How to Use the Claude API (2026 Guide)

Step-by-step tutorial • 8 min read

1. Get an API Key

1. Go to console.anthropic.com and create an account
2. Add billing information (you get some free credits to start)
3. Navigate to API Keys in the sidebar
4. Click Create Key and copy it — it starts with sk-ant-

2. Install the SDK

Python:

pip install anthropic

Node.js:

npm install @anthropic-ai/sdk

3. Your First Request (Python)

import anthropic

client = anthropic.Anthropic(api_key="sk-ant-...")

response = client.messages.create(
    model="claude-opus-4-7",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Explain quantum computing in 3 sentences."}
    ]
)
print(response.content[0].text)

4. Model Selection (2026)

Model	Speed	Cost	Best For
Claude Opus 4.7	Slower	Higher	Complex reasoning, coding
Claude Sonnet 4.6	Fast	Medium	Most tasks (default)
Claude Haiku 4.5	Fastest	Cheapest	Bulk, real-time

5. Key Features to Use

System Prompts

response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    system="You are a helpful coding assistant. Respond only with code.",
    messages=[{"role": "user", "content": "Reverse a linked list in Python"}]
)

Streaming

with client.messages.stream(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a poem"}]
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)

Prompt Caching (Save Costs)

Cache long system prompts or context — subsequent calls are 90% cheaper. Add cache_control markers to system messages.

Tool Use

Give Claude access to functions (weather, database queries, code execution). See docs.claude.com for the full tool_use spec.

6. Common Patterns

Multi-turn Conversation

messages = [
    {"role": "user", "content": "Hi"},
    {"role": "assistant", "content": "Hello! How can I help?"},
    {"role": "user", "content": "Explain recursion"}
]
response = client.messages.create(model="claude-sonnet-4-6", max_tokens=1024, messages=messages)

File Input (PDF, Images)

import base64
image_data = base64.b64encode(open("chart.png", "rb").read()).decode()
response = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": [
        {"type": "image", "source": {"type": "base64", "media_type": "image/png", "data": image_data}},
        {"type": "text", "text": "What does this chart show?"}
    ]}]
)

7. Pricing (2026)

• Opus 4.7: $15/M input, $75/M output
• Sonnet 4.6: $3/M input, $15/M output
• Haiku 4.5: $0.80/M input, $4/M output

Prompt caching reduces input cost by 90% for cached tokens.

8. Common Mistakes

• Not using system prompts — huge quality improvement
• Ignoring streaming — better UX for long responses
• Not caching long contexts — leaves 90% savings on the table
• Using Opus for everything — Sonnet is enough for most tasks
• Skipping prompt engineering — 30 minutes of prompt tuning beats model swaps

Related: ChatGPT vs Claude · Latest Anthropic News · AI Glossary