High-IQ Societies Compared: Mensa, Intertel, Triple Nine and More

Mensa is the one everybody has heard of, but it is the most accessible of a whole tier of organizations built on the same idea: admit people above a stated percentile on a recognized intelligence test, and ignore everything else about them. Above Mensa's 98th percentile sit Intertel at the 99th, the Triple Nine Society and the International Society for Philosophical Enquiry at the 99.9th, the Prometheus Society at the 99.997th, and the Mega Society at a claimed one in a million. This guide compares the thresholds, the admission routes, and the measurement problem that makes the highest tiers far shakier than their decimal places suggest.

1. What a high-IQ society actually is

A high-IQ society is a membership organization with a single entry criterion: documented performance at or above a defined percentile on an accepted test of general reasoning ability. There is no interview, no nomination, no requirement to have achieved anything. You either have the paperwork or you do not.

The model was invented in Oxford in 1946, when Roland Berrill and Lancelot Ware founded Mensa as a society with no political, religious or social agenda — a round table, as the Latin name suggests, where your background does not determine your seat. The idea proved durable enough that successive generations built higher-threshold versions of it, usually because a group of Mensa members felt the top 2 % was too broad a club to produce the conversations they wanted.

What these societies offer is almost always the same package: a journal or newsletter, online discussion, special-interest groups, occasional gatherings, and a directory of members. What they do not offer is a credential. No employer, university or licensing body treats membership in any of them as a qualification, and the societies themselves generally do not claim otherwise.

2. Percentiles, not IQ numbers

Every society states its requirement as a percentile, and for a good reason. An IQ score is a position on a bell curve expressed in arbitrary units, and not all tests use the same unit. Virtually all modern tests put the mean at 100, but the standard deviation — the width of the curve — differs between instruments. The same percentile therefore converts into different IQ figures depending on the scale.

Percentile	Rarity	≈ IQ on an SD-15 scale	≈ IQ on an SD-16 scale
98th	1 in 50	130	133
99th	1 in 100	135	137
99.9th	1 in 1,000	146	149
99.997th	~1 in 30,000	160	164
99.9999th	1 in 1,000,000	171	176

These are two ways of writing the same thing. A 176 on an SD-16 scale is not higher than a 171 on an SD-15 scale — it is the identical percentile in different units. This is why the figure quoted for a given society varies between sources, and why no society will accept "I scored X online" as evidence: without a known norming sample and a known standard deviation, the number has no percentile attached to it at all.

3. The major societies compared

The table below covers the organizations that come up most often. Founding years and thresholds are as the societies themselves have published them; membership figures for the smaller groups are rough, because most do not publish current counts.

Society	Founded	Threshold	≈ SD-15 IQ	Rarity	Journal
Mensa International	1946	98th percentile	~130	1 in 50	national magazines (e.g. Mensa Bulletin)
Intertel	1966	99th percentile	~135	1 in 100	Integra
International Society for Philosophical Enquiry (ISPE)	1974	99.9th percentile	~146	1 in 1,000	Telicom
Triple Nine Society (TNS)	1978	99.9th percentile	~146	1 in 1,000	Vidya
Prometheus Society	1982	99.997th percentile	~160	1 in 30,000	Gift of Fire
Mega Society	1982	99.9999th percentile	~171	1 in 1,000,000	Noesis

Mensa is in a category of its own by scale: a federation of national groups in roughly 100 countries with well over 100,000 members between them. It is the only one of these organizations with enough members to sustain local meetups in most large cities, paid staff, and foundation-run programmes.

Intertel was founded in 1966 as a programme of the International Legion of Intelligence, aimed at the top 1 %. It is global but much smaller than Mensa, and is known mainly for its journal Integra and its member round-robin correspondence groups.

ISPE was founded in 1974 and has a distinctly different flavour: it was conceived around scholarly and creative exchange rather than socializing, and has historically recognized higher membership grades for members who submit original scholarly or creative work. Its journal Telicom publishes member essays.

The Triple Nine Society — the name refers to the 99.9th percentile — was founded in 1978 and is the best-known of the one-in-a-thousand societies. It is run entirely by volunteers, publishes Vidya, and maintains a comparatively long and transparent list of accepted tests.

