What Is the Average Typing Speed?

Live distribution of typing speeds

Every bar below is a 10-WPM slice of how fast real visitors typed the passage in the 60-second test. Taller bars mark the most common speeds; the long right tail is the small number of very fast typists.

What this means

Typing speed is a trained motor skill, so a spread of results — not a single fixed number — is the honest description of how fast people type. Casual typists who never learned touch typing tend to cluster lower, trained touch typists sit in a broad middle band, and a thin tail of very fast typists stretches the top end. Because this test counts only correctly typed characters, the score already blends raw speed with accuracy: hammering keys quickly but sloppily does not beat a calm, clean run. That said, the numbers on this page are our dataset — from visitors to this tool, on one specific passage, not a controlled lab sample — so treat them as a live, informal measurement rather than a peer-reviewed result.

Methodology & reproduce this

What's measured

Everyone types the same fixed passage for 60 seconds. net_wpm = (correct_chars / 5) / minutes, where a character is correct only if it matches the passage at the same position, and accuracy = correct_chars / typed_chars as a percent. The timer starts on your first keystroke. Runs shorter than a few seconds or with only a handful of characters are discarded client-side before they're ever sent.

Headline statistic

The median net WPM across all stored results, which is resistant to the occasional wild outlier in a way a plain average is not, reported alongside the median accuracy. It only appears once at least 1,000 tests have been recorded; before that the page shows the running result count instead.

Underlying dataset

The full set also carries a trimmed mean and 10th/90th-percentile WPM bounds, available in the raw export below — the chart and headline on this page use only the median and the histogram.

Anti-spam & anti-cheat

Pasting into the box is blocked, each submission carries a short-lived signed token, results outside a plausible 1–250 WPM range are rejected, and a per-visitor daily limit caps how many runs count. Because the headline is the median rather than the mean, it stays robust even if a few stray values slip through.

Download the raw dataset (CSV) — timestamp, net WPM and accuracy per test, nothing else. That export, together with the methodology above, is the reproducible artifact: compute the median WPM and accuracy, and for the trimmed mean drop the lowest and highest 10% of rows before averaging — no proprietary steps or hidden data. Run the reproduce harness → — a small Node script that pulls the CSV and recomputes every number above using the site's own formulas.

Frequently asked questions