How does token-based pricing affect an application?
Why B is the answer
Longer prompts and outputs generally cost more
Input and output tokens drive cost, so longer prompts and longer outputs generally cost more.
In detail
- Pricing follows tokens, so longer prompts and outputs generally cost more.
- Both input and output tokens count.
Use case
An illustrative scenario, not a real incident.
A team trims a long prompt, and the monthly bill falls, because each call uses fewer tokens.
See it in different ways
mindmap
root((Generative AI))
Token
Unit of text
Drives cost
Embedding
Meaning as numbers
Chunking
Focused passages
Hallucination
Confident and false
Temperature
Varietymindmap
root(("Token cost"))
n1["Input"]
n2["Prompt"]
n3["Context"]
n4["Output"]
n5["Generated text"]Key point: More tokens usually cost more.
Why the other options are wrong
- A. Costs depend only on which AWS Region you are using
- Region is not the only factor.
- C. Every request costs the same amount, whatever its length
- Cost varies with length.
- D. Only the number of users affects the monthly price
- User count does not set token cost.