How much do DeepSeek models cost?

Current per-million-token input and output API prices for each DeepSeek model are listed here and linked to DeepSeek's official pricing page, which is the source of truth as prices change.

Latest › Models › DeepSeek

Vikshy Models · DeepSeek

DeepSeek models: pricing, context & what each is for

Q: What is DeepSeek's most capable model?

DeepSeek's highest-tier catalogued model, with its context window, API pricing, and intended use, is listed on this page, alongside every other DeepSeek model.

DeepSeek has 6 models in the Vikshy catalogue, spanning DeepSeek V4, DeepSeek, DeepSeek R1, DeepSeek V3, compared side by side on context window, API pricing, modality, and what each is best for, every model linked to DeepSeek's official docs and pricing. Prices are per million tokens and can change, so the linked pricing page is the source of truth. For how DeepSeek ranks against other labs, see its Vikshy Score. Updated 2026-07-21.

ModelContextInput /MOutput /MModalityAccess

DeepSeek-V4-ProFlagship

Hard reasoning, long documents and agentic tasks.

$1.74

$3.48

Text

API + open weights

Docs Price

DeepSeek-V4-FlashBalanced

Everyday chat, coding and high volume production use.

$0.14

$0.28

Text

API + open weights

Docs Price

deepseek-chatLegacy

Existing code that still calls the old chat name.

$0.14

$0.28

Text

API

Docs Price

deepseek-reasonerLegacy

Existing code that still calls the old reasoner name.

$0.14

$0.28

Text

API

Docs Price

DeepSeek-R1Legacy

Self hosted reasoning and research on open weights.

128K

Open

Text

Open weights

Docs Price

DeepSeek-V3.1Legacy

Self hosting the earlier V3 line.

128K

Open

Text

Open weights

Docs Price

DeepSeek's Vikshy Score →All models, compared →What is the Vikshy Score? →

DeepSeek models, common questions

What is DeepSeek's most capable model?

Its highest-tier model in the catalogue is DeepSeek-V4-Pro (Flagship tier), built for hard reasoning, long documents and agentic tasks.

Which DeepSeek model is cheapest?

By API input price, DeepSeek-V4-Flash is the lowest at $0.14 per million input tokens ($0.28 output).

What is the largest context window DeepSeek offers?

DeepSeek-V4-Pro has the largest listed context window, up to 1M tokens.

Does DeepSeek offer open-weight models?

Yes. 4 of DeepSeek's catalogued models are available as open weights.