Independent browser-based measurement

Measure the path your network actually takes.Choose DNS with evidence.

DNS Benchmark Pro runs validated DNS-over-HTTPS queries from your browser to public resolvers, then reports median response time, variability and verification rates in one defensible comparison.

Run the browser benchmark Review the methodology

No installation No account Results stay in this tab

Measurement engine ready Browser session

…Public resolvers

MedianPrimary ranking

MADVariability shown

Validated response profileIllustrative interface

Only readable, parseable responses that match the original DNS question enter the ranking.

Verified measurementLocal browser output

Validated responses only

Unreadable or malformed replies are excluded instead of being mistaken for useful latency results.

Median-led ranking

One stalled request cannot distort the headline figure the way it can distort a simple average.

Variability exposed

Spread, sample count and success rate remain visible beside every reported median.

Failures identified

Timeouts, refusals and browser limitations are named and kept outside the ranking.

Measurement integrity

A DNS result should be explainable, not merely fast.

Generic speed tests measure bandwidth after a connection exists. DNS Benchmark Pro isolates an earlier step: how long this browser takes to receive a verifiable response from each participating DNS-over-HTTPS endpoint on the network path you are using now.

Interleaved, shuffled rounds

Providers are not measured in one long block, reducing the chance that temporary conditions favor a single resolver.

Controlled test questions

Each query uses a controlled name under dnsbenchmarkpro.com and is checked against the response that was expected.

Uncertainty stays visible

When two results are closer than the test can reliably distinguish, the report labels them as tied.

Run the benchmark

Before you start — what this sends, and to whom

Your browser will make DNS-over-HTTPS requests directly to … third-party resolvers. Each provider you contact can see your IP address and the test name being queried, and may log it under its own privacy policy.
The queries use a controlled test name under dnsbenchmarkpro.com. Nothing about your browsing, device or identity is placed in a query name.
This tool does not request your location. No geolocation prompt, no coordinates, no reverse-geocoding service.
Results stay in this browser tab. Nothing is uploaded to our servers, and nothing is stored unless you export it yourself.

Provider privacy policies are linked in the resolver comparison. Our own data flow is documented in the privacy policy.

Test depth

Quick5 rounds · about 20–40 seconds Thorough11 rounds · a few minutes, tighter estimates

What to measure

A cached lookup is what you experience most of the time. An uncached lookup is an upper bound, not a typical result.

Results

Each dot in Every measurement is one validated response, on a scale shared by every row; the upright bar is the median. The scale is logarithmic, because the difference between 20 ms and 40 ms matters far more than the difference between 300 ms and 320 ms. Dots stacked together mean a resolver answered in much the same time each round. Dots spread across the track mean it did not — and that is usually more important than which median happens to be lowest.

Resolvers ranked by median browser-to-DoH latency, fastest first. Each row also reports the spread, the number of validated responses and the resolver addresses.
#	Resolver	Every measurement	Median	Spread	Verified	IPv4 addresses

Excluded from the ranking (0)

These resolvers did not return a response this browser could read and verify, so they carry no latency figure. In most cases this is a browser limitation, not a fault in the resolver: a DNS-over-HTTPS endpoint that does not send cross-origin headers simply cannot be measured from a web page. Their configuration addresses are still listed in the resolver comparison.

Resolvers excluded from the ranking and the reason for each
Resolver	Outcomes	Detail

Before you change your DNS settings

A low figure in the table above is a reason to try a resolver, not proof it will be faster for your whole system. What you measured here is an HTTPS request from this browser to a provider’s DNS-over-HTTPS endpoint. What you would configure in Windows, macOS or your router is usually that provider’s classic DNS service on port 53 — a different endpoint, transport and cache. Related, frequently correlated, but not the same measurement.

Write down what you have now

Before changing anything, record your current settings so you can put them back. On most systems the original setting is “Obtain DNS server address automatically” (DHCP) — returning to that undoes the change completely.

Where it can go wrong

On a corporate or school network, a custom resolver often breaks access to internal hostnames, intranet sites and VPN resources that rely on split DNS.
Public Wi-Fi captive portals may fail to load until you revert to the network’s own DNS.
Changing resolver changes what is filtered. A family or malware-filtering resolver will block things your current one allows, and an unfiltered one will stop blocking things you may rely on.
Your chosen resolver sees every domain you look up. Changing DNS moves that visibility from your ISP to that provider — it does not remove it.

Step-by-step guides

Each guide covers IPv4 and IPv6 separately, distinguishes classic DNS from DNS-over-HTTPS, and ends with a check that the change worked plus the exact steps to roll it back.