Reading Level Calculator
Paste any text to compute 6 readability scores — Flesch Reading Ease, Flesch-Kincaid Grade Level, Gunning Fog, SMOG, Coleman-Liau, ARI — plus long-sentence flags and plain-English suggestions.
Reading Level Calculator
Long sentences (> 25 words)
- Analysing…
Plain-English suggestions
Analysing…
How to use the Reading Level Calculator
Paste English text into the box
Drop in an article, an email, a product page or a policy. The page opens with a short passage we wrote for it: four long sentences built almost entirely from one-syllable words, which the five grade formulas score between grade 4.4 (Coleman-Liau) and grade 12.9 (Fog). Everything runs in your browser; nothing is sent to a server. Longer samples give steadier scores: Kincaid's Navy passages averaged 170 words, and McLaughlin normed SMOG on 30 sentences, so under 30 sentences the page marks the SMOG figure unreliable. Under 30 words it tells you that one sentence is deciding every score.
Read the range, not one number
Flesch Reading Ease runs the other way from the rest: higher is easier. The five grade-level formulas each report a US school grade, and they rarely agree. SMOG was fitted to complete comprehension, the others to a lower bar; and even formulas fitted to the same readers on the same passages disagree, because each counts something different. The box above the cards shows the lowest and highest of the five and the gap between them. The page does not average them, because an average of SMOG and Flesch-Kincaid mixes two different standards of understanding.
Use the long-sentence and substitution panels
Every formula here treats a longer sentence as harder, so the long-sentence panel lists any sentence over 25 words, which is where to start cutting. Flesch-Kincaid adds 0.39 of a grade for each extra word in the average sentence: taking the average from 25 words to 15 takes 3.9 grades off. The substitution panel flags common formal words with shorter equivalents: utilise → use, commence → start, terminate → end, facilitate → help, demonstrate → show.
Check what your own guidance actually asks for
Few bodies name a formula. The US Plain Writing Act of 2010 names none and sets no grade. GOV.UK's guidance said in March 2025 (archived copy) that it tells people to write "for a 9 year old reading age"; it named no formula, and the current version we checked names no reading age at all. Microsoft's help for Word says that for most standard files "you want the score to be between 60 and 70" on Flesch Reading Ease, and a Flesch-Kincaid grade "around 7.0 to 8.0". Florida requires insurance policies to reach 45 on Flesch Reading Ease, counted by its own rules. Edit, paste back in, and watch which scores move.
Why six readability formulas give one text different grades
Rudolf Flesch published his Reading Ease score in 1948, in "A new readability yardstick" in the Journal of Applied Psychology. In 1975 J. Peter Kincaid and colleagues refitted it for the US Navy as a school-grade number, the Flesch-Kincaid Grade Level, using 18 passages from Navy training manuals read by 531 enlisted personnel at four technical training schools. Microsoft Word reports both scores, though in older versions only once you tick "Show readability statistics". Florida wrote Flesch Reading Ease into law for insurance policies, with a minimum of 45. Both scores measure the same two things: how long the sentences are and how many syllables the words have.
Different standards, and different things counted
Two things separate the formulas. The first is the standard of understanding they were fitted to. Flesch-Kincaid gave a passage a grade when half of the Navy readers at that reading grade scored 35% or better on a cloze test, where the reader fills in deleted words; Kincaid chose that bar because earlier work had shown a cloze score of 35 to 40% to be comparable to 75% correct on multiple-choice questions, the criterion Flesch used. SMOG (Harry McLaughlin, 1969) was fitted to "the grade of subjects showing complete comprehension" of the McCall-Crabbs reading lessons, a stricter bar, and McLaughlin noted that SMOG grades "are generally two grades higher" than corrected Dale-Chall levels. The second is what each formula counts. Gunning Fog (Robert Gunning, 1952) counts words of three or more syllables; this page leaves out capitalised names and words that reach three syllables only through -ed, -es or -ing, which is how the rule is usually applied, though we have not checked it against Gunning's book. Coleman-Liau (1975) and the Automated Readability Index count letters instead of syllables. The counting alone is enough to split them: Kincaid fitted a new ARI, a new Fog Count and Flesch-Kincaid to the same readers, the same criterion and the same 18 passages, and his own table still puts the three a median 1.5 grades apart, and 4.0 apart on one passage.
A readability score measures sentence length and word length, nothing else. Shorten both and the grade falls; whether the reader then understands is still your job.
