VPN speed comparison
TOP 12 VPNs results
Why 90% of the VPN speed comparisons make no sense
VPN review and comparison often use speed test results. The screenshots look nice, and the visitors get an idea of which service performs better. Unfortunately, most of the time, the results are not based on any solid foundation and can even be deceptive.
How a regular speed test works
The basic problem of any regular VPN speed test is that too many factors influence the results. Your traffic takes a much longer route from your PC to the VPN server than it does from there to the speed test server. This makes it difficult to say how fast the VPN infrastructure is.
Here’s how it works:
- The user chooses a VPN server and clicks Connect
- His request travels through multiple hops, often in various countries, until it reaches the data center with the VPN server
- The data center sends the request to the VPN server, which establishes the connection
- The VPN server sends a request to the data center with the speed test server
- The request reaches the data center with the speed test tool
- The data center transfers the request to the speed test server
- The result goes back to the user along a similar chain, measuring the speed
As you can see, there are many points where the connection can slow down, affecting speeds. But that’s just one of the issues when testing your VPN’s performance.
More problems with regular VPN speed tests
The majority of VPN speed tests give you the numbers from a single moment in time. What you see is a single speed reading from a single location during a particular hour and minute. Changing any one of the parameters will have an impact on the results. Now imagine changing all three and others that weren’t mentioned yet.
Here are just some of the reasons VPN speeds may vary:
- Time of the day/week. Each VPN has many more users than servers, which are shared between everyone connected at that time. A bigger server load means slower speeds. If the speed test was made during peak hours and peak weekdays, the numbers would probably be lower than those you might get on a Saturday night.
- Location of the VPN server. VPN servers are spread across data centers all around the world. Even if all of them were of the same quality, the internet infrastructure varies from country to country. That’s why sometimes, you may get better speeds in a country that’s thousands of miles away than in one that’s nearby.
- The number of attempts. Two speed readings could be significantly different, even if you made them one after the other. A sudden VPN server usage spike or somebody starting to download a large file on your wifi can make an industry-leading VPN seem sluggish.
- Chosen service. The internet is a network of interconnected devices constructed of various protocols. This means that connecting to a VPN server in the United Kingdom and downloading a file from a major UK website will result in different speeds than streaming Netflix from the United States. This also means speeds will vary depending on whether you’re running a speed test or trying to stream something, for example.
Hopefully, this gives you an idea about the relevance of most VPN download and upload numbers that you encounter online.
Why our VPN speed test is more accurate
In short, our speed test is better because it removes more variables from the equation than your regular speed reading. Let’s discuss each of these variables in a bit more detail.
How our speed test works
The diagram below should give you an idea about the differences in the VPNpro speed testing tool compared to any regular online speed test.
Here’s how it works, step by step:
- Our speed test connects to the internet and requests a connection with the VPN server
- The request travels through as few hops as possible because the VPN server and the speed test server are always within the same country
- The request reaches the data center with the VPN server
- The VPN server sends back the results along a similar path
Eliminating the user’s ISP from the test
For starters, we take the measurements and compare them automatically, which is not possible for a user unless you dedicate your life to VPN speed testing. By eliminating the user’s ISP from the test, we exclude the following factors:
- Interconnect links chosen by the ISP
- The infrastructure of the user’s ISP
- The user’s selected bandwidth plan
These three factors alone can influence the speed test results more than the quality of the VPN.
Measuring within the same country
Let’s say you live in Europe, and you connect to a VPN server in the USA. Then you run an online speed test and get the results. In such a scenario, the route your traffic takes from your PC to the VPN server in the US is a lot longer than the route from the VPN server to the speed test server.
Having all this in mind, we measure VPN speed within the same country. This allows us to have as few hops as possible between the data centers that host our speed test and the VPN server. Even if the user is in the US and runs a speed test in the same country, his traffic would likely still travel through more hop than a dedicated VPN speed testing tool like ours.
Maintaining the “user perspective”
Even though we try to remove as many unnecessary steps as possible from our VPN speed tests, we don’t try to get better results than a regular user could. We use a precise algorithm that takes each VPN provider’s OVPN configuration file (where available) and uses real account credentials.
At the moment, this method has one disadvantage – VPNs use different tunneling protocols. For example, Surfshark has WireGuard, while NordVPN offers its WireGuard-based NordLynx protocol. Therefore, these services are way faster than any other that uses OpenVPN.
Other VPNs that have introduced next-gen tunneling protocols include:
- VyprVPN (WireGuard)
- Private Internet Access (WireGuard)
- CyberGhost VPN (WireGuard)
- ExpressVPN (Lightway, in beta)
We hope to integrate all of them into our VPN speed test as soon as possible.
