The best personality tests in 2026, compared

How should you judge a personality test?

Four questions settle most of it. Is the model one that has been replicated across samples and languages? Does the test give the same answer when nothing has changed? Do its scores relate to real behaviour and outcomes? And does the result keep the information, or round you to a type? A test can be enjoyable and fail all four. The Big Five and HEXACO models have a replicated structure behind them and scores that relate to real outcomes (Goldberg, 1993; Roberts et al., 2007; Ashton & Lee, 2007), and any test that reports continuous scores keeps the information rather than rounding it.

We make Wellington, so read our entry with that in mind. The others are described as fairly as we can manage.

A fifth question is worth asking after the other four: can you use the result? An instrument can be well validated and still hand you a table of numbers that sits in a drawer, which is the gap most free scientific inventories leave open. If you want to see what you will be asked before you start, personality test questions sets out what the items look like and how to answer them.

The tests, compared

How strong is the evidence for the free Big Five inventories?

Strong, and unusually well documented, because the people who built them published the construction as well as the result. The BFI-2 was developed as a hierarchical instrument with three facets under each of the five domains, and the published development work reports its reliability, its factor structure and its predictive power against the earlier version it replaced (Soto & John, 2017). You can score it yourself with a key, which is both its charm and its limit.

The IPIP is a public-domain pool of personality items created so that researchers would stop paying licence fees to measure well-understood constructs, and it has been used and re-used for decades (Goldberg et al., 2006). The most convenient version for an individual is the IPIP-NEO-120, a 120-item inventory measuring all 30 facets, whose four-item facet scales showed acceptable reliability in two large internet samples and converged with the NEO PI-R and with acquaintance ratings (Johnson, 2014). If you want the deepest free facet report in existence, that is where it is. Free Big Five personality tests, compared puts the free options side by side.

What these instruments do not give you is a reading. You get numbers, a facet table and, at best, a paragraph of generic description per scale. For many people that is exactly right. For others it is the moment the interest dies. The Big Five, explained and the facets page are a decent substitute for the missing interpretation.

What about the HEXACO-PI-R and the VIA survey?

The HEXACO-PI-R measures six factors rather than five, and the extra one is the point. Lexical studies conducted from scratch in seven languages produced a six-factor solution in which Honesty-Humility stood as a factor of its own rather than folding into Agreeableness (Ashton et al., 2004). The inventory built on that structure has published psychometric properties: its 24 facet scales were internally consistent and fitted the six-factor structure in a validation sample of over 400 (Lee & Ashton, 2004).

The case for bothering with the sixth factor is that it is not a restatement of the fifth. A review of research on Honesty-Humility sets out why it is distinguishable from HEXACO Agreeableness, and reports that Five-Factor Model instruments capture only part of the Honesty-Humility variance (Ashton et al., 2014). If you are choosing an inventory and you care about integrity as well as temperament, that is the reason to prefer six factors. What is Honesty-Humility covers the trait itself.

The VIA Survey measures something different: 24 character strengths organised under six virtues, set out in a handbook that defined each strength, its criteria and its measurement (Peterson & Seligman, 2004). Several strengths, notably hope, zest, gratitude, curiosity and love, have been linked with higher life satisfaction (Park et al., 2004), and a study of data from 75 nations found that the rank order of strengths is broadly similar from country to country (McGrath, 2015). A meta-analysis of interventions that ask people to use their signature strengths found small to moderate gains in life satisfaction and positive affect, from a small evidence base of 29 effect sizes (Schutte & Malouff, 2019).

The honest caveat is that a ranked list is not a measurement against other people in the way a percentile is. Read it as an ordering of what is strongest in you, which is what it is good at.

Which test is best for what?

• For a free, quick, reliable read of the broad dimensions: the BFI-2 or the Wellington Snapshot. Both take under ten minutes. Wellington adds a written page per dimension and a sixth dimension, Honesty-Humility.
• For the most detailed free facet scores and you do not mind reading a table: the IPIP-NEO 300.
• For Honesty-Humility specifically: the HEXACO-PI-R, or Wellington, which includes it in the free Snapshot.
• For character strengths: the VIA Survey, or the Wellington Membership, which measures all 24 alongside the Big Five facets.
• For a written report that reads like it was written for you, and a plan to act on it: the Wellington Membership. This is the gap we built Wellington to fill; the free inventories above give you numbers, not pages.
• For a memorable label to talk about with friends: 16Personalities. Enjoy it; do not make decisions on it.