The Prometheus Society and the Mega Society were both established in 1982 by the philosopher Ronald K. Hoeflin, at the 99.997th and one-in-a-million levels respectively. Both are tiny — the Mega Society has at times had a membership countable in dozens — and both exist primarily as journals with a discussion list attached.

Beyond these, dozens of smaller societies have come and gone, including Hoeflin's Top One Percent Society and One-in-a-Thousand Society, the Cerebrals Society, OLYMPIQ, and the Giga Society, which claims a one-in-a-billion threshold. The further up the scale you go, the more the "society" tends to be one dedicated organizer, a self-published test, and a mailing list.

4. How admission works

The routes are broadly the same everywhere, but their practicality changes dramatically with the threshold.

	Mensa (98th)	99th–99.9th tier	Above 99.9th
Supervised society test	Yes, most national groups run one	Rarely; Intertel has offered its own	No
Prior evidence from a standard test	Yes, from a long approved list	Yes, the main route	Possible but limited by test ceilings
Unsupervised specialist tests	Not accepted	Generally not accepted	Commonly the only route
Realistic documentation	Straightforward	Straightforward if you have been tested	Difficult and contested

For Mensa, you either sit your national group's supervised admission test — usually a once-in-a-lifetime opportunity — or submit a qualifying score from one of the roughly two hundred standardized tests on its accepted list. The one-in-a-hundred and one-in-a-thousand societies lean almost entirely on the second route: you send documentation of a score you already hold, typically from an individually administered instrument such as the WAIS or the Stanford-Binet, or from a recognized group ability test.

Above the 99.9th percentile the ordinary route runs out, which is why those societies accept unsupervised, untimed "power tests" written specifically for the purpose. Hoeflin's own Mega Test, published in Omni magazine in 1985, and his later Titan Test are the canonical examples: long sets of extremely difficult verbal analogies and spatial problems, taken at home with no time limit, scored by the author. They are ingenious puzzles. They are not standardized instruments in the psychometric sense, and their norms come from self-selected respondents rather than a representative sample.

5. The ceiling problem above the 99.9th percentile

This is the part most comparison articles skip, and it matters more than any of the thresholds.

Norming a test means administering it to a representative sample and recording the distribution of scores. The WAIS-IV standardization sample contained roughly 2,200 people — large by clinical standards, and enough to describe the middle of the curve precisely. But the number of people in a 2,200-person sample who sit at one in a thousand is about two, and the number at one in a million is zero. Scores that far out are therefore not measured; they are extrapolated from the shape of the curve, assuming the curve stays perfectly normal in a region where almost nobody has been observed. The major clinical tests acknowledge this by simply stopping: the WAIS-IV's reported full-scale range tops out at 160, and the Stanford-Binet's standard range does the same.

Three consequences follow:

• The precision is cosmetic. A threshold stated as the 99.997th percentile implies a resolution that no available instrument delivers. The confidence interval around any individual score at that level is wide enough to swallow the distinction between neighbouring societies.
• Different tests disagree wildly at the extremes. Someone can qualify for a given society on one specialist test and fall well short on another, because the two were normed on different self-selected pools.
• The distinctions are socially real but psychometrically thin. The person at the 99.9th percentile and the person at the 99.99th are not reliably distinguishable by any supervised test in common use.

Mainstream psychometrics takes essentially this view, and so do many members of the societies concerned — the argument about whether high-range tests mean anything has been running inside their own journals for decades. It is also why Guinness World Records retired its "highest IQ" category in 1990, on the grounds that such claims could not be reliably verified.

6. What membership gives you — and what it does not

Members who get value out of these societies almost always describe the same thing: people. Local gatherings, special-interest groups covering everything from cryptography to opera, annual conventions, and journals where members argue at length in print. For the smaller societies the journal is the society, and the appeal is correspondence with a few dozen people who will take an unfashionable argument seriously.

What membership is not:

• Not a professional qualification. It carries no weight in hiring, admissions or licensing, and reads on a CV roughly the way a chess club does.
• Not a measure of achievement. It records performance on reasoning tasks, usually on a single day, and nothing about what has been done with that performance.
• Not a full picture of a mind. Creativity, judgement, domain expertise, conscientiousness, social skill and sheer persistence are all outside the admission criteria, and between them they explain most of what people actually accomplish.
• Not representative of high scorers. The top 2 % of the world's population is tens of millions of people; only a tiny, self-selected minority ever apply anywhere. Society members are joiners, not a sample of the able.

