DeepSeek OpenAI-Compatible API: SDK Integration & Migration

from openai import OpenAI

client = OpenAI(
    api_key="<DeepSeek API Key>",
    base_url="https://api.deepseek.com"
)

response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=[{"role": "user", "content": "Hello"}],
    reasoning_effort="high",
    extra_body={"thinking": {"type": "enabled"}},
    stream=False
)

import OpenAI from "openai";

const client = new OpenAI({
    baseURL: "https://api.deepseek.com",
    apiKey: "<DeepSeek API Key>",
});

const completion = await client.chat.completions.create({
    model: "deepseek-v4-pro",
    messages: [{ role: "user", content: "Hello" }],
});

import anthropic

client = anthropic.Anthropic(
    base_url="https://api.deepseek.com/anthropic",
    api_key="<DeepSeek API Key>"
)

response = client.messages.create(
    model="deepseek-v4-pro",
    max_tokens=4096,
    system="You are a helpful assistant.",
    messages=[{"role": "user", "content": "Hello"}]
)

# Before
client = OpenAI(api_key=openai_key)

# After
client = OpenAI(
    api_key=deepseek_key,
    base_url="https://api.deepseek.com"
)

# Before
model="gpt-4o"

# After
model="deepseek-v4-pro"

# GPT doesn't have a native reasoning mode — CoT is prompt-based
# DeepSeek: Enable thinking mode
response = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=messages,
    reasoning_effort="high",
    extra_body={"thinking": {"type": "enabled"}}
)

# Before
client = anthropic.Anthropic(api_key=anthropic_key)

# After
client = anthropic.Anthropic(
    base_url="https://api.deepseek.com/anthropic",
    api_key=deepseek_key
)

from langchain.chat_models import ChatOpenAI

llm = ChatOpenAI(
    model="deepseek-v4-pro",
    openai_api_key="<DeepSeek API Key>",
    openai_api_base="https://api.deepseek.com"
)

# Robust streaming pattern
stream = client.chat.completions.create(
    model="deepseek-v4-pro",
    messages=messages,
    stream=True,
    extra_body={"thinking": {"type": "enabled"}}
)

for chunk in stream:
    delta = chunk.choices[0].delta
    if delta.reasoning_content:
        # Thinking tokens arrive first
        yield {"type": "reasoning", "content": delta.reasoning_content}
    elif delta.content:
        # Answer tokens arrive after reasoning
        yield {"type": "content", "content": delta.content}

from openai import APIError, RateLimitError

try:
    response = client.chat.completions.create(...)
except RateLimitError:
    # 429 — concurrent request limit exceeded (Flash: 2500, Pro: 500)
    time.sleep(1)
    retry()
except APIError as e:
    if e.status_code == 400 and "reasoning_content" in str(e):
        # Missing reasoning_content in tool-call loop
        # Fix: Always pass full message object
        pass
    elif e.status_code == 402:
        # Insufficient balance
        pass