How the DNS speed test works
The script sends standard DNS-over-HTTPS requests to your chosen public resolvers. It warms up every provider/hostname pair, then interleaves repeated timed queries with at most two in flight. Timing starts immediately before a request and ends after the DNS response body is read. Warm-up samples are excluded from the statistics. A and AAAA select the DNS record type; AAAA does not force the HTTPS connection to use IPv6.
Each timed response is checked as a DNS message, including its question, record type, and record boundaries. Valid negative answers such as NXDOMAIN are timed too; expand the provider to see them. Filtering policies can produce different answers. The ranking is a latency comparison, not a claim that one provider offers better privacy or filtering.
What a fast result means
A lower median indicates a quicker typical DoH exchange in this run. The minimum and maximum show the observed range; the average is more sensitive to slow outliers. Repeat at different times, keep VPN state consistent, and compare reliability before changing settings. Browser scheduling, DNS and HTTP caching, resolver routing, and other traffic affect the measurements.
A failed request can reflect CORS policy, an extension, a VPN, filtering, or a network problem. It does not by itself prove that a DNS service is down. Stopping, leaving the page, hiding the tab, or reaching the run deadline cancels outstanding requests; interrupted output remains labeled partial.
Use the result thoughtfully
Start with how DNS works, compare DoH and DoT, and understand DNS caching and TTL. DNS response time affects lookups; it does not increase the capacity of an established transfer. See bandwidth versus latency for that distinction.