Artifact Security’s VPN performance report 2026: a closer look at the results

To demonstrate their commitment to security and privacy, certain VPN providers have undergone independent security audits. These audits are carried out by unaffiliated third parties and examine specific areas of the VPN's technology, infrastructure, and policies, with the scope varying from one provider to another.

Claims such as 'no logs' or 'highly secure' are nothing but pure marketing unless they have been independently verified. Rather than asking customers to accept the VPN provider's word on the matter, audits provide them with reliable evidence.

The purpose of this article is to examine the Artifact Security report 2026, to determine how the audit was conducted, what the researchers discovered, and the test's shortcomings.

About Artifact Security and the VPN performance report

Artifact Security is an independent cybersecurity testing lab based in the UK, with over 15 years of testing experience. The company has examined over 100 products.

The VPN performance report evaluated 5 VPN providers across a range of criteria, including connection speeds, data leaks, website accessibility, and VPN features. This report represents Round 1 of an ongoing testing program, with additional VPN platforms scheduled for evaluation in Round 2.

What Artifact Security tested

Artifact Security tested 5 top VPN providers: NordVPN, ExpressVPN, Proton VPN, Surfshark, and Mullvad VPN. The testing looked at both everyday performance and core security features to give a balanced picture of how each service performs in practice.

The testing covered several areas:

• Connection speed. Measured VPN speeds across 5 origin regions and 55 different highways.
• Connection time. Tested how quickly each VPN established a connection.
• Leak testing. Checked whether VPNs caused IP, DNS, IPv6, or WebRTC leaks.
• Website accessibility. Tested whether the VPNs caused CAPTCHAs or site blocks.
• Core features. Tested main features like auto-connect, launch on boot, kill switch, and split tunneling.
• Additional features. Validated whether other features advertised by the providers actually work.

To keep the comparison fair and consistent, Artifact Security used a tiered scoring system and gave each VPN an overall score based on its results across the different tests.

Artifact Security’s testing follows the VPN testing guidelines published by AMTSO, which provide a consistent standard for testing and comparing VPN services. All tests were carried out in controlled, repeatable environments, and the VPN products were purchased commercially. This helps ensure that the results are independent, consistent, and can be reproduced.

Which of the VPNs came out on top?

The Artifact Security tests covered several areas, from speed and connection performance to leak protection, site blocks, and core VPN features. Below is a table with the aggregated VPN results.

Product	Overall score	Award
NordVPN	91	Platinum
ExpressVPN	84	Gold
Proton VPN	83	Gold
Surfshark	81	Gold
Mullvad VPN	70	Silver

Below, I dive deeper into the results for each category to see how each VPN performed.

Comparing VPN speed

The report includes a wide range of speed tests. What’s nice is that Artifact Security doesn’t simply test the connection to a single location. Instead, it tested each VPN from 5 locations and connected to 11 VPN servers to get a better picture of how each provider performs in practice. Here’s a quick summary of the key results:

• Download speed. NordVPN delivered the strongest overall download performance, although it wasn’t the fastest in every individual location. Surfshark led in two of the five regional tests, Frankfurt and Singapore, showing that performance can vary depending on the test location.
• Android speeds. NordVPN showed the best download results here. It also recorded the highest median upload speed, at 223.3Mbps, outperforming the second-best result, ExpressVPN, by about 35%.
• Latency and consistency. Proton VPN stood out here, recording the lowest latency and jitter among the tested providers. This suggests it offered the most consistent connection performance in the tests.
• Connection performance. NordVPN had the fastest connection time in the test, while ExpressVPN was the clear outlier, taking substantially longer to connect.

VPN leak protection

The test also simulated various real-life situations like regular VPN connection, using split tunneling, and stress cases.

All 5 VPNs passed the tested IP, DNS, IPv6, and WebRTC leak checks. This means these providers can protect your connection, so you don’t risk being exposed while using a VPN that’s supposed to protect you.

CAPTCHAs, blocks, and the browsing experience

VPNs can sometimes trigger CAPTCHAs or other forms of site blocking. That can quickly become frustrating, especially when it gets in the way of browsing or makes you think twice about using a VPN.

According to the Artifact Security report, none of the 5 VPNs caused hard blocks, and 98% of the tested site visits were completed without a challenge.