7. Common misconceptions

"The higher the society, the more rigorous it is." Usually the opposite. Mensa and the 99th-percentile societies rely on supervised, properly normed tests. The ultra-high tiers rely on unsupervised take-home tests with homemade norms, because nothing else reaches that far.

"An online test can get me in." Not into any of them. Unsupervised online tests — including Brambin's — have no verified conditions and, in most cases, no published norming sample, so they carry no percentile that a society could accept. If you are simply curious roughly where you sit before paying for a supervised session, you can try Brambin's free quick IQ check — 12 puzzles for a rough, just-for-fun estimate (not an accurate IQ, with no medical or scientific basis).

"Qualifying for one means qualifying for the next one up." No. Each society sets its own accepted-test list, and a score that satisfies one may not be on another's list at all, especially where test dates and editions have cut-offs.

"These are secret elite networks." They are volunteer-run hobby organizations with annual dues, newsletters and AGMs. The membership lists of several are effectively public.

"Members are geniuses." The word has no psychometric definition. Every threshold here is a statistical convenience, and the person just below any of them is not meaningfully different from the person just above — the gap is comfortably inside the measurement error of the test used to draw it.

Frequently asked questions

Which high-IQ society is the hardest to get into?

By stated threshold, the Mega Society at a claimed one in a million, with groups such as the Giga Society claiming to go further still. In practice "hardest" is ambiguous, because those admissions rest on unsupervised specialist tests whose norms are not independently verified. Among societies using properly normed supervised tests, the 99.9th-percentile tier — Triple Nine and ISPE — is the realistic top end.

What is the difference between Mensa and the Triple Nine Society?

The threshold and the scale. Mensa admits at the 98th percentile, roughly 130 on an SD-15 test, and has well over 100,000 members worldwide with local chapters almost everywhere. Triple Nine admits at the 99.9th percentile, roughly 146, and is a far smaller volunteer organization whose activity centres on its journal and online forums rather than in-person meetups.

Can I join a high-IQ society without taking a new test?

Often yes. Most of these societies accept documented prior scores from approved standardized tests, which is the main route for everything above Mensa. The catch is documentation: you need an official record from a qualified administrator, the test must be on that society's current accepted list, and old editions are sometimes excluded by date.

Do high-IQ societies cost money?

Yes, all of them charge annual dues, and the societies that run their own admission tests charge a testing or evaluation fee as well. The amounts are modest and vary by country and organization, but none of these groups is free to belong to. Dues typically fund the journal and administration rather than any benefit to the individual member.

Is membership in a high-IQ society worth putting on a résumé?

Generally no. Recruiters do not treat it as a qualification, and surveys of hiring opinion have long suggested it can read as self-promotional rather than impressive. The exception is a context where it is directly relevant — writing about psychometrics, for instance, or applying to something run by the society itself.

Why do different sources give different IQ numbers for the same society?

Because they are quoting different test scales. The 98th percentile is about 130 on a test with a standard deviation of 15, about 133 on an SD-16 test, and about 148 on the Cattell Culture Fair, which uses a much wider scale. All of those figures describe the same level of performance, so a society with one fixed percentile requirement can honestly be described with several different IQ numbers.

Summary

High-IQ societies form a fairly orderly ladder on paper: 98th percentile for Mensa, 99th for Intertel, 99.9th for Triple Nine and ISPE, 99.997th for Prometheus, one in a million for Mega. The numbers are all percentiles rather than scores, which is why the IQ figures quoted alongside them shift depending on the test scale being used.

The ladder gets less solid the higher you climb. Up to about the 99.9th percentile, admission rests on supervised tests with real norming samples. Above that, the instruments are unsupervised, self-normed and openly contested, and the decimal places promise a precision that no test can deliver. That does not make the societies pointless — plenty of members get exactly what they came for, which is a journal and a few dozen people worth arguing with. It does mean the thresholds are best read as club rules rather than as facts about the people inside.

---

Brambin is an entertainment app. Its IQ test is a game-like puzzle challenge, and the score it returns is a rough estimate from our own simple scoring model — not an accurate IQ, and with no medical or scientific basis. It is not a clinical assessment, it is not accepted as evidence for membership in Mensa, Intertel, the Triple Nine Society or any other organization, and it has no affiliation with any of them. Treat any online score, ours included, as a starting point for curiosity rather than a verdict.