Are LLM input and output tokens priced equally?
Not necessarily. Multiply input tokens by the published input rate and output tokens by the output rate, then add the two charges. Using one price for all tokens can misstate the bill.
The API price pages show both rates. Cached input, batch usage, long contexts and separately billed reasoning tokens need their applicable rate rules rather than the ordinary input/output example.
Use the job cost and fee table, model memory estimates, training calculator, and price trends for the relevant inputs.
Related questions
Numbers on this page come from today's verified snapshot. Full table on the homepage; method in the methodology.