How do you choose in two minutes?

Answer three questions in order and stop at the first one that decides it.

• What is the result for? If someone else will act on it, a hiring panel or a school, stop. None of these instruments is built for that, and the MBTI's own publisher says so about its own test.
• How much time do you have? Under ten minutes points to the BFI-2 or the free Wellington Snapshot. Half an hour opens up the IPIP-NEO-120 and the HEXACO-PI-R, which are the deepest free options.
• Do you want numbers or a reading? A table of percentiles is what the free scientific inventories hand you, and it is enough if you are comfortable interpreting scores. If you want the interpretation written for your band, that is what Wellington sells.

Two shortcuts are worth knowing. If a test gives you a four-letter code, treat the result as a conversation rather than a measurement. And if it never says what it compared you with, the number cannot mean much; how to read personality test scores explains why a percentile without a reference sample is not a percentile.

If your real question is what is my personality type, the better version of it is where you sit on each of six traits, and any of the trait inventories above will answer that. The Big Five personality test covers how to take one and how to read the result, and how accurate are personality tests covers how much to trust any single sitting.

What about Wellington?

Wellington is a science-based personality assessment from Therabot Labs LLC. It measures personality as continuous traits, built on the Big Five and HEXACO models, and writes the results back as a personal report rather than a type.

6 dimensions in the free Snapshot and, with the membership, 90 traits growing toward all 255, from a bank of 1,605 questions. In the written report, every trait has three pages, one per band, and your answers choose which you read. Read the methodology before deciding anything. Wellington is a wellness tool for reflection, not a medical or psychological diagnosis, and it does not replace care from a qualified professional.

The Snapshot is free and takes about 7 minutes. The first ten questions are answered before anything is asked of you; after question ten you sign in with an email so your answers are saved and your results can be shown. No card is needed. Every trait is reported as a percentile against the reference sample, in one of three bands, Quiet below the 30th percentile, Balanced to the 70th, High above. The bands and the Portrait names exist to make the result easier to talk about; the percentile is the measurement, and it is always shown alongside. Our norms are provisional until the norming sample is complete, which the methods page states plainly.

Where we differ from the free inventories is not the model, which is theirs, but what happens after the scoring. Where we differ from the type tests is that nothing is rounded. Whether that is worth anything to you depends on which of the three questions above decided your choice.

Questions people ask

Is there a free personality test that is actually scientific?

Yes, several. The BFI-2 measures five domains and fifteen facets, and its development and validation are published in full (Soto & John, 2017). The IPIP-NEO is built from public-domain items with decades of use behind them (Goldberg et al., 2006), and the HEXACO-PI-R adds a sixth factor with published psychometric properties (Lee & Ashton, 2004). The Wellington Snapshot is free and uses the same models, with a band and a percentile for each dimension.

Is the MBTI scientifically valid?

Partly. Its four scales relate to four of the Big Five dimensions, so the questionnaire is measuring something real. The same study found no support for the view that those scales measure dichotomous preferences or distinct types: they are continuous dimensions cut at a midpoint (McCrae & Costa, 1989). A critical review concludes the four-letter formula does not support the inferences commonly drawn from it (Pittenger, 2005). Treat the letters as a vocabulary rather than a measurement.

How long should a good personality test be?

Long enough to ask several statements for each trait it reports. Very abbreviated scales, single-item ones worst of all, understate how much traits matter and inflate the apparent importance of rival constructs, and slightly longer scales are substantially better for little extra time (Credé et al., 2012). The authors of the ten-item Big Five measure say plainly that it should be used only where a longer instrument will not fit (Gosling et al., 2003). Sixty questions for five or six broad dimensions is a reasonable floor.

Which free test gives the most detail?

The IPIP-NEO-120 or its 300-item version. Both are public domain and report all 30 facets beneath the five domains, and the 120-item version's facet scales reached acceptable reliability in two large internet samples (Johnson, 2014). The trade is time and presentation: you get a thorough table and very little interpretation of what the numbers mean for you.

Which test measures character strengths rather than traits?

The VIA Survey, which ranks 24 strengths defined in a published classification (Peterson & Seligman, 2004). Several of them, including hope, zest and gratitude, relate to life satisfaction (Park et al., 2004), and interventions built on using signature strengths show small to moderate gains in life satisfaction and positive affect (Schutte & Malouff, 2019). Wellington measures all 24 with the membership.