Regular testing
To ensure that we exclude any network or other technical anomalies that can influence any single measurement result, we test 1 provider per location approximately every 5 minutes. This roughly translates to 1 new reading per provider/location every hour. This allows us to average out the results and give you comparable data on how each provider performs.
It would be near impossible for the user to run the same number of tests. And even if they did, the factors mentioned above would deem the results subjective. That’s why it’s not even worth the trouble, and that’s why you should always remember to look at the tests that other users have taken with at least five pinches of salt.
Multiple servers
VPN providers usually have multiple VPN servers per country, which is why we always use more than one server for our tests – any specific server may be under more load at the time of testing, which can skew the results. Some providers offer a recommended server based on the load, distance, and other factors. We use this feature whenever possible for the best results.
Finally, VPN providers constantly upgrade their infrastructure and move servers around. For that, we have implemented sophisticated monitoring measures, which help us react to any unusual readings and provide you with the most accurate data possible.
VPN speed test FAQ
How do we measure the VPN speed?
We run regular tests for each VPN and provide the average speed for the last three months or overall. When testing, we try to remove as many factors as possible, such as location, infrastructure, and time of the day/week. We always take measurements within the same country to minimize the number of hops.
How is our test different from the others?
As you may have seen from the illustration above, there’s a significant difference in the number of steps involved in a regular and our in-house speed test. We remove the unnecessary steps and conclude only after gathering enough data, so there won’t be daily changes in the number one spot.
Why do we merge download and upload?
It’s just one of the ways to display the average VPN speed. You can also check these stats separately and make your own calculations.
Why do we test these particular VPNs?
Currently, we only test the best and most popular VPNs on the market. These are the VPNs most users are interested in – and they’re also faster than most VPN providers outside of our Top 10.
With that said, eventually, we hope to include more VPN services in our speed test.
hey please review strongvpn and hotespot shiled please
Hello, there. We have a written StrongVPN review and Hotspot Shield review
interesting vpn speed comparison and I can see that the tests are very recent. Do you conduct them automatically? How does the process work?
Greetings. Yes, the tests are done automatically. The process is explained below the test and in the FAQ.
I didn’t know before how those vpn speeds tests can be so inaccurate. Your method seems interesting but would it accurately represent what it’s like to browse as a user?
Howdy. I wouldn’t say that it accurately represents what it’s like to browse as a user since not everyone has such fast internet connections. However, the test does show the full potential of each provider, which is relevant for users with exceptional needs.
I use speedtest vpn for mac, very good results!
Oh wow TorGuard performed amazing in this vpn speed test, I really didn’t know but It still is not as good as NordVPN. I wonder how are they so good. I wanna know the secret sauce!! Haha
Greetings. I think the WireGuard tunneling protocol is an essential ingredient in the secret sauce of speed. After all, Nordlynx is based on it.
How can I test vpn connection speed independently? I would like to compare results to your ones! And maybe try out different vpn servers and countries, probably the speed could differ greatly
Hello there. When our speed test is unavailable, we usually use the Ookla speed test. We choose the fastest tunneling protocols and closest servers for optimal performance. Naturally, your results will differ as your base internet speed will not be the same.
That vpn speed test comparison graph at the top is really cool, played around with it for over 20 min haha good job
Hello, I am getting 300mbps, download and upload speed with VyprVPN for now already more than 1 week.
Can you explain to me how is it that you’re getting those really slow speeds ? thank you.
Hi there. Our test is still using the OpenVPN protocol. Meanwhile, if you’re using the WireGuard protocol, it’s not surprising that you’re getting faster speeds. We are trying to fix these issues ASAP, but they still take time.
I was using ookla speed test to check my VPN speed, and the results are very odd. Results shows that VPN which I am using is twice slower than it decleared in your comparison table. Maybe there is some sort of mistake ?
Hello, Pia. VPN speed test results will always vary since they depend on a lot of factors e.g. your ISP, baseline connection speed, device, location, server load, etc. If you look at some providers over a longer period of time you’ll see that their speeds fluctuate over time. It’s normal to get different results.
Will a VPN speed up my internet if I assume that my ISP might be slowing down the internet on purpose by downgrading the firmware of the router? I didn’t run any speed test because internet providers can detect speed tests and artificially inflate your speed to make it appear that they’re not throttling you, so it’s purely my assumption.
Greetings, Andie. Yes, there is a possibility that a VPN will get around speed throttling. You could test that out with a free VPN. However, I don’t think that ISPs have the power to inflate speed test results. At least if you’re using a speed that that isn’t affiliated with your ISP.
How much of your speed do you lose with a double VPN when you are far from the server you want to connect to. I’m talking about CactusVPN in this situation. I don’t see it in your list. Somehow I am sure, that the drop might be significant, due to double connection.
