Skip to content
trending

Benchmarks

1.1.1.1 vs 8.8.8.8: We Benchmarked 10 Public and ISP DNS Resolvers

We tested 10 DNS resolvers head-to-head. Cloudflare 1.1.1.1 matched local ISP speeds at ~20 ms, while Google 8.8.8.8 proved more than twice as slow.

By 5 min read
1.1.1.1 vs 8.8.8.8: We Benchmarked 10 Public and ISP DNS Resolvers

👁 128 reads

We measured the response times of 10 DNS resolvers from a broadband connection: Cloudflare 1.1.1.1 and the local ISP resolver finished nearly neck and neck (median 19.6-21.3 ms), Quad9 and Netia were right behind them, while Google 8.8.8.8 took more than twice as long (45-48 ms). On “cold” queries - when the resolver must perform full iterative lookups from scratch - Cloudflare’s lead over Google widened even further. Read more about the provider in our Cloudflare guide.

Benchmark run on October 5, 2026, rounds at 9:02 AM and 9:25 AM

  • Connection: Orange Polska (AS5617), Windows 11.
  • Methodology: 2 rounds × 50 domains × 2 query types per resolver, UDP protocol, 2 s timeout.
  • Fastest: 1.1.1.1, 1.0.0.1, and both ISP servers - median 19.6-22.1 ms for popular domains.
  • Slowest among responding: AdGuard DNS - median 53-55 ms and 3 failed queries in each round.
  • Did not respond externally: Tested DNS addresses belonging to Play, T-Mobile, Vectra, and one Netia address.

Why DNS Speed Matters

Every web visit begins with a question to a DNS resolver: “what is the IP address for this domain?”. If the resolver responds in 20 ms, you will never notice. If it takes 80 ms and a web page pulls assets across 15 separate domains, those lookups add up to noticeable latency, especially on your first visit. By default, your connection uses your internet service provider’s DNS. The question searched thousands of times every month - “Cloudflare DNS vs Google DNS” - comes down to whether switching is truly worth the effort.

Our Testing Methodology

We wrote a Python script utilizing the dnspython library to query each target resolver directly, entirely bypassing the local operating system cache:

  • 50 popular domains - top regional and global destinations (including youtube.com, chatgpt.com, onet.pl, allegro.pl, and others). Resolvers usually hold these active in their cache, making this a test of raw network routing distance to the server.
  • 50 “cold” queries - a randomized, non-existent subdomain for each domain (generated uniquely for every resolver). Because nobody has requested these hostnames before, the resolver must walk the entire DNS hierarchy from the root servers down. This tests resolver engine performance.
  • Resolver order was randomized on every single domain query to prevent sequential bias.
  • Two separate rounds conducted roughly 20 minutes apart.

Median response time - popular domains, round 1 (ms, lower = better)

Cloudflare 1.1.1.119.6 ms
Orange 194.204.159.121.0 ms
Orange 194.204.152.3421.1 ms
Quad9 9.9.9.922.0 ms
Cloudflare 1.0.0.122.1 ms
Netia 62.233.128.1725.0 ms
Google 8.8.4.439.3 ms
Google 8.8.8.845.3 ms
OpenDNS47.6 ms
AdGuard DNS53.3 ms
Server Median R1 Median R2 90th Percentile R1 / R2 Errors
Cloudflare 1.1.1.1 19.6 ms 21.3 ms 32.3 / 34.0 ms 0
Cloudflare 1.0.0.1 22.1 ms 20.7 ms 29.4 / 32.9 ms 0
Orange 194.204.159.1 21.0 ms 21.2 ms 42.0 / 35.8 ms 0
Orange 194.204.152.34 21.1 ms 20.9 ms 41.2 / 41.4 ms 0
Quad9 9.9.9.9 22.0 ms 24.1 ms 57.0 / 48.6 ms 0
Netia 62.233.128.17 25.0 ms 25.8 ms 40.6 / 40.1 ms 0
Google 8.8.4.4 39.3 ms 40.4 ms 51.3 / 59.0 ms 0
Google 8.8.8.8 45.3 ms 48.1 ms 73.6 / 56.6 ms 0
OpenDNS 208.67.222.222 47.6 ms 49.2 ms 68.2 / 66.6 ms 0
AdGuard 94.140.14.14 53.3 ms 55.4 ms 112.2 / 100.2 ms 3 / 3 (out of 100)

Benchmark Results: Cold Cache Queries

Here is how quickly each resolver resolves a domain name that is entirely absent from its cache:

Median response time - cold queries, round 1 (ms)

Cloudflare 1.0.0.126.2 ms
Cloudflare 1.1.1.127.0 ms
Netia 62.233.128.1729.8 ms
Orange 194.204.159.129.9 ms
Orange 194.204.152.3432.8 ms
Quad9 9.9.9.942.8 ms
OpenDNS53.2 ms
Google 8.8.4.456.9 ms
Google 8.8.8.863.7 ms
AdGuard DNS74.9 ms

Round 2 showed a very similar distribution: 1.0.0.1 recorded 28.1 ms, Orange 30.8 and 35.1 ms, 1.1.1.1 recorded 32.6 ms, Netia 35.1 ms, Quad9 39.6 ms, and Google 59.4 and 62.5 ms. The worst single-query outliers reached up to 1.6 s for AdGuard, 0.4 s for Google, and 0.34 s for 1.1.1.1.

