Superscript Generator

The most common complaint about Unicode superscript is that the digits look inconsistent with each other, with one or two sitting at a different height or weight from the rest. This isn't a font bug and it isn't avoidable by choosing better characters. It's a consequence of how these characters were added to Unicode.

Three digits come from a different neighborhood

Superscript one, two and three were encoded early, in a legacy block inherited from older Western European character sets where they already existed for fractions and footnotes. The remaining digits, zero and four through nine, were added later in the dedicated Superscripts and Subscripts block.

Because those two groups arrived separately, fonts often draw them separately, and not always to the same metrics. A sequence mixing them, such as an exponent of twelve or a year, is where the mismatch becomes obvious: the digits from the older block can appear slightly larger, bolder, or set at a different height than their neighbors. Fonts with careful coverage harmonize them, and many fonts don't.

The letters aren't really superscripts

Superscript letters have a further complication. Most of them are classified in Unicode as modifier letters, characters that exist to support phonetic notation in linguistics rather than to raise ordinary text. They look like superscripts and they are categorised as letters that modify the sound of what precedes them.

That classification has practical effects. Coverage is uneven, since the set was chosen for phonetic usefulness rather than to complete an alphabet, so some letters have no superscript form at all. It also explains why assistive technology sometimes announces them by their phonetic names rather than reading them as raised letters.

When to use markup instead

Unicode superscript exists to survive plain-text fields where formatting is stripped: social bios, usernames, chat messages, spreadsheet cells. In any environment that supports real formatting, proper superscript markup is the better choice. It uses the actual letters. So the text stays searchable and readable to assistive technology, and the typeface handles the raising consistently rather than depending on which Unicode block each character came from.