There were still some differences between providers. Proton VPN triggered the most CAPTCHAs and other site changes, while Mullvad had the fewest issues. So, although all 5 performed well overall, some may provide a smoother browsing experience than others.

Core VPN features

Features such as auto-connect, launch on boot, kill switch, and split tunneling have become standard expectations for leading VPN providers. Artifact Security tested these core features as part of its evaluation.

All 5 providers passed the Windows feature checks included in the test. That means the VPNs covered by the report all delivered the core functionality tested, rather than falling short on basic features that users increasingly expect from a modern VPN.

The value of independent VPN testing

There’s a good reason to pay attention to independent research reports when choosing a VPN. When you’re shopping for one, a lot of what you see comes down to marketing claims. Some of those claims may sound impressive but, when you look closer, don’t actually tell you much.

Independent labs, on the other hand, use their own testing methodologies to evaluate VPN providers more objectively. They can put those marketing claims to the test and show whether they actually hold up. For users, that means getting a better idea of how a VPN performs in real-world conditions, rather than relying on ideal scenarios presented by the product developers.

It’s also worth remembering that VPNs have different strengths and weaknesses. One might offer excellent speeds but lack some of the extra privacy features you’re looking for. Another might not top the speed charts but could do a better job of avoiding website blocks, for example.

That’s why the overall ranking isn’t always the most important thing. If you know what you actually need a VPN for, independent test results can help you look beyond the headline score and choose the provider that fits your needs best.

Where the report falls short

As with any test, there are some limitations to keep in mind. The Artifact Security report is useful, but there are a few important gaps worth highlighting:

• Only 5 VPNs were tested. That’s enough to get a useful snapshot, but it’s a relatively small sample compared with the number of VPN providers available. The results, therefore, shouldn’t be taken as representative of the VPN market as a whole.
• The testing focuses heavily on performance. The report does a good job of looking at things like speed and performance, and it touches on some additional VPN features, but it doesn’t cover many of the factors that matter when evaluating a VPN’s overall security and trustworthiness. For example, there’s little or nothing on logging practices, jurisdiction, transparency, past security incidents, server infrastructure security, source-code security, account security, or how providers respond to government requests. In other words, the test can tell you a fair amount about how well a VPN performs, but it can’t give you a comprehensive picture of how secure or trustworthy the provider’s wider infrastructure and practices are.
• The testing is mainly focused on Windows, with some Android testing. A VPN that works well on one platform doesn’t necessarily perform or behave the same way on another. VPN apps can use different protocols, features, and security implementations depending on the operating system, so results from Windows shouldn’t automatically be assumed to apply to macOS, iOS, Linux, or other platforms.

How to evaluate independent test results

Independent VPN testing can be useful for comparing providers, but the headline score rarely tells the whole story. To understand what a report actually says about a VPN, start by looking at how the test was conducted and what was (and wasn’t) measured.

• Start with the methodology, not the results. Before even looking at scores, check how the VPNs were tested, including the test setup, locations, devices, protocols, servers, and conditions. A result is only meaningful in the context of the methodology used to produce it.
• Understand the scoring. Find out what each score or rating represents and how it was calculated. A high overall score may combine several very different factors, so it’s important to know which areas carry the most weight.
• Look at results per category, not aggregated results. Break the report down into individual categories such as speed, privacy, security, streaming, or usability. This can reveal strengths and weaknesses that an overall score hides.
• Look at the difference in results. Consider how large the gap is between the results, rather than focusing only on which provider ranks first. You may not notice a small difference of 5Mbps speed in use, so don’t overvalue small differences.
• See what’s not tested. No test can cover every aspect of a VPN, so check which features, use cases, or privacy claims were outside the scope of the assessment. An excellent result does not necessarily mean the VPN performs well in areas the test didn’t examine.
• Check the test date. VPN performance, apps, protocols, server networks, and privacy practices can change over time. For VPNs, I recommend relying on reports no older than 2 years.

Conclusion

The Artifact Security VPN report gives us valuable insight into how the 5 major VPN providers perform and compare in real-world conditions. NordVPN came out on top, earning the highest overall score and a Platinum award, while the results also showed that each provider has its own strengths and weaknesses.

Independent lab testing is important because it puts VPN providers’ claims to the test under consistent, repeatable conditions. It gives users objective evidence to look beyond marketing claims and make a more informed choice.