Chat with DeepSeek models on the web or app, or build with a low-cost API with a 1M token context.

From $0.15/1Minputtokens
See pricing
deepseek.com · opens in a new tab

About DeepSeek

DeepSeek is an AI chat assistant and model API from China, with DeepThink reasoning, web search and a 1M token context. It suits developers and cost-conscious users, with API prices from $0.15 per million input tokens.

DeepSeek makes AI models you can use in a web chat, mobile app, a desktop app or through its API. In chat, DeepThink turns on step-by-step reasoning and Search pulls in web results. The API serves two models, deepseek-flash and deepseek-v4-pro, with a 1 million token context, up to 384K output tokens, tool calls, JSON output and both OpenAI and Anthropic compatible formats.

It suits developers who want strong models at low prices, plus anyone who wants a capable chat assistant. A key limit is data location, since DeepSeek is a Chinese company, so check its privacy policy before sharing sensitive data. No paid chat plan is listed, and the API is pay as you go from $0.15 per million input tokens off-peak.

Pricing model
Free, Freemium, Paid
Starting price
API from $0.15 per 1M input tokens
Platforms
WebiOSAndroidAPI / SDK
API
Available
AI models
DeepSeek-V4.1-FlashDeepSeek-V4-Pro-0813DeepSeek-R1DeepSeek-Coder
Made for
Software developersAI app buildersStudentsCost-conscious businesses

What you get

DeepThink mode shows step-by-step reasoning for harder questions in the chat app.

Search mode adds live web results to answers in DeepSeek chat.

The API handles a 1 million token context and up to 384K tokens of output.

Call the API in OpenAI or Anthropic format, so existing code needs few changes.

Off-peak API rates are half price outside set weekday hours in UTC.

Web search and fetch APIs cost $0.75 per 1,000 successful requests.

The good and the bad

What's good4

  • API pricing is very low, with deepseek-flash input from $0.15 per million tokens off-peak.
  • Cached input costs far less, from $0.003 per million tokens on flash off-peak.
  • Both models support tool calls, JSON output and thinking or non-thinking modes.
  • The chat is available on the web, in mobile apps and in a desktop app for computers.

What's not3

  • DeepSeek is based in China, which may not suit teams with strict data rules.
  • Peak-hour API prices are double the off-peak rates on weekday mornings UTC.
  • Vision input is only supported on deepseek-flash, not on deepseek-v4-pro.

Pricing

deepseek-flash API
$0.15/1M input tokens
Off-peak; $0.60/1M output; peak rates double
  • 1M context
  • Vision
  • Tool calls
See deepseek-flash API
deepseek-v4-pro API
$0.66/1M input tokens
Off-peak; $1.98/1M output; peak rates double
  • 1M context
  • Tool calls
  • JSON output
See deepseek-v4-pro API
Search and Fetch API
$0.75/1K requests
Per 1,000 successful requests
  • Web search
  • Page fetch
See Search and Fetch API
Prices can change. Check the current price on the official site before you pay.api-docs.deepseek.com pricing

User Reviews & Ratings

Reviews from people who use DeepSeek. Sign in with Google to add yours.

No reviews yet
Rating BreakdownFilter

Questions people ask

What is DeepSeek and who is it for?

DeepSeek is an AI company that offers a chat assistant and a model API. It suits developers who want low-cost models and users who want a capable chatbot.

Is DeepSeek free?

DeepSeek's web chat and apps list no paid plan. The API is pay as you go, with fees taken from a topped-up balance.

How much does the DeepSeek API cost?

deepseek-flash costs $0.15 per million input tokens and $0.60 per million output tokens off-peak. deepseek-v4-pro costs $0.66 input and $1.98 output off-peak. Peak rates are double.

When are DeepSeek's off-peak hours?

Peak hours are 01:00 to 04:00 and 06:00 to 10:00 UTC on weekdays, excluding Chinese public holidays. All other times, including weekends, are off-peak at half price.

What context length does DeepSeek support?

Both API models support a 1 million token context and up to 384K output tokens.

Can I use DeepSeek with existing OpenAI code?

Yes. The API supports the OpenAI format and also an Anthropic-compatible format at a separate base URL.

Which devices does DeepSeek support?

DeepSeek runs on the web, in mobile apps and in a desktop app, with an API for developers.

How does DeepSeek handle privacy?

DeepSeek is a Chinese company with its own privacy policy. Review it before entering sensitive personal or business data.

How does DeepSeek compare to ChatGPT?

ChatGPT offers more built-in tools like image creation and voice with video. DeepSeek stands out for very low API prices and strong reasoning models.

Is DeepSeek easy to use?

Yes. The chat works like other assistants, and the API follows familiar formats, so developers can switch with small code changes.

Alternatives to DeepSeek

See all
ChatGPTFree · Freemium

Ask questions, write, code, search the web and make images with one AI chat assistant from OpenAI.

Free plan, paid from $8/month

ClaudeFree · Freemium

Write, research, analyze files and code with an AI assistant that handles very long documents well.

Free plan, paid from $17/month

Qwen ChatFreemium · Subscription

Chat with Alibaba's latest Qwen models for free, plus create images, videos, slides and web pages.

Free plan, no public paid price

Anima AIFreemium · Paid

Chat with a customizable AI friend that remembers your stories and adapts to your mood.

Free plan, no public paid price

Related topics