> ## Documentation Index
> Fetch the complete documentation index at: https://docs.u-kiyo.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Pricing

> Use current inventory and the checkout quote, not fixed example rates.

GPU availability and rates change. There is no permanent three-GPU price table.

* [Live GPU catalog](https://pay.u-kiyo.ai/api/v1/gpus): public JSON grouped by model.
* [Available offers](https://pay.u-kiyo.ai/api/v1/offers): current individual rental offers.
* [Rent in Telegram](https://t.me/UkiyoComputeBot): choose an offer and review the quote before paying.

## Reading prices

API monetary values use USD minor units (cents). Divide by 100 to display dollars. A value of `100` means \$1.00; this is a unit example, not an advertised rate.

Catalog `minPricePerHourMinor` and `maxPricePerHourMinor` describe the **whole instance**, including its GPU quantity. Do not treat them as per-GPU prices. VRAM is per GPU, not automatically pooled memory.

The checkout quote is authoritative for your selected configuration. It shows your budget, hourly price, and estimated runtime. A higher budget buys more runtime within the offer's limits; it does not change the hardware.

## Budget and runtime

Rentals start at \$5. The upper budget limit depends on current availability and allowed duration. The app calculates and rounds the runtime; review its displayed result rather than relying on a hand calculation.

No automatic renewal or subscription is implied. Read [billing](/guides/billing) and [expiration](/guides/lifecycle) before starting long jobs.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.