DeepSeek models: off-peak rates halve an already low price

Prices from official vendor price lists · October 2026
DeepSeek has 9 models in this rating, led by DeepSeek V4.1 Flash and DeepSeek V4 Pro. 7 of them have a public per-token price, from $0.177 to $3.15 per 1M output tokens, and context windows reach 1.31M. 8 have a free tier; all ship open weights.
DeepSeek
Services found: 9
#AI modelCapabilitiesAPI $/1M (in / out)Context windowPrice / moOutput $/1MOverviewVisit site
1

DeepSeek V4.1 Flash

DeepSeek
Text & chatVisionReasoning
$0.3 / $1.21M tokensAPI only1.2View
2

DeepSeek V4 Pro 0813

DeepSeek
Text & chatReasoningAdjustable thinking
$1.0494 / $3.14821.05M tokensAPI only3.148View
3

DeepSeek V4 Flash 0731

DeepSeek
Text & chatVisionReasoning
$0.065 / $0.181.31M tokensAPI only0.18View
4

DeepSeek V4 Pro 0423

DeepSeek
Text & chatReasoningAdjustable thinking
$0.95526 / $1.910521.05M tokensAPI only1.911View
5

DeepSeek V4 Flash 0423

DeepSeek
Text & chatVisionReasoning
$0.088606 / $0.1772121.05M tokensAPI only0.177View
6

DeepSeek V3.1 Terminus

DeepSeek
Text & chatStructured JSON outputPrompt caching
$0.27 / $1163.8K tokensAPI only1View
7

DeepSeek V4 Pro

DeepSeek
Text & chatReasoningAdjustable thinking
1M tokensAPI onlyView
8

DeepSeek V4 Flash

DeepSeek
Text & chatVisionReasoning
1M tokensAPI onlyView
9

DeepSeek V3.2

DeepSeek
Text & chatStructured JSON outputPrompt caching
$0.269 / $0.4163.8K tokensAPI only0.4View

Browse by need

Reasoning models169 modelsCoding assistants17 modelsVision and images156 modelsLong context 1M+82 modelsFree tier available80 modelsOpen weights88 models

DeepSeek delivers the most aggressive price-to-capability ratio here. V4 Flash runs at roughly $0.09 input and $0.18 output per 1M tokens with a 1.05M window; V4 Pro sits above it for harder reasoning. Both support thinking and non-thinking modes and a 384K maximum output.

Two quirks are worth knowing. DeepSeek publishes its price list in yuan, so dollar figures here come from Alibaba Model Studio resale. And billing splits into peak and off-peak windows on Beijing time, with off-peak rates at half — which turns scheduling into a real cost lever.

Cheapest at scaleSub-cent per thousand tokens on the Flash line.
Off-peak halvingBeijing-hours scheduling cuts the bill in half.
Huge output ceiling384K maximum output, far above most competitors.

Models from this vendor

9 DeepSeek models in this rating. Prices are input and output per 1M tokens from the vendor official price list; for dirhams multiply by 3.6725. Tap a card to open the model page.

DeepSeek V4.1 Flash
#1
DeepSeek
in $0.3out $1.21M tokensText & chatVision
DeepSeek V4 Pro 0813
#2
DeepSeek
in $1.0494out $3.14821.05M tokensText & chatReasoning
DeepSeek V4 Flash 0731
#3
DeepSeek
in $0.065out $0.181.31M tokensText & chatVision
DeepSeek V4 Pro 0423
#4
DeepSeek
in $0.95526out $1.910521.05M tokensText & chatReasoning
DeepSeek V4 Flash 0423
#5
DeepSeek
in $0.088606out $0.1772121.05M tokensText & chatVision
DeepSeek V3.1 Terminus
#6
DeepSeek
in $0.27out $1163.8K tokensText & chatStructured JSON output
DeepSeek V4 Pro
#7
DeepSeek
price on request1M tokensText & chatReasoning
DeepSeek V4 Flash
#8
DeepSeek
price on request1M tokensText & chatVision

How to choose

  • Schedule batch work into off-peak hours — the rate halves.
  • Use Flash for volume and reserve Pro for reasoning that Flash visibly fails.
  • Convert prices yourself: DeepSeek publishes in yuan, so dollar figures are resale estimates.
  • Check the 384K output ceiling if you generate long documents in one pass.

Common mistakes

Assuming the dollar price is officialIt is not — DeepSeek publishes in yuan only.
Running peak-hour batchesIdentical work costs twice as much between 09:00 and 18:00 Beijing time.

Who this suits

Cost-driven productsThe lowest credible per-token rates in this rating.
Long-output work384K maximum output, far above most competitors.
ExperimentationCheap enough that testing many prompts costs almost nothing.

Related filters

Text & chat228 models for text generation and chat
Reasoning169 models that think before answering
Adjustable thinking79 models with a reasoning depth dial
Tool calling166 models that call your functions
Alibaba (Qwen)34 models in this rating
OpenAI30 models in this rating
Google DeepMind17 models in this rating

Frequently asked questions

Why are DeepSeek prices shown in dollars here?

DeepSeek publishes in yuan only; the USD figures come from Alibaba Model Studio, which resells the same models.

What is the difference between Flash and Pro?

Flash is the volume model; Pro costs roughly six times more and handles harder reasoning.

Is DeepSeek open weight?

Earlier generations were published. The V4 documentation makes no licence statement.

What are peak and off-peak hours?

Peak is 09:00–12:00 and 14:00–18:00 Beijing time on weekdays; off-peak rates are half.

Is DeepSeek V4 open weight?

The V4 documentation makes no licence statement, unlike earlier generations.

Frequently searched

By capabilityReasoningCodingVisionLong context 1M+Free tierOpen weightsLow latencyRealtime voiceImage generation
By vendorAnthropic (13)OpenAI (30)Google DeepMind (17)xAI (8)Meta (6)DeepSeek (9)Mistral AI (11)Alibaba (Qwen) (34)Z.ai (12)Moonshot AI (5)
Popular modelsClaude Fable 5.1GPT-6 AstraClaude Opus 5Claude Fable 5GPT-5.6 SolGLM 5.3Grok 4.6Kimi K3GPT-5.6 Terra
By deliveryBatch APIPrompt cachingPriority tierFrontier tierProprietary APITool callingStructured JSONWeb searchMultilingual

Ratings 0

Avg. score 0

Rate this page

Help us improve