DeepSeek

Model

API for DeepSeek's V4 reasoning and chat models

Price: Pay per use
Access: API key

Docs

About

DeepSeek's API serves its V4 models with a 1M-token context through OpenAI- and Anthropic-compatible endpoints. It is for developers who want low per-token prices, and off-peak hours cost half. The model weights are published under the MIT license.

What you can do with it

Chat, summarize and reason over long documents with a 1M-token context
Generate and complete code, including fill-in-the-middle
Point OpenAI or Anthropic SDK code at DeepSeek by changing the base URL

Get started

Create an API key on the DeepSeek platform
Top up your account balance
Set DEEPSEEK_API_KEY in your environment

Example

curl https://api.deepseek.com/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer ${DEEPSEEK_API_KEY}" \
  -d '{
    "model": "deepseek-flash",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ],
    "thinking": {"type": "enabled"},
    "reasoning_effort": "high"
  }'

Details

Hosting: Hosted service, Self-hosted
Available in: Worldwide
MCP server: None