Claude API Access: Complete Setup Guide for Sonnet, Opus & Haiku (2026)

Call Claude Sonnet, Opus, and Haiku from Python, Node.js, Cursor, LangChain, and any OpenAI-compatible tool — with a single API key.

Claude by Anthropic is one of the most capable AI model families available today. It excels at reasoning, coding, creative writing, long-document analysis, and following complex instructions. This guide shows you everything you need to access the Claude API — from setup to advanced features — with practical, copy-paste-ready code examples.

Claude Model Family Overview

Anthropic offers three Claude model tiers, each optimized for different use cases and budgets:

Which Claude Model Should You Use?

• Claude Sonnet is the recommended default for most developers. It offers the best balance of intelligence, speed, and cost. Use it for coding, content generation, data analysis, and production chatbots.
• Claude Haiku is ideal for high-volume tasks where cost matters most: classification, sentiment analysis, simple Q&A, and bulk content moderation.
• Claude Opus is reserved for the hardest problems: deep research, complex multi-step reasoning, and tasks where maximum accuracy justifies the premium cost.

What Makes Claude Special?

Claude stands out from other AI models in several ways:

1. Exceptional Instruction Following

Claude is widely regarded as the best model for following complex, multi-step instructions. If you give it a detailed system prompt with specific formatting requirements, it adheres more reliably than most alternatives.

2. 200K Token Context Window

Claude can process up to 200,000 tokens of context — equivalent to a 500-page book. This makes it ideal for analyzing long documents, codebases, or multi-turn conversations without losing track of earlier context.

3. Nuanced Writing Quality

Claude's writing style is natural and sophisticated. For content creation, marketing copy, and creative writing, many developers prefer Claude's output over GPT's more predictable patterns.

4. Strong Coding Performance

Claude Sonnet consistently ranks at or near the top of coding benchmarks. It excels at understanding entire codebases, generating complex functions, and debugging subtle issues.

Prerequisites

• A Token1.US account with a topped-up balance (start with $1+)
• Your API key from the dashboard
• The OpenAI SDK for your language (or any HTTP client)

Python Setup

Install the SDK

pip install openai

Basic API Call

from openai import OpenAI

client = OpenAI(
    api_key="your-token1-us-key",
    base_url="https://discount-token.com/v1"
)

response = client.chat.completions.create(
    model="claude-sonnet-4-20250514",
    messages=[
        {"role": "system", "content": "You are a senior Python developer."},
        {"role": "user", "content": "Explain decorators with a practical example."}
    ],
    max_tokens=1024
)

print(response.choices[0].message.content)
print(f"\nTokens: {response.usage.total_tokens}")

Streaming Responses

stream = client.chat.completions.create(
    model="claude-sonnet-4-20250514",
    messages=[{"role": "user", "content": "Write a haiku about APIs."}],
    stream=True
)

for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

Multi-Turn Conversation

messages = [
    {"role": "system", "content": "You are a helpful coding assistant."},
    {"role": "user", "content": "What is a context manager in Python?"},
]

response = client.chat.completions.create(
    model="claude-sonnet-4-20250514",
    messages=messages
)

# Add Claude's response to history
messages.append({"role": "assistant", "content": response.choices[0].message.content})

# Ask a follow-up question
messages.append({"role": "user", "content": "Show me an example."})

response2 = client.chat.completions.create(
    model="claude-sonnet-4-20250514",
    messages=messages
)
print(response2.choices[0].message.content)

Node.js Setup

npm install openai

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "your-token1-us-key",
  baseURL: "https://discount-token.com/v1",
});

const response = await client.chat.completions.create({
  model: "claude-sonnet-4-20250514",
  messages: [{ role: "user", content: "Hello, Claude!" }],
});

console.log(response.choices[0].message.content);

Using Claude API with Popular Tools

Cursor IDE

Open Cursor Settings → Models → set OpenAI Base URL to https://discount-token.com/v1 and enter your Token1.US API key. Select claude-sonnet-4-20250514 as the model. Cursor will use Claude for code generation, refactoring, and chat.

LangChain

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="claude-sonnet-4-20250514",
    api_key="your-token1-us-key",
    base_url="https://discount-token.com/v1"
)

# Simple invocation
response = llm.invoke("What is 15 * 37?")
print(response.content)

# With prompt template
from langchain_core.prompts import ChatPromptTemplate

prompt = ChatPromptTemplate.from_messages([
    ("system", "You are a helpful assistant that translates English to French."),
    ("user", "Translate: {text}")
])

chain = prompt | llm
result = chain.invoke({"text": "Hello, how are you?"})
print(result.content)

Dify

Add a model provider in Dify settings with base URL https://discount-token.com/v1 and your Token1.US API key. You can then use Claude Sonnet in any Dify AI app.

curl

curl https://discount-token.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your-token1-us-key" \
  -d '{
    "model": "claude-sonnet-4-20250514",
    "messages": [{"role": "user", "content": "Hello!"}],
    "max_tokens": 500
  }'

Claude API Pricing Deep Dive

Understanding Claude's pricing helps you estimate costs accurately:

Use our interactive cost calculator for your own scenarios.

Troubleshooting Common Claude API Errors

Claude vs Other AI Models

How does Claude compare to GPT and DeepSeek? Here is a quick summary:

For a full comparison, see our detailed model comparison page.

Best Practices for Claude API

• Use system prompts effectively: Claude responds well to clear, detailed system prompts. Set expectations and formatting rules up front.
• Leverage the long context window: Claude's 200K token window means you can pass entire documents, codebases, or long conversation histories without summarization.
• Be specific: Claude follows instructions precisely. If you want a specific format (JSON, markdown, bullet points), say so explicitly.
• Use streaming for better UX: Claude's streaming is smooth and reduces perceived latency for end users.
• Choose the right tier: Use Haiku for simple tasks, Sonnet for most work, and Opus only when you need maximum reasoning power.

Frequently Asked Questions

Related Guides

• What Is an AI API Gateway?
• GPT API Access Guide
• GLM API Pricing Guide
• DeepSeek API Guide
• AI Model Comparison
• Cheapest AI API Guide
• 5-Minute Quick Start