What Are Tokens in AI and Why Do They Matter for Your Bill?

If you use AI APIs, you'll hear a lot about "tokens." Let's explain what they are and why they're so important.

What is a Token?

A token is the basic unit of text that language models process. It's not exactly a word:

• "Hello" = 1 token
• "Intelligence" = 2-3 tokens
• "PromptRouter" = 2 tokens
• A space also counts

General rule: 1 token ≈ 4 characters in English, or ¾ of a word.

Why Do They Matter?

Because you pay for them. AI providers charge per million tokens:

Model	Input (per 1M)	Output (per 1M)
GPT-4 Turbo	$10	$30
Claude Opus	$15	$75
GPT-3.5	$0.50	$1.50
Claude Haiku	$0.25	$1.25

Input vs Output Tokens

• Input tokens: Your question + system context
• Output tokens: The model's response

Output tokens are usually more expensive because they require more computation.

Practical Example

Imagine this prompt:

System: You are a customer service assistant.
User: What is your return policy?

• Input: ~20 tokens
• Estimated output: ~100 tokens

With GPT-4 Turbo:

• Input: 20 × $0.00001 = $0.0002
• Output: 100 × $0.00003 = $0.003
• Total: $0.0032 per request

If you have 100,000 requests/month: $320/month

With an economical model like GPT-3.5:

• Total: $0.0002 per request
• 100,000 requests/month: $20/month

How to Optimize Token Usage

1. Concise prompts: Eliminate unnecessary words
2. Limit responses: Use max_tokens to control output
3. Choose the right model: Don't use a cannon to kill a fly
4. Use intelligent routing: Let PromptRouter choose for you

Savings with PromptRouter

PromptRouter analyzes each request and decides:

• Is it a simple question? → Economical model
• Does it require complex reasoning? → Premium model

Result: You use premium tokens only when you really need them.

---

Optimize every token. Get started with PromptRouter.