The number says how wide a code unit is. However, it doesn't say anything about how many code units are required to encode a single code point. UTF-8 needs 1–4,

The number says how wide a code unit is. However, it doesn't say anything about how many code units are required to encode a single code point. UTF-8 needs 1–4, UTF-16 needs 1–2 and UTF-32 needs 1. They all are able to represent all encoded characters of Unicode, it's just that the individual bytes are different.