The Greek alphabet in order, both cases, with the accented letters Greek writing uses, each carrying its real codepoint and official Unicode name. The notes cover final sigma, the unassigned slot in the capital run, and the letters that are not their Latin lookalikes.
Tap any tile to copy it. Nothing is sent anywhere — the copy happens in your browser.
Twenty-four letters, two cases, and a hole in the middle of the run
The capitals run from Α U+0391 to Ω U+03A9 and the small letters from α U+03B1 to ω U+03C9, but the capital run is not continuous. U+03A2 is unassigned: the slot sits empty where a capital final sigma would go, and no such letter exists, because final sigma is a positional form and a capital never takes it.
ς and σ are one letter in two positions
σ U+03C3 GREEK SMALL LETTER SIGMA is used inside a word, ς U+03C2 GREEK SMALL LETTER FINAL SIGMA at the end of one. They are separate characters, and the case mapping does not round-trip: uppercase ς and you get Σ; lowercase that Σ and you get σ, not what you started with. Casefolding — the operation a case-insensitive comparison ought to be using — maps both to σ, so a search that casefolds matches a word spelled either way and a search that merely lowercases does not.
The accented letters belong to Greek writing, not to notation
ά, έ, ή, ί, ΰ, Ϊ and Ϋ carry the tonos and the dialytika. They are precomposed characters and NFD takes them apart: ά becomes α U+03B1 followed by U+0301 COMBINING ACUTE ACCENT, and ΰ becomes three characters. NFC puts them back together. So the same visible word can be one character or two per accented letter depending on which normalisation last touched it — which is the usual explanation when two strings that look identical refuse to compare equal.
These are letters, and software treats them as letters
Category Lu and Ll, not Sm or So. A validator that permits “letters only” permits them. A sort that follows the Unicode default order files them after the Latin letters, not among them. A spellchecker running over an English document has a Greek word in front of it. This is the opposite of how a maths operator character behaves, and it is why a Greek letter pasted into a username or an identifier gets through checks that a symbol would fail.
Α is not A
Α U+0391, Β U+0392, Ε U+0395, Ζ U+0396, Η U+0397, Ι U+0399, Κ U+039A, Μ U+039C, Ν U+039D, Ο U+039F, Ρ U+03A1, Τ U+03A4, Υ U+03A5 and Χ U+03A7 are drawn like Latin capitals and are not Latin capitals. They never compare equal, and no normalisation form merges them: NFKC leaves Α as U+0391. The same holds in lower case for ο, ν and ρ against Latin o, v and p. Paste one into a password, a URL or a code identifier and you have a different string that looks the same, which is a debugging session nobody enjoys.
Width is not uniform across the alphabet
The twenty-four base letters in both cases have East Asian width Ambiguous, so a terminal set to render ambiguous characters wide gives each of them two cells. The accented forms and ς are width Neutral and take one. A Greek word therefore changes width depending on whether it happens to contain an accent, which is enough on its own to break a monospace column that was padded out with spaces.
The wing index is the maths and science symbols list. If you came for a single letter because of what it denotes — delta for a change, sigma for a sum — those readings are on greek letter symbols, which also covers the italic maths forms a formula needs. Other scripts are on foreign characters.