- There are only about 4.3 billion IPv4 addresses, and the free ones ran out long ago, so they are expensive
- There are almost endless IPv6 addresses, but many websites treat a whole /64 subnet as one user
- A website without IPv6 support will not open through a proxy that exits only over IPv6
- With an IPv4 proxy, requests that bypass it, for example through WebRTC, can reveal your IPv6 - check this
- For accounts and websites without IPv6 you need IPv4, for bulk tasks on websites with IPv6 it will also do
When you choose a proxy, sooner or later you run into the question: IPv4 or IPv6. Some offers are noticeably cheaper than others, and the difference is often exactly the address version. But cheapness has a price: some websites do not open over IPv6 at all, and many treat such addresses more strictly than it seems.
Below: what IPv4 and IPv6 are, why second-version addresses are cheap, where they work and where they do not, and how to check that the website sees exactly the address you expect.
Two versions of one address
An IP address is a device's number on the internet. A website uses it to know where to send the response and, along the way, which country and network the request came from. There are two versions of these numbers.
IPv4 is the old and most widespread version. An address looks like four numbers separated by dots, for example 203.0.113.25. There are about 4.3 billion such addresses in total, which is fewer than there are devices in the world.
IPv6 is the new version. An address is long, made of eight groups separated by colons, for example 2001:db8:85a3::8a2e:370:7334. There are so many addresses in this system that there are enough for every device with a huge margin.
Long IPv6 addresses are usually written in a shortened form. Leading zeros in a group are dropped, and several all-zero groups in a row are replaced with two colons. So the same address can look different, and this is not a mistake.
Why IPv4 addresses ran out
Free IPv4 addresses started running out back in 2011, when the central pool handed out the last blocks to the regional registries. The European registry gave out its last free block in 2019. Since then, a new address can only be bought or rented from someone who already has it.
That is why providers economize: one IPv4 address is often shared by many subscribers at once through a common exit. And blocks for datacenters cost real money, which is reflected in the price of IPv4 proxies.
Why IPv6 proxies are cheap
With IPv6 the situation is the opposite. A provider usually gives a customer not one address but a whole subnet, most often a /64. It contains more addresses than the entire IPv4 internet. A server owner can hand out millions of addresses from such a subnet at almost no cost.
Hence the low prices for IPv6 proxies: there are many addresses, and each costs pennies. But this abundance has a downside, and it is known not only to sellers but also to websites.
One subnet, one user
Since a home subscriber usually gets a whole /64 subnet, it makes no sense for a website to count each address in it as a separate person. Many websites do exactly that: limits, blocks and reputation are counted for the whole /64 subnet at once.
For proxies this means a simple thing. A thousand IPv6 addresses from one subnet can look like one user to such a website. If the subnet gets a restriction, all addresses in it get it at once.
That is why, when you choose IPv6, what matters is not the number of addresses but the number of different subnets. A hundred addresses from a hundred different subnets are more useful than a million from one.
With IPv4 the logic is similar, only the scale is different. Websites also look at neighboring addresses and at who owns the network. But a home subscriber with IPv4 usually gets one address, not millions, so each such address weighs more for a website. This is exactly why platforms with accounts trust home IPv4 addresses more than addresses from a large IPv6 subnet.
Not all websites understand IPv6
Large platforms, search engines and social networks have long worked over both versions. But many smaller websites, some online stores, local services and old systems are still available only over IPv4. To open them over IPv6, the website simply has no address for it.
If a proxy goes online only over IPv6, such a website will not open through it. You will get a connection error or a timeout, even though the proxy works fine. This is a common cause of confusion when "the proxy does not work" only on some websites.
You can check whether a website has IPv6 with the nslookup or dig command: if the domain has an AAAA record, it has an IPv6 address. If it has only an A record, the website is available only over IPv4.
Dual-stack: when both versions work
Most modern networks work in dual-stack mode: a device has both IPv4 and IPv6. When you open a website that supports both versions, the system usually chooses IPv6 and switches to IPv4 if there are problems.
