DNS Speed Test — Find Your Fastest DNS in 30 Seconds

We test each resolver using DNS-over-HTTPS (DoH) from your browser. For every resolver, we run cached lookups (domains likely in the resolver cache) and uncached lookups (unique subdomains that force fresh resolution). We report median latency (p50), tail latency (p95), jitter (standard deviation), and availability. The composite score weights all four metrics.

Average latency hides the spikes that make browsing feel slow. A resolver averaging 20ms with occasional 400ms spikes feels worse than one averaging 35ms with rock-solid consistency. Jitter measures that consistency — it is the standard deviation across all test queries. Lower jitter means a more predictable, smoother browsing experience.

Cached lookups test how fast a resolver returns domains it already knows about — this is the best-case scenario. Uncached lookups force the resolver to query upstream authoritative servers from scratch, which is what happens when you visit a site for the first time. Most tools only test cached performance. We test both because uncached speed is what you actually feel during normal browsing.

Results reflect your real-world experience because tests run from your browser, on your network, from your location. We use high-resolution timing for sub-millisecond accuracy. The uncached test uses best-effort cold-cache methodology — some resolvers may have recently cached our test domains from other users, which can slightly understate the cached/uncached difference. We document this limitation openly.

Most people never change their DNS and use whatever their ISP assigned. We detect your current DNS provider when possible, then show which public DoH resolver tested fastest from your network. Your ISP DNS is shown as context, but it is not scored directly unless it also exposes a public DoH endpoint you add manually.

No. Resolver speed queries go directly from your browser to the resolver being tested using encrypted DNS-over-HTTPS (DoH). We never see, log, or proxy that DNS traffic. To identify your current ISP DNS, the page also contacts our DNS detection service (bigdig.energy) and makes a few browser probes to generated test subdomains before the speed test runs.

Different use cases care about different metrics. Gaming prioritizes low jitter and tail latency (p95) because spikes cause lag. Streaming prioritizes availability and median speed. General balances everyday browsing signals, and Power User applies equal weight to p50, p95, jitter, and availability.

Yes. Use the custom resolver field to add any DNS-over-HTTPS (DoH) endpoint. Enter the full URL (e.g., https://your-resolver.example/dns-query) and it will be included in the next test run.

DNS performance varies with network conditions, resolver load, cache state, and routing changes. This is normal and expected. We store your previous results locally so you can compare over time — the trend matters more than any single test.

After running the test, expand the setup instructions below the results for step-by-step guides for Windows, macOS, Linux, iOS, and Android. The easiest method for most people is changing DNS in your router settings, which applies to every device on your network.

Usually yes — Cloudflare (1.1.1.1) is faster than Google DNS (8.8.8.8) for most users because Cloudflare operates a larger edge network with more points of presence globally. However, the gap is location-dependent. In some regions, Google's infrastructure routes queries faster. Run the test above to see which is actually faster from your location — the answer varies by city and ISP.

For cached lookups, 10–30ms is excellent, 30–80ms is acceptable. For uncached lookups, 30–80ms is good, 80–150ms is typical. If your cached latency exceeds 100ms, the resolver's nearest point of presence is likely far from you — try a geographically closer provider. If uncached latency exceeds 200ms, the resolver has slow upstream recursion or routing issues. Jitter matters as much as the raw number: a resolver at 40ms with ±2ms jitter feels faster than one at 25ms with ±60ms jitter.

DNS does not affect file download speeds or streaming throughput — once a domain resolves to an IP address, your data transfer speed is unrelated to DNS. What DNS does affect is how quickly pages start loading. Every domain lookup must complete before a connection can open, and a modern page can trigger 50–100 lookups. Slow DNS makes pages feel unresponsive even on a fast connection because the browser stalls waiting for addresses before it can begin fetching resources.

DNS-over-HTTPS (DoH) sends your DNS queries inside standard HTTPS connections on port 443, making them indistinguishable from regular web traffic. This encrypts your queries against eavesdropping, prevents tampering by ISPs or networks, and hides your DNS lookups from anyone monitoring your connection. Traditional DNS uses plain UDP on port 53, which is visible to any network observer. DoH is supported by Windows 11, macOS, iOS, Android, Firefox, Chrome, and all resolvers tested here.

Yes. Changing DNS is one of the safest network settings you can modify — it is instantly reversible and does not affect your connection speed, VPN, or firewall. Choose a resolver that publishes a clear privacy policy: Cloudflare, Google, Quad9, and Control D all publish annual audits and explicit data-retention policies. Avoid random or unknown DoH endpoints, which could log or manipulate your queries. If you are on a work device, check whether your employer requires the company DNS resolver for internal services before switching.