Proxy pricing comes in two main shapes: pay by the gigabyte of traffic, or pay for a fixed number of IPs. Which is cheaper depends entirely on your workload — but for most rotating residential and mobile use, per-gigabyte billing is the model that scales with what you actually consume.
Per-gigabyte (bandwidth) pricing
You pay for the data your requests transfer, regardless of how many IPs you touch. This fits rotating pools perfectly: you might cycle through thousands of residential IPs in an hour, but you're only billed for the gigabytes moved. Bigger prepaid packs lower the rate per GB.
On this network, bandwidth doesn't expire — a pack depletes only as you use it — so you can buy in bulk for a better rate without a clock running against you.
Per-IP pricing
Here you rent a set number of dedicated IPs for a period, with unlimited bandwidth on them. This can win for steady, high-volume traffic concentrated on a few IPs — for example a fixed set of datacenter proxies running flat out around the clock.
It loses when you need IP diversity: paying per IP for a large rotating pool is far more expensive than paying for the gigabytes you actually move through it.
Estimating your usage
To budget per-GB, estimate the average size of a response and multiply by your request volume. Plain HTML pages and API JSON are small (often tens of kilobytes); rendering full pages with images in a headless browser costs far more bandwidth.
A practical lever: scrape only what you need. Block images and media, request compact endpoints, and cache — each cuts gigabytes directly off the bill.