Industry-Standard Speed Test
Speedtest by Ookla — free, no account required
Test against thousands of global servers · Compare VPN servers directly
Marketing claims about VPN speed are nearly useless. Here's how to actually measure what you'll get on your own connection.
"Blazing fast speeds" and "minimal slowdown" are phrases every VPN provider uses, and they tell you almost nothing about what you'll actually experience on your specific connection, in your specific location, on the specific server you'll actually use day to day.
VPN speed depends on too many variables — your baseline connection, distance to the server, server load at that moment, the protocol used, and even time of day — for any single marketing number to mean much. The only number that actually matters is the one you measure yourself, on your own setup.
Three ways to measure, each telling you something slightly different.
| Method | What It Measures | Reliability | Best For |
|---|---|---|---|
| Manual Speed Test | Raw throughput to nearest test server | High, but server choice matters | Quick comparison across VPN servers |
| Built-In App Test | Provider-selected "optimal" server | Lower, provider-biased selection | Quick check within the VPN app |
| Real-World Transfer Test | Actual sustained transfer speed | Highest for real usage | Validating specific use cases |
The table tells you what's measured. This is when each test actually earns its keep.
A few extra minutes here saves you from judging a VPN on a misleading single result.
The protocol you connect with matters as much as the server you connect to.
| Protocol | Typical Speed Loss vs No VPN | Notes |
|---|---|---|
| WireGuard | 5-15% | Modern, fast, increasingly the default for most providers |
| OpenVPN (UDP) | 15-25% | Older but well-tested, still widely supported |
| OpenVPN (TCP) | 25-40%+ | More reliable on restrictive networks, noticeably slower |
For a quick gut-check, a manual speed test on the specific server you'll actually use is enough. For a genuinely important decision — like committing to an annual plan — run the slower real-world transfer test instead.
Either way, never trust a marketing speed claim without your own measurement on your own connection — too many variables differ for someone else's number to apply to you.
View Our Full VPN Speed RankingsServer load fluctuates with the number of simultaneous users, and that load tends to peak during evening hours in most regions — testing at different times reveals this variation directly.
Generally yes — more physical distance and network hops between you and the server tend to increase latency and reduce throughput, though server load and routing quality can outweigh distance in some cases.
In the large majority of cases yes, due to its more efficient design, though specific network conditions or provider implementation quality can occasionally narrow or reverse the gap.
Some ISPs do throttle traffic patterns associated with VPN or encrypted connections, particularly during network congestion, though this varies significantly by provider and region.
A well-optimized VPN on a modern protocol like WireGuard typically loses somewhere in the 5-15% range under good conditions — meaningfully more than that on a server you'd expect to perform well is worth investigating.
Yes, if you use the VPN on mobile regularly — mobile network conditions differ enough from home wifi that a separate test gives a more accurate picture of real mobile performance.
Often yes — a browser extension typically only encrypts browser traffic, while a full app encrypts your device's entire connection, which can produce different speed characteristics depending on what's being tested.
This guide covers the testing methodology — our category page covers current independently-verified speed benchmarks across every VPN we've reviewed.