How close the formulas get to real readers
Kincaid's report prints, for each of its 18 Navy passages, both a measured reading grade (his scaled estimate from the sailors' cloze scores and reading tests, itself not a perfect measure) and the Flesch-Kincaid grade. On the very passages the formula was fitted to, its grade misses the measured one by 2.1 grades on average and by as much as 4.3. A passage titled "Communiques" was measured at grade 16.0, while Flesch-Kincaid put it at 12.3. Kincaid, citing earlier work, reports that readability formulas generally have validity coefficients of about .70, which "account for about 50 percent of the variability", and that Flesch Reading Ease, "the most widely validated and used of all readability formulas, has never been validated for use by military personnel." McLaughlin's own estimate for SMOG is that it predicts a passage's grade "within one and a half grades in 68 percent of cases". Read a grade as a rough position, not a measurement of your reader.
What counts as a sentence moves the score
Every formula divides by the number of sentences, so the rule for ending one matters as much as the formula. This page ends a sentence at a full stop, question mark or exclamation mark, at a blank line, a list item or a short heading line, and not after a title such as "Dr." or an initial. Florida's insurance rule counts differently: "A unit of words ending with a period, semicolon, or colon, excluding headings and captions, shall be counted as one sentence." Scored here, subsections (1) and (2) of that statute come out at a Flesch Reading Ease of 10.0 counting full stops (four sentences) and 48.2 counting semicolons and colons as Florida does (ten). The statute is not an insurance policy and we did not apply its exclusion of defined words, but on text punctuated like a legal clause the sentence rule alone moved the score by 38 points.
English only: what the page does with Malay and Chinese
The two calibrations we could check were both made with American readers and American English texts: Navy enlisted personnel for Flesch-Kincaid, and the McCall-Crabbs school reading lessons for SMOG. None was built or tested for Malay, Chinese, Tamil or Filipino, and the page does not pretend otherwise. Text that reads as Malay or Indonesian is refused with a note rather than scored: a Malay translation of plain English bus directions scores Flesch-Kincaid grade 14.9 here, against 3.5 for the same English, almost entirely because the Malay words carry more syllables. Text that is mostly Chinese, Thai, Tamil or another non-Latin script is refused too, as is Latin-script text heavy with accents such as Vietnamese. In mixed text, the page scores the English words and sentences and says how much it left out. Readability formulas built for those languages do exist, among them a Malay formula in Khadijah Rohani Md. Yunus's 1982 dissertation, the CRIE analyser for Chinese (Sung and colleagues, 2016), a formula for Chinese as a second language in Teaching Chinese Language in Singapore (Soh, 2020), and a 2015 study of Fry and SMOG on Filipino texts. Microsoft Word makes a similar call: for some European languages inside an English document it shows "only information about counts and averages, not readability."
The disagreement between tools is a known problem in the region. A 2014 study in the Malaysian Journal of Learning and Instruction ran six formulas over 64 English texts and found "significantly different readability indices for the same text using different formulae". Singapore's Hack for Public Good 2026 project SpeakEasy reported that across five key government sites "only 13% of content is written at lower secondary level or below", and quoted a UX agency: "Different tools give different results." That is why this page shows every formula and the gap between them.
McLaughlin named SMOG after Gunning's Fog, and Florida counts a semicolon as a full stop
Flesch published his Reading Ease formula in 1948, in "A new readability yardstick", Journal of Applied Psychology 32(3):221–233.
Florida Statutes §627.4145 deems an insurance policy readable only if its text scores at least 45 on Flesch Reading Ease, counting a unit of words ending in a period, semicolon or colon as one sentence.
Flesch-Kincaid came from a 1975 US Navy report: 18 passages from Navy training manuals, read by 531 enlisted personnel at four technical training schools.
McLaughlin wrote that he called his system SMOG Grading "in tribute to Gunning's Fog Index", and that the name also refers to London, his birthplace. The 1969 paper does not spell out "Simple Measure of Gobbledygook".
McLaughlin credited Robert Gunning, author of The Technique of Clear Writing (1952), as the first to count polysyllabic words.
The US Plain Writing Act of 2010 requires federal agencies to write clearly, defining plain writing as "clear, concise, well-organized" — it sets no readability score.
Coleman and Liau titled their 1975 paper "A computer readability formula designed for machine scoring": it counts letters, which a machine can do exactly, instead of syllables.
