) released 13 August; peak and off-peak pricing since 16 August.

Checked 27 Sep 2026 against DeepSeek pricing and DeepSeek updates .

Other DeepSeek models

DeepSeek retired V4 Flash on 10 September 2026 and the deepseek-chat and deepseek-reasoner endpoints (V3.2 and R1) on 24 July.

Model Input / 1M Output / 1M Cached input Context Released
DeepSeek V4.1 Flash
 $0.30 $1.20 $0.006 1M 10 Sep 2026

DeepSeek V4.1 Flash details

DeepSeek V4.1 Flash

New 10 Sep

$0.30 / $1.20 per 1M tokenscached $0.006 · off-peak 50% off · 1M context

Released 10 Sep 2026 · deepseek-flash

  • Peak price shown. Off-peak requests cost $0.15 / $0.60.
  • Replaced V4 Flash on 10 September; the old model names now route here.

Checked 27 Sep 2026 against DeepSeek pricing and DeepSeek announcement .

Compare DeepSeek V4 Pro

Compare with any model

Price history

All changes
  • DeepSeek
    DeepSeek V4.1 Flash replaced V4 Flash at $0.30 / $1.20 peak ($0.15 / $0.60 off-peak). Source
  • DeepSeek
    DeepSeek V4 Pro moved to peak and off-peak pricing: $1.32 / $3.96 at peak, down from $1.74 on input and up from $3.48 on output. Source
  • DeepSeek
    DeepSeek discontinued the deepseek-chat (V3.2) and deepseek-reasoner (R1) endpoints. Source
  • DeepSeek
    DeepSeek launched V4 Pro ($1.74 / $3.48) and V4 Flash ($0.14 / $0.28). Source

DeepSeek pricing rules

  • Off-peak requests cost half the peak price. DeepSeek lists peak hours as 01:00–04:00 and 06:00–10:00 UTC on weekdays.
  • There is no batch discount.
  • Prices shown are peak rates.

All DeepSeek models

Retired DeepSeek models

  • Retired 10 Sep 2026DeepSeek V4 Flash was $0.14 / $0.28 per 1M tokens. Use DeepSeek V4.1 Flash instead.
  • Retired 24 Jul 2026DeepSeek V3.2 was $0.28 / $0.42 per 1M tokens. The deepseek-chat endpoint was discontinued on 24 July 2026. Use DeepSeek V4.1 Flash instead.
  • Retired 24 Jul 2026DeepSeek R1 was $0.55 / $2.19 per 1M tokens. The deepseek-reasoner endpoint was discontinued on 24 July 2026. Use DeepSeek V4 Pro instead.

Questions

How much does DeepSeek V4 Pro cost?

As of 27 Sep 2026, DeepSeek V4 Pro costs $1.32 per million input tokens and $3.96 per million output tokens, and $0.044 per million cached input tokens. Prices checked 27 Sep 2026.

What is the cheapest DeepSeek model?

DeepSeek V4.1 Flash, at $0.30 / $1.20 per million input / output tokens.

What is DeepSeek V4 Pro's context window?

1M tokens.

Does DeepSeek V4 Pro support prompt caching?

Yes. Repeated input read from the cache costs $0.044 per million tokens, 97% less than normal input.