Hey, Trick. You’re right, the speed drop will be significant, especially since CactusVPN isn’t known for its speed when connecting to far-off lands. It’s difficult to say for sure, but the speed reduction could vary from 50% to 80%.
Hey! How to effectively test vpn connection speed at home? Maybe there are some apps that you could suggest? Before buying, I want to check certain providers with my internet. Your tool is great to see the full picture of each provider, but as far as I know, the speed you have, doesn’t assure, that I will have the same result.
Hi, Zhane. It’s a great idea to test several VPNs with your setup. While there will be more hops than in our speed tool, you’ll be testing those services on the same device using the same connection. This should give you some insights on how fast each VPN is. Good luck!
I use nord and get nowhere near that.
I have GB internet. Regardless of what server I use, the absolute most I can get is 230Mbps if im lucky (downloading 23 megabytes/sec). Thats even with this new nordlynx everyone is raving about that changed nothing except seem to slow down the legacy protocols for comparison.
Thats speed is essentially a quarter of what I pay for. Ive used the service for 3 years now, and it was a huge mistake. Finally, my contract is up this month and I stumbled across this site, which I suspect is just a sponsored shill site. Can anyone give raw #s of their experience with other VPN’s?
1gb down translates to 230 down on any (nord) US server for me. Changing server only affects ping. I think nord specifically caps it at this rate. Its not the technology because 3 years ago I used PIA up until Nord and was getting around 500mbps. Was going to go back to PIA but I heard they sold out to some shady shell company, gonna have to look further into it. One thing is for sure, there is no way in hell Nord is 10x faster than PIA. No way.
My advice is to trial run your vpn before you commit to some lengthy contract like I did.. it was a horrible mistake.
Dear Mr. Souper,
Thanks for the comment. I’ll start from the end of your letter by saying that we use WireGuard for NordVPN and OpenVPN for PIA testing. This automatically means much better results for the former.
Additionally, we have never seen the numbers that we got from all WireGuard tests before. However, Mbps that you see on our speed test tool page shouldn’t be compared with any “local test.” As we tried to explain, our method is totally different and you will often get lower upload and download numbers with all these additional hops.
With that said, this tools gives you a chance to compare multiple providers. But I don’t recommend doing that with two different protocols involved. Good luck!
This speed test revealed interesting facts about certain providers. I was always pretty sure that expressvpn is faster than mullvad. however, as far as I can see mullvad speed for the last 30 days from us is very high. Interesting and shocking actually.
Hello, VPNpro! I have never tried watching Netflix with a vpn before. Is privatevpn speed enough for watching it without buffering or for streaming it’s a bit slow and I need to consider other vpn provider?
Hi, Chris! Although PrivateVPN does not have the WireGuard tunneling protocol, it’s still quite fast and should be able to unblock Netflix without any lags or stuttering. However, if you want to be certain that there should be no lags, consider a provider with the WireGuard tunneling protocol. Something like NordVPN or Surfshark.
it seems that vyprvpn speed is pretty stable and below the normal. can you comment why the speed radically differs from other providers? maybe there is some mistake in measurements? i’m planning on buying this vpn, so it is very important
Hi, Robin. In our test, we’ve tried to remove as many speed-changing factors as possible. Doing that for all tested VPNs allowed us to get the precise results. However, we don’t do tests for all tunneling protocols, because configuring some can be troublesome. In case of VyprVPN, we use OpenVPN, even though they have been offering WireGuard for quite a while. Therefore, I’d suggest you to test it yourself during the money-back guarantee period and see if it’s fast enough for you.
i saw one youtuber was measuring surfshark speed simply by downloading a torrent file and using a stopwatch. needless to say, that’s 100% inaccurate. how do you take those measurements? do you take different protocols, location accuracy into consideration?
Hi, suren. This method might be good when comparing Surfshark with other VPNs. However, testing this service using different protocols would be pretty inaccurate. That’s because the users hardware and distance to the server play a major role.
In our case, we remove as much variables as we can. We start up a server in the same country and use as few hops as possible. Moreover, we use different servers as get new results every hour.
There is an expressvpn speed test in the vpn app. However, I’ve noticed, that there is a slight difference in speed. The built-in shows a faster speed(150mbs) on the USA server, than it’s showed here (126mbs). Is it normal for such a difference or built in speed test is simply not that accurate?
Hi, Dominic. Talking about which speed test is more accurate. Obviously, ExpressVPN uses a difference approach than we do. However, I expect our speed test to show better results because it removes the user factor. Also, we are yet to implement the Lightway protocol. So if you used this or IKEv2, that would explain the difference.