On the 18 passages Flesch-Kincaid was fitted to, Kincaid's own grades miss the measured reading grade by 2.1 grades on average and by up to 4.3.
McLaughlin's own page described SMOG as "widely used, particularly for checking health messages".
SMOG was calibrated on samples of 30 sentences; on a shorter passage its grade is unreliable.
Frequently Asked Questions
-
None of them on its own. SMOG was fitted to complete comprehension and the others to a lower bar, and formulas fitted to the same readers still disagree because each counts something different. So the highest score is not more honest and the lowest is not more generous. SMOG also needs 30 sentences to mean what McLaughlin meant by it. Even on the passages it was built from, Flesch-Kincaid misses the measured reading grade by about two grades on average. Use the range the page shows, the same formula each time you compare drafts, and a real reader when it matters.
-
Use whatever your own organisation or regulator sets, and check which formula and counting rules it means. Published figures differ: GOV.UK's guidance said in March 2025 (archived copy) "we tell people to write on GOV.UK for a 9 year old reading age", named no formula, and no longer gives a reading age in the current version we checked; Microsoft's Word help suggests a Flesch Reading Ease "between 60 and 70" and a Flesch-Kincaid grade "around 7.0 to 8.0" for most documents; Florida requires 45 on Flesch Reading Ease for insurance policies. The US Plain Writing Act of 2010 names no formula and no grade.
-
With rules, not a dictionary. It counts groups of vowels (a, e, i, o, u, y), drops a silent final "e" (but keeps the "e" of "table" and "people", which is the syllable), drops a final "-es" unless it follows s, x, z, ch, sh, c or g (so "makes" has one syllable and "sentences" three), drops a silent "-ed", and never returns less than one. Hyphenated words are counted part by part. On a 3,030-word sample of the CMU Pronouncing Dictionary it matches the dictionary's count for 86.7% of words. Fog and SMOG are the most sensitive to a miscount, because a single syllable decides whether a word counts as long.
-
Because sentence length is one of the two inputs every one of them was fitted on, alongside word length. Flesch-Kincaid, Fog and ARI use words per sentence directly; SMOG uses sentences in its count of long words per 30 sentences; Coleman-Liau uses sentences per 100 words. A common explanation is working memory: George Miller's 1956 paper on "the magical number seven, plus or minus two" and Nelson Cowan's 2001 revision to about four items describe how little a reader holds at once, and a long sentence asks the reader to hold its start while parsing its end.
-
No. All six were built for English, and none was built or validated for Malay, Chinese, Tamil or Filipino. The page refuses to score text that reads as Malay or Indonesian, text that is mostly in a non-Latin script such as Chinese, Tamil or Thai, and Latin-script text heavy with accents such as Vietnamese, and says why instead of printing a number. A Malay translation of plain English bus directions would otherwise score Flesch-Kincaid grade 14.9 against 3.5 for the same English. For mixed text, common in Singapore and Malaysia, the page scores only the English words and sentences and tells you how much it left out. Formulas built for Malay and for Chinese exist; these six are not them.
-
The formulas were fitted on passages, not sentences: Kincaid's Navy passages averaged 170 words, and McLaughlin normed SMOG on samples of 30 sentences and wrote that tables for texts of fewer than 30 sentences "are statistically invalid". Under 30 sentences the page marks SMOG unreliable, and under 30 words it says that one sentence decides every score. On very short or very long sentences the numbers can also leave their usual ranges: "I go. We run. You sit. He ran." scores 120.2 on Flesch Reading Ease, which is printed as computed rather than cut to 100.
-
Same inputs, different scale. Both use words per sentence and syllables per word. Flesch Reading Ease gives a score where higher is easier; the bands the page prints beside it (90 and above very easy, 60 to 70 standard, below 30 very difficult) are the conventional reading of the scale, usually attributed to Flesch's 1948 paper, which we have not been able to check. Flesch-Kincaid Grade Level turns the same inputs into a US school grade with Kincaid's 1975 coefficients: 0.39 × words per sentence + 11.8 × syllables per word − 15.59.
-
Google's guidance asks for "people-first content" and names no reading grade, so a readability score is not a ranking factor it publishes. A lower grade is also not the same as clearer writing: the sample passage on this page is made almost entirely of one-syllable words and still scores up to grade 12.9, because its sentences are long. Write for the reader you have, and use the scores to find the sentences that are doing too much.
