LLM pricing glossary
The concepts behind every number in the LLM pricing explorer, in plain English. If you are not sure what a cached token, a batch rate, or a long-context surcharge does to your bill, start here.
Input Tokens
The tokens in everything you send the model — the prompt, system message, tools, and prior turns — billed at the input rate.
Output Tokens
The tokens the model generates in its response — billed at the output rate, typically several times higher than input.
Cost per 1M Tokens
The standard unit for LLM API pricing: dollars per one million tokens, quoted separately for input and output.
Prompt Caching (Cached Tokens)
Reusing a previously processed prompt prefix at a steep discount — often 90% off the input rate on a cache hit.
Cache Write
The one-time fee some providers charge to store a prompt prefix in the cache before you can read it back cheaply.
Batch API
An asynchronous processing mode that trades immediacy for a large discount — commonly 50% off input and output.
Context Window
The maximum number of tokens — prompt plus response — a model can consider at once, ranging from 200K to over 1M.
Max Output Tokens
The ceiling on how many tokens a model will generate in a single response — separate from the context window.
Reasoning Tokens
The hidden "thinking" tokens a reasoning model generates before its answer — billed at the output rate even though you never see them.
Tokenizer
The component that splits text into tokens — and the reason two models with the same per-token price can cost different amounts.
Long-Context Pricing
A surcharge that raises the per-token rate once a prompt crosses a token threshold — common on large-window models.
Off-Peak Pricing
Time-of-day pricing where the standard rate applies off-peak and a multiplier raises it during busy hours.
Blended Price
A single headline figure that averages input and output rates so you can rank models at a glance.
Open-Weight vs Proprietary
Whether a model’s weights are published (so any host can serve it) or kept private and sold only by the vendor.
Ready to compare? Open the LLM pricing hub, or browse by provider —Anthropic, OpenAI,Google. Working with AWS too? TheEC2 pricing glossary covers On-Demand, Reserved and Savings Plans.