Skip to content
AI Playbook

DeepSeek API & SDKs

📖 3 min read deepseekapisdkdevelopment
Complete guide to the DeepSeek API — dual format (OpenAI + Anthropic compatible), thinking mode, tool calls, context caching, streaming, rate limits, and code examples in Python and cURL.
Key Takeaways
  • Dual API format: use either OpenAI SDK (api.deepseek.com) or Anthropic SDK (api.deepseek.com/anthropic) — no code changes needed
  • Thinking Mode: configurable reasoning with 'enabled'/'disabled' toggle and reasoning_effort (low/medium/high)
  • Context Caching: automatic KV cache with cache hits at 1/10 of input price
  • Tool Calls, JSON Output, FIM Completion, Chat Prefix Completion all supported

The DeepSeek API is uniquely versatile — it speaks both OpenAI and Anthropic API formats. You can switch to DeepSeek with zero code changes: just change the base_url and api_key.

Getting Started

Terminal window
# Get your API key from https://platform.deepseek.com/api_keys
export DEEPSEEK_API_KEY="your-api-key"

Option 1: OpenAI-Compatible Format

pip install openai


from openai import OpenAI


client = OpenAI(
    api_key="your-deepseek-key",
    base_url="https://api.deepseek.com"
)


response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[{"role": "user", "content": "Hello!"}],
    thinking={"type": "enabled"},
    reasoning_effort="high"
)
Terminal window
# cURL — OpenAI format
curl https://api.deepseek.com/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $DEEPSEEK_API_KEY" \
  -d '{
    "model": "deepseek-v4-pro",
    "messages": [{"role": "user", "content": "Hello!"}],
    "thinking": {"type": "enabled"},
    "reasoning_effort": "high"
  }'

Option 2: Anthropic-Compatible Format

pip install anthropic


from anthropic import Anthropic


client = Anthropic(
    api_key="your-deepseek-key",
    base_url="https://api.deepseek.com/anthropic"
)


message = client.messages.create(
    model="deepseek-v4-pro",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}]
)
Terminal window
# cURL — Anthropic format
curl https://api.deepseek.com/anthropic/v1/messages \
  -H "x-api-key: $DEEPSEEK_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-pro",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Thinking Mode

# Enable thinking mode (V4 Pro — default on)
response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[{"role": "user", "content": "Design a rate limiter"}],
    thinking={"type": "enabled"},
    reasoning_effort="high"
)


# Disable thinking (V4 Flash only)
response = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "What is 2+2?"}],
    thinking={"type": "disabled"}
)

Tool Calls (Function Calling)

response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[{"role": "user", "content": "What's the weather in Tokyo?"}],
    tools=[{
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get current weather for a city",
            "parameters": {
                "type": "object",
                "properties": {
                    "city": {"type": "string"}
                },
                "required": ["city"]
            }
        }
    }]
)

Streaming

stream = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Write a haiku about programming"}],
    stream=True
)
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

JSON Output

response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[{"role": "user", "content": "List 3 programming languages with their creator"}],
    response_format={"type": "json_object"}
)

Context Caching (KV Cache)

DeepSeek’s KV cache is automatic — no special parameters needed. Repeated prompt prefixes are cached server-side:

# First call: full price (cache miss)
# Subsequent calls with same prefix: 1/10 input price (cache hit)
for query in queries:
    response = client.chat.completions.create(
        model="deepseek-v4-flash",
        messages=[
            {"role": "system", "content": system_prompt},  # Cached after first call
            {"role": "user", "content": query}
        ]
    )
OperationV4 FlashV4 Pro (promo)
Standard Input$0.14 / 1M$0.435 / 1M
Cache Hit$0.0028 / 1M$0.0036 / 1M

Rate Limits

ModelConcurrency Limit
V4 Flash2,500
V4 Pro500

Rate limits are shared across API keys per workspace. For higher limits, contact DeepSeek.

Where Next