Per-GB, per-port, or unlimited: which pricing model fits your workload

There is a lot written about which proxy network to choose and almost nothing honest about how to pay for one. That is not an accident: most providers sell a single billing model, so their advice arrives at the same conclusion every time. We sell both — pay-as-you-go per gigabyte on Residential / ISP and Mobile, flat and port-based on Budget Unlimited and Premium Unlimited — which means we have no particular reason to push you either way. Here is the actual math.

The two models, and what each one is really charging for

Per-GB charges for data. You pay for what leaves the network, and nothing when you are idle. The number of IPs you can touch is effectively unbounded; the meter is on volume.

Per-port (or flat) charges for concurrency. You buy a number of simultaneous connections, and the bandwidth through them is unlimited. The meter is on how much you do at once, not how much you move.

That distinction is the whole article. One model prices the size of the job, the other prices the shape of it. Almost every bad proxy invoice comes from paying for the wrong dimension.

The break-even is one division

You do not need a spreadsheet. Take your monthly data volume and multiply it by the per-GB rate. If that number is bigger than the flat plan, the flat plan wins.

Residential / ISP starts at $1.40/GB. Budget Unlimited starts at $51.99/mo for 5 ports. So:

$51.99 ÷ $1.40/GB ≈ 37 GB

Under roughly 37 GB a month, per-GB is cheaper. Over it, the port-based plan is — and the gap widens fast, because on the flat plan the next gigabyte is free and on the metered plan it never is. At 200 GB/month you would be looking at about $280 on per-GB against $51.99 flat.

The same division works at the top end. Premium Unlimited starts at $149/day, which breaks even against $1.40/GB at roughly 106 GB per day. If you are moving that much daily, flat is the cheaper shape. If you are moving 5 GB a day, it is an absurd waste and you should be on the meter.

Do the division with your own numbers before you talk to anyone in sales, ours included.

But volume is only half the question

Here is where the simple math stops. The flat plan is not unlimited in every dimension — it is unlimited in bandwidth and finite in concurrency. Five ports means five simultaneous connections. If your job needs forty at once, five ports will not do it no matter how cheap the gigabytes are, and you are buying 25 ports or more.

So the real decision is two-dimensional:

• Low volume, low concurrency → per-GB. You are paying pennies and skipping the commitment.
• High volume, modest concurrency → port-based. This is the sweet spot and the biggest saving on the page.
• Low volume, high concurrency → awkward. Per-GB will be cheap but you may hit practical limits; port-based will be underused but predictable. Price both.
• High volume, high concurrency → flat, and size the ports properly.

We wrote a whole post on sizing that second dimension — how many proxies do you actually need — because it is the number people get wrong most often. The short version: it has almost nothing to do with total volume and almost everything to do with how much happens simultaneously.

What “unlimited” should mean, and what to check

“Unlimited” is not a regulated word in this industry, and you should treat it as a marketing term until someone tells you what the cap actually is. There is always a cap. The only question is whether it is stated.

On our port-based plans the honest description is: unlimited bandwidth, capped by concurrency. You buy 5, 15, 25, 50 or 100 ports and the data through them is not metered. The ceiling is how many requests you can have in flight, not how many gigabytes you can pull.

When you evaluate anyone’s unlimited plan — again, ours included — ask these three things and get the answers in writing:

1. What is the concurrency limit? If it is not stated, it is enforced anyway, just invisibly.
2. Is there a throughput cap, and is it per port or per account? “Unlimited GB at 1 Mbps” is not unlimited in any useful sense.
3. What happens when you exceed it — throttle, queue, or error? Silent throttling is the one that quietly wrecks a pipeline.

A provider that answers those three plainly is more useful to you than one with a bigger number on the pricing page.

Where per-GB genuinely wins

It is not just for small jobs. Per-GB is the right model when precision matters more than volume.

City-level and ASN targeting, and the cleaner IP quality that comes with Residential / ISP, live on the metered plans. If you are doing ad verification from twelve specific cities, or checking local rankings from a named network, you are moving very little data and you need the targeting to be exact. Paying $1.40/GB for a few gigabytes is nothing, and the flat plan would not give you the precision anyway.

Mobile is metered-only for the same reason, at $4.00/GB: real 4G/5G carrier IPs are a scarce resource and the price reflects what they cost to keep. Nobody sells them flat, and you should be suspicious of anyone who claims to.

And IPv6 is its own model — $0.05/thread/day — because the thing that is scarce there is concurrency, not addresses. You buy threads.

The mistake that costs the most

It is not choosing the wrong model. It is choosing the right model and then feeding it work it should never have been given.

On a metered plan, the single biggest line item is usually data you never wanted: images, fonts, video, ad creative, and tracking calls pulled down by a headless browser because nobody told it not to. Blocking third-party hosts and media in a browser routinely removes more than half the bytes on a page. That is a direct, immediate cut to a per-GB bill and it takes a few lines of code — there is a snippet in our post on measuring success rates, which is the same fix for a different problem.

The corollary: if you have done that and the meter is still expensive, that is the signal to move to flat. Not before.

A rule of thumb to leave with

• Moving under ~40 GB/month → stay on per-GB; the flat plan is a solution to a problem you do not have.
• Moving hundreds of GB with country-level targeting → port-based, and it is not close.
• Need city, state, or ASN precision → per-GB, whatever the volume, because that is where the targeting lives.
• Need guaranteed throughput and dedicated resources → flat, and buy the speed tier rather than the gigabytes.

Work out your two numbers first — gigabytes per month, and requests in flight at peak. Those two decide it. Everything else, including which provider you pick, is downstream of them.

If you are between two models with real numbers in hand, tell us both figures and we will tell you which of ours is cheaper, including when the answer is the smaller plan.