Global Internet Speed Statistics - Real-World Broadband Data

Every number on this page comes from real speed tests run by real visitors using the speedmbps.com tool - and it updates live. Unlike the speeds your ISP advertises in its marketing, these figures reflect what actual connections delivered during real-world use: real downloads, real uploads, real latency, measured across more than 90 countries and 100 network providers.

This dataset includes historic results that were automatically saved under the site's policy before August 13, 2026. From August 13, a visitor must choose to create a share link before non-identifying measurement fields are included. Tests kept only in the visitor's browser are not added. The figures are a useful field snapshot, not a representative census of every connection.

What These Statistics Show

The four headline figures at the top summarize the whole dataset at a glance:

• Total Tests - the number of completed speed tests recorded in this dataset. It grows every time a visitor runs the tool.
• Average Download Speed (Mbps) - how fast data, on average, travels from the internet to users' devices.
• Average Upload Speed (Mbps) - how fast data travels back from users' devices to the internet.
• Average Latency (ms) - the average round-trip delay between a user's device and the test server. Lower is better.

Think of download speed as how quickly content reaches you, upload speed as how quickly you can send content out, and latency as how responsive your connection feels. All three matter - a connection can have huge download speeds and still feel sluggish if its latency is high.

Internet Speeds by Country

This table ranks the countries in our dataset by the number of tests run from each location. Countries consistently near the top - such as Hong Kong and Finland - tend to have dense fiber-optic infrastructure and low-latency networks. Countries with larger populations or wider rural coverage, such as the United States and India, show a bigger spread of results because their samples include everything from fiber city connections to rural wireless.

Important: the sample size differs enormously between countries - from thousands of tests for the United States down to a handful for smaller nations. That means a small country with very few tests can show a misleadingly high or low average. Always check the test count before drawing conclusions.

Internet Speeds by Provider (ISP)

Here you can see how individual network providers compare on download speed, upload speed, and latency. This is the closest thing most people will ever get to an independent, side-by-side ISP comparison.

One thing worth knowing: the provider names are reported by the network itself during the test, so a provider can appear under slightly different names (for example, "Optimum" and "Optimum Online"). Some entries are also corporate or data-center networks rather than home internet providers - that's why you'll occasionally see names like "Cloudflare" or "Private Customer" in the list. These are legitimate test results, just from a different kind of connection.

Internet Speeds by City

Drilling down to metro areas shows how much broadband performance can vary within a single country. In this dataset, cities with strong fiber deployment - Hong Kong, Helsinki, Toronto - frequently post the best combination of high speed and low latency. The table also reveals how a few very fast tests in a small city can lift its average, which is why we always display the number of tests alongside each figure.

Methodology and Accuracy

We believe transparent data beats impressive-sounding claims, so here is exactly how this dataset works:

• Source: every included result comes from the speedmbps.com speed test tool, which runs entirely in the visitor's browser using Cloudflare edge infrastructure. No app download or registration is required. Historic records were saved automatically; from August 13, 2026 onward, inclusion requires the visitor to create a share link.
• Coverage: The dataset is neither a controlled national sample nor independently audited. Cloudflare operates server locations in hundreds of cities worldwide, which enables measurements from many countries but does not make results nationally representative.
• What each test measures: download speed, upload speed, latency (ping), plus advanced metrics like jitter and packet loss that most speed tests do not show.
• Real-world conditions: tests are run on whatever connection the visitor is actually using - WiFi, Ethernet, mobile data, or even a VPN or corporate network. We do not filter these out, because the goal is to reflect real-world performance, not idealized lab conditions.
• Identifiers: city, country, and provider are reported by the network at the time of a shared test. They can be wrong, incomplete, or inconsistent; small samples and unusual network labels should be treated as indicative rather than conclusive.

These figures are field measurements for informational and educational purposes. They are not official ISP line tests, and they should not replace a technician's diagnostic tools. Read more in our privacy policy and terms.

Why This Data Matters

Real-world speed statistics are useful for more than curiosity:

• Choosing an ISP - see how providers actually perform across thousands of tests before you sign a contract.
• Checking your own plan - compare your test results against the national and provider averages above.
• Troubleshooting slow internet - if your speeds sit far below your country's average, there may be an issue with your router, WiFi, or plan.
• Remote work and gaming decisions - latency and upload speed matter more than raw download for video calls, VPN work, and online gaming.

Frequently Asked Questions

How is this statistics page updated?

Historic results were saved automatically. From August 13, 2026 onward, a visitor must deliberately create a share link after the test before a result is included. The aggregated figures update as new shared results are completed.

Why is the average different from my own speed?

Your speed is influenced by many factors: your plan, your router, whether you're on WiFi or Ethernet, the time of day, and how congested your area is. The averages above blend thousands of different conditions, so your result naturally varies from the mean.

Why does one country or provider have so few tests?

Sample size simply reflects how many visitors ran a test from that location. Treat small samples (under about 50 tests) as indicative rather than conclusive.

Can my speed test result appear here?

From August 13, 2026 onward, only if you choose to create a share link after the test. If you do nothing, your test stays in your browser history and is not included in the public aggregate.