This is convenient for ordinary internet use, but with proxies it creates surprises. A website may see one version of the address or the other, depending on how the proxy, the program and the website itself are set up. So it is better not to rely on guesses but to check.
This is especially true of mobile internet. Many carriers in Europe already give subscribers IPv6 and provide IPv4 through a common exit shared by many people. A phone sharing its internet with a laptop can quietly bring IPv6 into the picture where there was none at home.
How to check which address the website sees
- Open the my IP page through the proxy. An address of four numbers separated by dots is IPv4, a long address with colons is IPv6.
- Compare the country and network type: they should match what you chose in the generator in the dashboard.
- Look at the WebRTC block on the same page. Your real address must not be there, neither IPv4 nor IPv6.
- Check the website you need separately: if it opens directly but not through the proxy, the cause may be the address version.
In Prokso there is one connection address, gw.prokso.com, and you can check which address the website will see on the "my IP" page. The check takes a minute and removes most of the questions before you start working.
IPv6 leak with an IPv4 proxy
There is also the opposite problem. You have connected a proxy, the website sees the proxy's address, but your device also has IPv6 from your home provider. If some requests go past the proxy, they go out directly over IPv6 and show your real address.
The most common causes of such a leak:
- WebRTC in the browser. It can find out the device's addresses bypassing the proxy, including IPv6, and the website will see them through a script.
- Programs that do not use the system proxy settings and go online directly.
- The proxy is set up in only one browser, while the account is also open in an app on the same device.
- DNS requests go through the provider rather than through the proxy and reveal which websites you open.
What to do about it: in an antidetect browser, set WebRTC to spoofing or disabled mode; in an ordinary browser, use an extension to manage WebRTC. If you do not need IPv6, you can turn it off in the system's network connection settings. After any changes, open the "my IP" page again and make sure your real address is not there.
When IPv6 fits and when you need IPv4
| Task | What to choose | Why |
|---|---|---|
| Accounts on social networks and marketplaces | IPv4 | Platforms are stricter with IPv6 subnets, and linking accounts by a /64 subnet is risky |
| Websites without an AAAA record | IPv4 | Such a website simply will not open over IPv6 |
| Bulk collection on websites with IPv6 | IPv6 will do | If subnet limits do not get in the way of the task |
| Availability and geo results checks | Depends on the website | Check which version the website serves content over |
| Banks and government websites | Not through a proxy | Some of these websites are closed to the service's addresses |
The last row is not only about the address version. Some websites are closed to all of the service's addresses - the list is on the blocked sites page.
If you need your "own" permanent IPv4 rather than an address from a shared pool, read the article IPv4 and static proxies: alternatives. And how home addresses work and why platforms trust them is explained on the residential proxies page.
A short checklist before you start
- Find out whether the website you need has IPv6: an AAAA record in DNS.
- For accounts, rely on IPv4 and a consistent browser profile.
- For IPv6, count subnets, not individual addresses.
- Check the address and WebRTC on the "my IP" page in every profile.
- If you do not need IPv6 on the device, turn it off to avoid leaks.
FAQ
What is the difference between IPv6 and IPv4 in simple terms?
These are two versions of addresses on the internet. IPv4 addresses are short, made of four numbers, and there are few of them. IPv6 addresses are long, made of eight groups, and there are enough of them for practically everything. A website can work over one version or over both at once.
Why are IPv6 proxies cheaper?
Because there is a huge number of IPv6 addresses, while an IPv4 block today has to be bought or rented. But a cheap address is not always useful: some websites do not support IPv6, and many restrict the whole subnet at once.
How do I find out which version a website sees me on?
Open the "my IP" page through the proxy. If it shows an address of four numbers separated by dots, the website sees IPv4. If it is a long address with colons, it is IPv6.
What is an IPv6 leak?
It is a situation where some requests go past the proxy directly over IPv6 and show the website your real address. Most often this happens through WebRTC in the browser or in programs that do not use the proxy settings.
Which address does Prokso give: IPv4 or IPv6?
There is one connection address: gw.prokso.com. Check which address a particular website will see on the "my IP" page by opening it through the proxy.
