Token Counting and LLM API Pricing Quiz

Answers and explanations appear as you take the quiz.

1. What is a common rule of thumb for English text and tokens?
   • One token equals one full sentence
   • One token is about four characters, or three-quarters of a word
   • One token equals exactly one word
   • One token is about 100 characters
2. Why are output tokens usually priced higher than input tokens?
   • Output tokens are longer words
   • Providers pay authors for output text
   • Output is generated one token at a time, while input can be processed in parallel
   • Output tokens are stored permanently
3. With made-up prices of $3 per million input tokens and $15 per million output tokens, what does a call with 6,000 input and 800 output tokens cost?
   • $0.03
   • $0.30
   • $0.018
   • $3.00
4. Why do chatbot costs often grow as a conversation goes on?
   • Providers raise prices mid-conversation
   • Output tokens become more expensive
   • Models slow down over time
   • The whole conversation history is resent as input on every turn
5. How are a reasoning model's internal thinking tokens typically billed?
   • They are free
   • As output tokens, even when not shown in full
   • As input tokens
   • As a flat monthly fee
6. Which tool does Anthropic provide for counting tokens before a request?
   • The tiktoken library
   • A countTokens method on Gemini
   • A token-counting endpoint that accepts messages, system prompt and tools
   • A browser extension
7. What do batch API endpoints typically offer?
   • Around half price in exchange for asynchronous, slower results
   • Free output tokens
   • Faster responses at double the price
   • Unlimited context windows
8. Why is the lowest per-token price not a reliable guide to total API cost?
   • Tool definitions count as input
   • Actual usage fields should be logged
   • Non-English text can use more tokens
   • The model with the lowest per-token price is always cheapest per task