Key Takeaways

  1. Cloudflare and local ISP DNS are a statistical tie. The 1-2 ms differences fall within normal measurement noise. Remember that testing occurred directly on the Orange network - an ISP’s DNS enjoys a home-field advantage inside its own infrastructure. For users connecting from different networks, that ISP’s resolvers would perform much worse.
  2. Google 8.8.8.8 is over twice as slow as Cloudflare for popular domains and 30-35 ms slower on cold lookups from this location. If you configured it simply because someone on a forum suggested 8.8.8.8 years ago, migrating to 1.1.1.1 offers an immediate, tangible improvement.
  3. Quad9 is a compelling security choice - it provides built-in malware blocking while remaining competitive on popular domains, though cold queries incur a modest speed penalty.
  4. AdGuard DNS trades speed for ad-blocking - it came in slowest overall and was the only resolver to experience request timeouts.
  5. Real-world gains are subtle. Saving 20-25 ms on an initial lookup matters, but subsequent requests are served out of your browser or local OS cache anyway. Do not expect faster download speeds - this improves snappy domain negotiation, not overall bandwidth.

How to Change Your DNS to 1.1.1.1

Windows 11

  1. Navigate to Settings → Network & internet → Ethernet or Wi-Fi → connection properties.
  2. Next to “DNS server assignment”, click Edit and switch the dropdown to Manual.
  3. Toggle IPv4 to On, enter 1.1.1.1 (Preferred) and 1.0.0.1 (Alternate). Set DNS encryption to “Encrypted only (DNS over HTTPS)”.
  4. For IPv6: use 2606:4700:4700::1111 and 2606:4700:4700::1001.

Android

  1. Open Settings → Network & internet → Private DNS.
  2. Select “Private DNS provider hostname” and type one.one.one.one.

iPhone

iOS lacks a universal native setting for cellular DNS. The cleanest path is downloading the official 1.1.1.1 app and toggling the connection to “1.1.1.1” mode (DNS only, without WARP routing).

Router

Setting custom DNS servers directly on your router protects every device on your local network automatically. Look for the WAN or DHCP configuration section inside your admin interface and locate the “Primary DNS” and “Secondary DNS” entries. If your ISP-provided gateway locks these fields down, configure DNS on individual devices instead.

Filtered DNS Versions

  • 1.1.1.2 and 1.0.0.2 - automatically block recognized malware domains.
  • 1.1.1.3 and 1.0.0.3 - block both malware and adult content (useful for parental controls).

Privacy and Security

Your DNS resolver sees every single domain name you look up. Cloudflare states that it does not retain client IP addresses permanently and purges query logs within 25 hours - commitments audited by independent third-party assessors. Your ISP can still monitor network traffic destinations by IP address, so changing DNS does not make you anonymous, but enabled DNS-over-HTTPS (DoH) keeps the plaintext lookup queries away from snooping ISP monitors. Since September 10, 2026, 1.1.1.1 also supports post-quantum DNSSEC signatures using the ML-DSA-44 algorithm.

Test Limitations

  • Single ISP connection (Orange), single European location, two rounds during one morning. Results will vary depending on your geographic region and peering.
  • Third-party ISP resolvers (such as Play, T-Mobile, or Vectra) generally restrict access to their own subscribers. Publicly cited forum addresses - 151.80.103.98, 213.158.199.1, 31.11.202.254, and 62.233.233.233 - refused all external inbound queries.
  • Measurements evaluated plain DNS over UDP. Encrypted protocols (DoH, DoT) introduce a few milliseconds of initial TLS handshake overhead.
Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →

Sources

Kinetyra

// read next

More on this from QUANTSLANT

Piotr Olszewski

Editor and publisher · Quantslant

I edit and publish Quantslant. AI tools help with research, drafts and translation; choosing topics, running the tests, checking facts and the decision to publish are mine, and so is the responsibility for what appears here.

read nextEvilTokens: How AI Helped Hijack Over 12,000 Mailboxesstate of aiEvilTokens: How AI Helped Hijack Over 12,000 MailboxesMicrosoft uncovered EvilTokens, an operation using AI to sift through hijacked email in over 10,000 organizations. Here is what security teams must learn.

Most read

  1. 1Best AI Video Generator Apps for Your Phone (2026)
  2. 2Claude Opus 5.5 Motion Graphics: 60s Showreel from One Prompt
  3. 3Wan 2.2 on RTX 4090: timings, VRAM and four test clips
// site map

This page in other languages5

Quantslant in numbers 1,427posts in archive26news from the last 7 days18video models in the observatory
🎬 AI Video Benchmarks: Higgsfield DoP vs OpenAI Sora • Higgsfield vs Runway Gen-3 • AI Video Credit Calculator
× ‹ enlarge ›