Cheapest residential proxies in 2026

On published prices, the cheapest residential proxies in our database are Evomi at $0.49/GB on a subscription and DataImpulse at $1/GB, with DataImpulse falling to $0.80/GB at 1 TB. Both come with conditions, which is why this ranking sorts by effective price at about 100 GB rather than by the first number on a homepage.

Pool sizes are vendor claims, and they differ: DataImpulse advertises 90M IPs and Evomi 54M, against 400M at Bright Data. A smaller pool is not automatically worse. It matters when you need one narrow city or ASN, because a small slice of a small pool means you will see the same IPs more often.

Thordata advertises a headline from $0.65/GB, but we have not verified its volume tiers or minimum commitments, so treat that figure as a lead to check, not a price to budget on. The table above lists the providers at or below $2.75/GB at 100 GB. For the wider market, including pricier networks with deeper targeting, see the best residential proxies.

The cheap rate is often the price of the stripped-down product, and the real cost shows up once you turn on the options you need. Four traps cause most of the surprises.

Price the exact configuration you need, not the base rate. A $1/GB rate that doubles with city targeting is really $2/GB, which puts it level with plans that include targeting from the start. SOAX, for example, includes city and ASN targeting on every plan, so its $25 entry pack (5 to 12.5 GB) is a different kind of offer from a bare country-level rate. Write down the filters you need first, then compare.

Pay as you go suits irregular jobs, and a monthly subscription suits steady usage, as long as you do not pay for traffic you never send. The right choice depends on how predictable your volume is.

At several providers it does not, and that matters more than a few cents per GB. IPRoyal, DataImpulse, NodeMaven and Rayobyte are listed in our data with non-expiring traffic, so you can buy a larger, cheaper pack and spend it over months.

On a monthly subscription, unused GB usually vanish at renewal, so the effective price rises every month you under-use the plan. If you are not sure about your volume, a non-expiring pack removes that risk.

Test before you buy in bulk, and compare trial terms the way you compare price. They differ a lot between providers and the low-price ones are not always the most generous.

Byteful gives a 1 GB residential trial, Bright Data offers a no-card free trial, SOAX sells a paid trial at $1.99, Oxylabs gives a $5 credit towards any product, and Proxy-Seller has a 3-day $1.99 residential trial. IPRoyal has no trial, and DataImpulse trial terms are not verified in our data. Rayobyte offers only a 2-day refund window, so test quickly if you buy there.

Use the trial to check the boring things first: that your authentication method works (credentials or IP whitelist), that the protocol you need is supported, that the targeting syntax returns the city or ASN you asked for, and that sticky sessions hold for as long as your job needs. A trial that passes those checks is worth more than a lower price on a product you cannot configure.

The cheapest residential proxies are the ones with the lowest bill per successful page, which is not always the lowest price per GB. A network that gets blocked on your target forces retries, and every retry is paid traffic.

Run 100 URLs from your real target list through each shortlisted provider with the same settings. Count the successes, then divide the traffic bill by that count. Here is the arithmetic with made-up round numbers: if the test used 0.5 GB at $2/GB, the bill is $1, and if 80 of the 100 pages came back correct, each good page cost $0.0125. A provider at $1/GB that only returned 40 good pages would cost the same $0.0125 per page, so the cheaper rate bought you nothing. Block images, fonts and video in your client where you can, since they often account for most of a page's weight. If you want to model the numbers first, the proxy cost calculator does the arithmetic.

Free residential proxies from public lists are rarely worth using, because you do not know who runs the exit or what they log. If you only need to test a script, a vendor's free allowance is the safer route.

Webshare's free plan includes 10 datacenter IPs and 1 GB per month, and several of the trials above give a small residential allowance at no cost. They are enough to check authentication, targeting syntax and a first sample of pages, but not enough to judge success rates at scale.

Choose on effective cost at your volume, with the add-ons you need switched on. Write down your monthly GB, the targeting you require and whether you need traffic to expire or not, then compare two or three providers on those terms.

For a small or irregular project, a non-expiring pack from DataImpulse or similar is a low-risk start. For steady volume, check Evomi's profile and read its footnotes on filters before assuming the base rate applies. If your targets are tolerant, a residential vs datacenter proxies comparison may show you do not need residential bandwidth at all.