💰How AI Pricing Works

Understand tokens, pricing tiers, and how to estimate and reduce your AI costs

How AI Pricing Works

When you use Claude via the API, you pay based on the amount of text you send and receive. Understanding how pricing works helps you build cost-effective applications.

What Are Tokens?

Tokens are the basic unit of text that AI models process. A token is not exactly one word — it's more like a "chunk" of text.

Rules of Thumb:

1 token ≈ 4 characters in English
1 token ≈ ¾ of a word in English
100 tokens ≈ 75 words
1,000 tokens ≈ 1.5 pages of text
Non-Latin scripts (Arabic, Chinese, etc.) use more tokens per word (roughly 1.5x-2x)

Examples:

Text	Approximate Tokens
"Hello"	1
"Hello, how are you today?"	7
A short email	50-100
A full page of text	500-700
A full book (80,000 words)	~100,000

Input Tokens vs Output Tokens

Every API call involves two types of tokens:

Input tokens — what you send to Claude (your question, context, system instructions)
Output tokens — what Claude sends back to you (the response)

Important: Output tokens are always more expensive than input tokens (typically 3-5x more).

Claude Model Pricing Table

Model	Input (per 1M tokens)	Output (per 1M tokens)	Context Window
Claude Opus 4	$15.00	$75.00	200K
Claude Sonnet 4	$3.00	$15.00	200K
Claude Haiku 3.5	$0.80	$4.00	200K

How to Calculate Costs

Formula:

Cost = (input_tokens × input_price) + (output_tokens × output_price)

Practical example with Claude Sonnet 4:

Sending: 1,000 input tokens
Receiving: 500 output tokens
Cost = (1,000/1,000,000 × $3.00) + (500/1,000,000 × $15.00)
Cost = $0.003 + $0.0075 = $0.0105 (just over one cent!)

Code Example: Tracking Costs

JavaScript
// Track cost after each API call
const response = await client.messages.create({
  model: "claude-sonnet-4-20250514",
  max_tokens: 1024,
  messages: [{ role: "user", content: "What is the capital of France?" }],
});

const inputTokens = response.usage.input_tokens;
const outputTokens = response.usage.output_tokens;

// Claude Sonnet 4 pricing
const inputCost = (inputTokens / 1_000_000) * 3.0;
const outputCost = (outputTokens / 1_000_000) * 15.0;
const totalCost = inputCost + outputCost;

console.log("Input tokens:", inputTokens);
console.log("Output tokens:", outputTokens);
console.log("Total cost: $" + totalCost.toFixed(6));

Python Example

Python
import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    messages=[{"role": "user", "content": "What is the capital of France?"}],
)

input_tokens = response.usage.input_tokens
output_tokens = response.usage.output_tokens

# Claude Sonnet 4 pricing
input_cost = (input_tokens / 1_000_000) * 3.0
output_cost = (output_tokens / 1_000_000) * 15.0
total_cost = input_cost + output_cost

print(f"Input tokens: {input_tokens}")
print(f"Output tokens: {output_tokens}")
print(f"Total cost: ${total_cost:.6f}")

Usage Tiers and Rate Limits

Anthropic has different usage tiers:

Tier	Monthly Spend Limit	Requests per Minute
Free	Very limited	5 RPM
Tier 1	$100	50 RPM
Tier 2	$500	100 RPM
Tier 3	$1,000	200 RPM
Tier 4	$5,000+	400 RPM

Tips to Save Money

Use the right model — Don't use Opus for simple tasks that Haiku can handle
Minimize input tokens — Remove unnecessary text from your prompts
Set max_tokens wisely — Don't request more than you need
Use prompt caching — Save up to 90% on repeated prompts
Batch processing — 50% discount when you don't need real-time responses
Monitor usage — Track costs daily to avoid surprises

Estimating Real Project Costs

Let's say you're building a customer support chatbot:

Parameter	Value
Messages per day	1,000
Avg input tokens per message	500
Avg output tokens per message	300
Model	Claude Haiku 3.5
Daily cost	(500K/1M × $0.80) + (300K/1M × $4.00) = $1.60
Monthly cost	~$48

With Sonnet 4 instead of Haiku, the monthly cost would be ~$195. With Opus 4, it would be ~$975.

Cost Comparison: Claude vs Competitors

Provider	Model	Input/1M	Output/1M
Anthropic	Claude Sonnet 4	$3.00	$15.00
OpenAI	GPT-4o	$2.50	$10.00
Google	Gemini 1.5 Pro	$3.50	$10.50
Anthropic	Claude Haiku 3.5	$0.80	$4.00
OpenAI	GPT-4o mini	$0.15	$0.60

Prices change frequently. Always check the official pricing pages for the latest numbers.

Useful Tools

Anthropic Console — Dashboard showing usage and costs
Usage API — Programmatically track usage
Response usage field — Every API response includes token counts used

Key Takeaways

You pay per token, not per request
Output tokens cost 3-5x more than input tokens
Choose the smallest model that meets your quality needs
Use caching and batching to save significantly
Always monitor and set spending limits

Next: We'll learn how to choose the right Claude model for your specific use case.

Module 1

8/9

🧬 How LLMs Actually Work

AI Safety & Responsible Use 🛡️

8/9