DNS and WebRTC leak tests
- Screen recording of the DNS test
- BrowserLeaks WebRTC result
- What a clean result does and does not prove
Hands-on tests · September 2026
CyberGhost tested: leak checks first, then speed and streaming
We put CyberGhost through DNS and WebRTC leak checks, a two-route speed test and a streaming playback, and filmed all of it. Here is the footage, the numbers and the limits.
- 66% of baseline download kept (Europe)
- DNS and WebRTC: no leak seen
- BBC iPlayer: played
At a glance
The headline numbers for CyberGhost, each one explained on its own page.
66%download kept, nearest Europe194 of 295 Mbps
42 msping, nearest Europebaseline 4 ms
#9of 11 for download, EuropeUS route: #1
2/2leak checks cleanDNS and WebRTC
What we tested
One page per test, each with the raw captures and what they do and do not show.
Speed tests from two routes
- Strasbourg: 194 Mbps down, 42 ms
- New York: 210 Mbps down, 126 ms
- League table of all 11 VPNs
BBC iPlayer playback
- BBC iPlayer home page
- Which other VPNs we captured
How CyberGhost compares with ten other VPNs
Same line, a fresh baseline for each VPN, ranked by nearest-European download.
| # | VPN | Europe, Mbps | Ping | US, Mbps | Ping | EU kept |
|---|---|---|---|---|---|---|
| 1 | Proton VPN | 262 | 42 ms | 126 | 174 ms | 89% |
| 2 | hide.me | 222 | 47 ms | 200 | 163 ms | 74% |
| 3 | Windscribe | 222 | 44 ms | 203 | 108 ms | 75% |
| 4 | NordVPN | 210 | 31 ms | 200 | 111 ms | 72% |
| 5 | Surfshark | 210 | 79 ms | 203 | 112 ms | 71% |
| 6 | LetsVPN | 207 | 42 ms | 202 | 181 ms | 70% |
| 7 | PrivadoVPN | 200 | 42 ms | 160 | 132 ms | 68% |
| 8 | PureVPN | 200 | 39 ms | 209 | 177 ms | 68% |
| 9 | CyberGhost this site | 194 | 42 ms | 210 | 126 ms | 66% |
| 10 | ExpressVPN | 181 | 35 ms | 160 | 112 ms | 62% |
| 11 | AdGuard VPN | 136 | 46 ms | 200 | 171 ms | 46% |
Proton VPN was tested on its Stealth protocol and Windscribe on WireGuard 443; the other VPNs were left on their app defaults.
How we tested
The same short procedure for all 11 VPNs in the September 2026 session.
- Baseline first. Every VPN started with a Speedtest result with the VPN off, on the same line: 293–298 Mbps down and up, 4–5 ms ping.
- Two routes. The nearest European server and one long-distance US server, with the VPN app on its defaults except where we name the protocol.
- Leak checks. A DNS leak test and a BrowserLeaks WebRTC test, recorded while connected.
- Streaming. Playback of a real title on BBC iPlayer and/or Netflix, with the VPN app visible on screen.
- Raw captures only. Screenshots and screen recordings, re-encoded for the web. The only edits are masks over account identifiers and the tester's real IP address.
What these results do not show
Read this before you quote a number about CyberGhost.
Limits of this test
- One run per route, on one day, on one residential line of about 300 Mbps. Server load, time of day and your own ISP will change your numbers.
- A streaming clip that played during our test is not a promise. Services update their blocklists; we record what happened on the day.
- Services and checks that are not listed here (Disney+, Hulu, Max, Amazon Prime Video, torrenting and P2P, kill switch, IPv6, mobile apps, reachability from China or Russia) are not covered by this evidence. A missing result means we did not capture it, not that it failed.
Questions people ask
Short answers, based on our own captures and the data on this site.
Is CyberGhost fast?
It kept 66% of baseline download on the nearest-European route and 71% on the long-distance US route in our single-run test. Details and the full league table are on the speed page.
Who ran these tests?
We did, by hand, in September 2026, on a single residential connection. The raw screenshots and recordings are published as captured, with account identifiers masked.
How are these tests different from the user reviews on this site?
The user reviews are what thousands of customers wrote over months; the tests are one controlled session of ours. They answer different questions, so we show both.
Can I reproduce these results?
Yes: run a Speedtest with the VPN off, then on, on the same line, and compare. Expect different absolute numbers; the gap between baseline and VPN is what carries over.
Which tests did you not run?
Anything not on these pages: other streaming services, torrenting, kill switch, IPv6, mobile apps and reachability from China or Russia. See the limits box above.
Compare VPNs side by side
See current plans, prices and refund terms for the VPNs we recommend.