📛 Rarest Names in the World

🔒 The rarest names are the ones the data refuses to show you

Every statistics office that publishes name counts hides the smallest ones, because a count of one is a person you could identify. The US Social Security Administration excludes any name with fewer than five occurrences in a geographic area, which is why its state counts for a birth year, added together, fall short of the national count. The ONS releases England and Wales names from three upward; in the 2023 data, anything with a count of two or less was "redacted using S40 of the Freedom of Information Act to protect the confidentiality of individuals". France's INSEE applies two tests: a spelling appears on its own only if it was given at least 20 times to one sex in the period measured (1900 to 1945 and 1946 onwards count apart), or it is pooled into a row labelled _PRENOMS_RARES_; and a year in which it appears fewer than three times is replaced with XXXX. The Dutch Voornamenbank at the Meertens Institute just prints "< 5".

That is the hard limit on every name-rarity tool. For a rare name the honest output is a range, and a missing name is never a zero — the count fell under that office's floor: one or two in England and Wales, one to four in the US and the Netherlands.

⚖️ Why two rarity checkers disagree about the same name

Spelling is counted literally. ONS counts use the exact spelling of the first name on the birth certificate, which is why Muhammad, Mohammed and Mohammad hold three separate positions instead of one combined entry — on the 2015 list, 12, 29 and 68, with Muhammed at 121. The ONS notes that merging spellings distorts a list: combining Sophie, Sophia and Sofia that year would have pushed Amelia off the top. A site that merges variants calls your name common; one that keeps them apart calls it rare. Neither is wrong.

The base population differs. The ONS and SSA files count birth registrations, so they answer "how rare was this among babies born in 1994", not "among people alive today". A name that peaked in the 1930s looks extinct in recent birth data. The Voornamenbank runs the other way: its names are those in the Dutch population register as it stood on 1 January 2018, plus names from nineteenth-century marriage records.

Position in the name matters. The ONS counts the first name only; the Voornamenbank counts a name whether it sits first or later in the sequence — one reason it holds 443,801 distinct single forenames, in a country CBS counted at 18.1 million people on 1 January 2026.

There is no global register. Any "X million people share your name" figure extrapolates from the countries that publish frequency tables onto the ones that do not — and even the publishers leave gaps: SSA national data covers the 50 states and the District of Columbia, leaving out Puerto Rico, Guam and other US territories. Treat worldwide totals as orders of magnitude, not counts.

📮 Getting a real count for your own name

If you are in the UK, you can simply ask. Special extracts of baby names can be ordered from the ONS health data team, subject to disclosure control and its charging policy, and answers to single-name queries appear in the ONS "user requested data" series. Reference 3535, released on 21 July 2026, is a spreadsheet of nothing but the number of babies named Elm in England and Wales between 1996 and 2025. The INSEE file, the Voornamenbank and the SSA's state downloads check a name against a register rather than a model. Our data sources page lists which registers these calculators draw on.

In 2023, England and Wales birth registrations produced 29,560 unique names and spellings for boys against roughly 35,000 for girls — the ONS reads that as parents drawing boys' names from a smaller pool. US data agrees: for 2025 births the top 1,000 names represent about 72% of all names, 75.59% among boys but 67.25% among girls. Rarity is the default condition for a name, which is why "I've never met another one" is such a weak test of how unusual yours really is.