-
"Utilise" has three syllables; "use" has one. Every syllable-based formula here weights syllables per word, so replacing a three-syllable word with a one-syllable synonym lowers the score. The substitution panel flags a fixed list of common formal words that have shorter equivalents: utilise → use, commence → start, terminate → end, facilitate → help, demonstrate → show, ascertain → find out, leverage → use, optimise → improve. Check each swap in context: they usually keep the meaning, not always.
-
No. All analysis happens entirely in your browser via JavaScript — open DevTools → Network and watch as you type, there is zero outbound traffic. Your text never leaves your device. Because the text is processed only on your device, we collect no personal data from it: nothing is stored and nothing is logged. When you close the tab, the text is gone.
Method & sources
How it computes
Flesch Reading Ease, Flesch-Kincaid (Kincaid 1975), Gunning Fog, SMOG (McLaughlin 1969, every 3+ syllable word), Coleman-Liau and ARI (letters and digits), shown side by side with the spread between the five grades, never an average. Syllables by vowel-group rules (86.7% exact on a CMU dictionary sample). Text that is not English (Malay, Chinese, Vietnamese) is refused; mixed text is scored on its English sentences. Shared kernel with Text Statistics and Word Counter.
What this tool implements
- Flesch-Kincaid as Kincaid et al. 1975 Table 3 writes it: 0.39 × words/sentence + 11.8 × syllables/word − 15.59; ARI with the coefficients that table labels "Old" (4.71, 0.5, 21.43), counting letters and digits. Kincaid fitted Flesch-Kincaid to 531 Navy enlisted personnel at four technical training schools; a passage got a grade when half the readers at that grade scored 35% or better on a cloze test. On those 18 passages Kincaid's own grades miss the measured reading grade by 2.1 grades on average; the kernel reproduces his Flesch-Kincaid within 1.6 grades on each.
- SMOG as McLaughlin 1969 Table 1 equation (d): 1.0430 × √(polysyllables × 30/sentences) + 3.1291, counting "every word of three or more syllables" with each repetition and no exclusions; normed on 30-sentence samples, and the page marks SMOG unreliable below 30 sentences. McLaughlin named it "in tribute to Gunning's Fog Index"; the paper does not spell out "Simple Measure of Gobbledygook".
- The five grade scores are shown with their spread, never averaged: SMOG was fitted to complete comprehension and the others to a lower bar, and Kincaid's own Table 2 puts three formulas fitted to the same readers a median 1.5 grades apart.
- The formulas were built for English. Text that reads as Malay or Indonesian, text mostly in a non-Latin script, and Latin-script text heavy with accents are refused, not scored; in mixed text only the English sentences are scored. A Malay translation of plain English directions would score Flesch-Kincaid 14.9 against 3.5 for the same English.
- The Plain Writing Act of 2010 names no formula or grade; GOV.UK's guidance said "9 year old reading age" in March 2025 (archived); Florida §627.4145 sets Flesch Reading Ease 45 for insurance policies and counts a unit ending in a period, semicolon or colon as one sentence. The Flesch band names are conventional and unverified against the 1948 paper.
Sources
- Kincaid JP, Fishburne RP Jr, Rogers RL, Chissom BS. Derivation of New Readability Formulas (Automated Readability Index, Fog Count and Flesch Reading Ease Formula) for Navy Enlisted Personnel. Research Branch Report 8-75. Naval Technical Training Command; 1975 (DTIC ADA006655). https://stars.library.ucf.edu/istlibrary/56/
- McLaughlin GH. SMOG Grading — a new readability formula. Journal of Reading 1969;12(8):639-646. https://web.archive.org/web/20070710/http://www.harrymclaughlin.c…).pdf
- Flesch R. A new readability yardstick. J Appl Psychol. 1948;32(3):221-233. https://doi.org/10.1037/h0057532
- Coleman M, Liau TL. A computer readability formula designed for machine scoring. J Appl Psychol. 1975;60(2):283-284. https://doi.org/10.1037/h0076540
- Florida Statutes §627.4145, Readable language in insurance policies. http://www.leg.state.fl.us/statutes/index.cfm?App_mode=Display_St…
- Plain Writing Act of 2010, Pub. L. 111-274, 124 Stat. 2861. https://www.govinfo.gov/content/pkg/PLAW-111publ274/html/PLAW-111…
Pick up where you left off
Stored only in this browser — never sent to our servers.