What is Pay-per-token pricing?
Pay-per-token pricing bills by the number of input and output tokens a request uses, with no subscription or seat fee — you pay only for what you consume.
A token is roughly three-quarters of a word. Language models bill separately for input tokens (your prompt) and output tokens (their response), usually quoted per million tokens. Output is typically priced higher than input because it's generated one token at a time.
Pay-per-token is the dominant model for API access because it scales precisely with usage: a prototype that makes ten calls a day costs cents, while a production app paying for what it uses never hits a subscription ceiling mid-month. Cached input (see prompt caching) is billed at a steep discount.
How Kyma does it
Kyma bills language models per token, with cached input at 10% of the input rate, and media models per image, per second, or per character. There's no subscription, and signup includes $0.50 in free credit. Every response reports its exact cost.