How Proxy Pricing Works

Updated: 2026-11-12

Proxies are usually priced in one of three ways: per IP for a rental period (common for datacenter and static ISP proxies), per gigabyte of traffic (common for rotating residential proxies), or per port, meaning the number of simultaneous connections to a rotating gateway. The billing model shapes your invoice more than the headline number on a pricing page, because the same job can be cheap under one model and expensive under another. This guide explains each model, what makes prices go up or down, and how to estimate your cost before you buy instead of guessing.

Priced per IP, not per GB

View pricing Free Proxy Checker

The three common billing models

Each model follows from the infrastructure behind it. Knowing that helps you read pricing pages correctly and avoid comparing numbers in different units.

Model	You pay for	Typical for	Best fit
Per IP and period	Each IP for each rental period, usually a month	Dedicated datacenter, static ISP	Accounts that need a fixed IP, always-on tools, heavy traffic
Per GB	Data that passes through the proxy	Rotating residential	Collection that needs fresh IPs, with small, predictable traffic
Per port or thread	Simultaneous connections to a rotating port	Port-based rotating plans	Frequent IP changes with a fixed number of threads

Per IP per month: the most predictable

This is the model you meet with dedicated IPv4 and static ISP proxies. Each IP has a price per rental period; during that period the IP is yours, and traffic is usually not billed separately. Your bill depends on two variables only, the number of IPs and the number of periods, so costs stay predictable even if you download a lot, watch video or run tools all day.

Check whether “unlimited bandwidth” hides conditions such as throttling after a threshold. Prices per IP also vary by type, since ISP addresses are harder to source than datacenter ones, and by country. If the types are new to you, read our ISP vs datacenter vs residential comparison.

Per GB and per port: flexible but harder to predict

Rotating residential providers pay for a network of home devices, so they usually bill by traffic. You get frequent new IPs, but every page, image and video adds to the bill. Something as simple as browsing an image-heavy social feed can use a surprising amount of data. This model suits text-heavy data collection where each request needs a fresh IP.

Port-based plans charge for simultaneous connections while the IP behind each port changes on a schedule. Cost is fixed per thread, but you cannot keep one IP for long, which makes it a poor fit for logged-in accounts. See static vs rotating proxies for the trade-offs.

What makes a proxy cheaper or more expensive

IP type: datacenter is usually cheapest, ISP costs more, residential and mobile tend to cost the most per unit.
Exclusivity: dedicated IPs cost more than shared ones because one customer pays for each address. See private vs shared proxies.
Country: IP ranges are scarcer in some countries, so prices differ by location.
Rental length: some providers discount longer terms in exchange for paying more upfront.
Volume: many providers discount larger orders.
Operational quality: clean IPs, 24/7 support and a refund policy all cost money, which is why very cheap plans often lack one of them. See the risks of cheap proxies.

How to estimate your cost before buying

You do not need exact numbers, just the right formula for the model you pick.

Step 1, count your needs: how many accounts need their own IP and how many threads run at once. Our guide on how many proxies you need walks through this.
Step 2, choose a model: fixed IPs mean per-IP pricing; fresh IPs with light traffic may suit per-GB.
Step 3, apply the formula: per IP, cost equals IPs times months times unit price; per GB, it equals estimated data per run times runs times price per GB.
Step 4, add a buffer for failed IPs, new accounts or reruns.
Step 5, test and adjust: run a small batch for a week, measure real usage, then commit.

One job, two billing models

Imagine a group of seller accounts that log in daily, upload product images and watch short videos. Under per-GB billing the invoice grows with every image and clip and becomes hard to forecast as the group grows. Under per-IP billing you pay a fixed amount for one IP per account, traffic does not change the price, and each account keeps a stable address.

Now imagine collecting text prices from many pages, a few dozen kilobytes per request, with a new IP each time. Per-GB pricing may be cheaper there. No model is cheapest for everything; the right one matches how you actually work.

How VinaProxy prices proxies

VinaProxy charges per IP per rental period for both dedicated datacenter IPv4 (seven countries) and static ISP proxies (Vietnam and the US). Each IP goes to one customer and bandwidth is unlimited, so cost depends only on how many IPs you rent and for how long. You pay in VND from your account balance, topped up by bank transfer or card, with no traffic charges.

Current prices by type and country are on the pricing page. To keep an IP into the next period, renew it from the dashboard; see how to renew a proxy and keep the same IP.

Frequently asked questions

Is per-IP or per-GB pricing cheaper?

It depends on the job. Heavy traffic or accounts that need a fixed IP usually favour per-IP pricing. Light traffic that needs a new IP each time can favour per-GB.

Is unlimited bandwidth real?

Under per-IP pricing, many providers do not bill traffic separately. Ask whether speed is reduced after any threshold so there are no surprises.

Why do ISP proxies cost more than datacenter IPv4?

Clean ranges registered to consumer internet providers are much harder to obtain than hosting IPs, so supply is smaller and each IP costs more.

Should I rent for a longer period to get a better price?

Only after testing and when you are sure you will keep the IPs. With a new provider, start with a short period and a small quantity.

How do I estimate traffic for a per-GB plan?

Run a few test jobs through the proxy, read the total usage in the provider's dashboard or your own tooling, then multiply by expected runs and add a margin.