Arabic typing test

Right-to-left, with the 600 most frequent Arabic words and a keyboard you can click if you have no Arabic layout installed.

15 left
No keyboard installed? Click the keys.

Arabic 101 and 102 — the difference is smaller than you have been told

The usual advice is to work out whether you are on Arabic 101 or Arabic 102 before you learn anything. It is worth correcting: the Arabic letters sit in identical positions on both. What differs is the punctuation keys, one extra key near the left Shift, and which operating system defaults to which.

Arabic 101 follows the ANSI physical layout used almost everywhere, with a QWERTY Latin base underneath the Arabic layer.

Arabic 102 is the ISO-style board, carrying the extra key that European physical keyboards have. The variant common across the Maghreb — Morocco, Algeria, Tunisia — keeps the Arabic layer identical but puts AZERTY underneath rather than QWERTY, a legacy of French influence on which keyboards the region bought. So a Moroccan typist and a Gulf typist are using the same Arabic layout with different Latin layers beneath it.

The practical consequence: if you are learning Arabic letter positions, you are learning one thing, not two, and a 101 tutorial serves a 102 typist perfectly well. Where the layouts genuinely diverge is diacritics and symbols, which is also where most people never go.

What this word list actually measures

Arabic is the outlier of the six: codepoints and visual units are almost the same thing, and the letter inventory is the smallest here. Counted from the exact list served above:

600words in the test list
3.94visual units per word
4.02codepoints per word
53distinct letter shapes
7%words that stack

Those two middle numbers sit almost on top of each other, and that is the finding rather than a rounding artefact. A word here takes about 4.02 keystrokes to build but occupies 3.94 visual units on the page — a ratio of 1.02 to one. On this word list that convention lands within a few per cent of the figure shown here — Arabic barely stacks, so characters and visual units are nearly the same thing.

In Arabic, speed is a travel problem, not a building problem

Because Arabic barely stacks, practice here behaves the way Latin-script practice does: you are training your hands to find keys quickly, not to assemble letters out of several keystrokes. That makes raw speed gains faster to come by in Arabic than in any Indic script on this site, and it puts the ceiling somewhere different — in layout familiarity and finger travel rather than in sequence knowledge.

What does not come for free is choosing correctly between letters that look nearly identical. ة and ه, the four hamza forms, ى and ي. These are not speed problems and they do not improve by typing faster; if anything they get worse. They are a knowledge problem wearing a typing problem's clothes, and the fix is to slow down on the specific words that contain them until the right form is the one your hand reaches for.

The missed-character list on the result card is unusually useful in Arabic for exactly this reason. In Devanagari it tells you which sequences you have not drilled. In Arabic it tells you which letters you are confusing — which is a spelling report as much as a typing one.

Why Arabic counts differently from every other script on this site

Arabic is the outlier here, and the measurements make it obvious. Across this word list, Arabic averages 4.02 codepoints and 3.94 visual units per word — a ratio of about 1.02 to one. For the Devanagari and Bengali tests on this site that ratio runs closer to 1.7. Only 7% of Arabic words in this list contain a cluster that takes more keystrokes than it shows.

The reason is that unvocalised Arabic — ordinary written Arabic, the kind in newspapers and messages — has almost no stacking. One letter, one keystroke, one visual unit. The contextual shaping that makes Arabic look cursive happens in the font at render time, not in the text: ع alone and ع in the middle of a word are the same character drawn differently, and it costs you nothing extra to type. Add full harakat and this changes, but almost nobody types that outside the Qur'an, poetry and teaching material.

What this means for your score: Arabic is the one language on this site where the conventional character-based word count is roughly honest. It also means Arabic typing speed is limited by finger travel and layout familiarity rather than by building letters, so practice transfers to raw speed more directly than it does in an Indic script.

The genuinely Arabic difficulty is elsewhere — deciding which of several near-identical letters a word wants.

The letters that cost Arabic typists accuracy

ة

Taa marbuta against ه

مدرسة ends in ة, not ه. They are adjacent in sound and far apart in correctness, and this is the most common Arabic typing error there is.

أ

The hamza forms

ا أ إ آ are four distinct characters that all look like alif at a glance. Getting the wrong one produces a word that reads correctly and compares as wrong.

ى

Alif maqsura against ya

على ends in ى; علي is a different word ending in ي. One dot pair, two meanings.

ـ

Tatweel

The kashida stretches a word visually and is a real character in the text. Typed by accident, it silently breaks a match.

Sources

The claim that Arabic 101 and 102 place the Arabic letters identically — and that the Maghreb variant changes only the Latin layer beneath — was checked against layout documentation rather than repeated from the common advice, which usually says the opposite.

If you came here to write, not to be measured

For writing Arabic rather than racing it, the Arabic keyboard has the full layout, harakat, predictive text and an editor. Other scripts are on the typing test hub.

FAQ

Frequently Asked Questions

Does the test run right to left?
Yes. The passage, the caret and the input all run right to left, which is the part most general-purpose typing tests get wrong or simply do not attempt.
How is the speed calculated?
Correctly typed whole words per minute, with visual letter units per minute alongside. On this word list that convention lands within a few per cent of the figure shown here — Arabic barely stacks, so characters and visual units are nearly the same thing.
Do I need an Arabic keyboard installed?
No. The on-screen Arabic keyboard below the test is clickable, which is the point — most people searching for an Arabic typing test do not have the layout installed.
Is this for Arabic 101 or Arabic 102?
Both. The Arabic letters are in the same positions on each, so there is nothing to choose between.
I am in Morocco and my keyboard is AZERTY. Does that matter?
Not for the Arabic layer. The Maghreb 102 AZERTY board keeps Arabic letters where Arabic 101 has them and changes only the Latin layer beneath.
Does the test require harakat?
No. The word list is unvocalised, which is how Arabic is normally written. That is also why Arabic scores here are comparable to Latin-script scores in a way the Indic ones are not.
Are hamza forms treated as different letters?
Yes, because they are. أ إ آ ا are four characters, and the test compares exactly. The result card will show you which one you keep missing.
Where do the test words come from?
An Arabic frequency corpus — the same verified list our Arabic keyboard autocompletes from, not a hand-written set.
Why is my Arabic speed closer to my English speed than my Hindi speed is?
Because Arabic barely stacks. A ratio of about 1.02 codepoints per visual unit, against roughly 1.7 in Devanagari. Fewer hidden keystrokes per word.
What is the longest word in the list?
بالتأكيد, at 8 letters.
Does it work on a phone?
Yes, with your phone Arabic keyboard.
Is my score saved anywhere?
Your personal best is kept in your browser only.
How many distinct letter shapes does the list use?
53 — by far the smallest inventory of the six languages tested here, and a real reason Arabic typing is quicker to learn than its reputation suggests.
Copied!