AQ-10 test: free autism screening in about 2 minutes

The AQ-10 (Autism Quotient-10) is the shortest validated autism screening tool in widespread clinical use. Developed as a rapid-screening version of the full AQ-50, it's used by GPs and clinicians to quickly determine whether a patient should be referred for a full autism assessment.

If you're looking for more depth after the AQ-10, we also offer the full AQ-50 (50 questions) and the RAADS-R (80 questions designed specifically for adults). All screenings are free, instant, and completely private.

Common reasons people take the AQ-10

People arrive at this page from a handful of common starting points. Some have just watched a video or read a post that described an experience they recognized in themselves for the first time. Some are preparing for a conversation with a doctor and want a concrete number to open with rather than a vague description of years of feeling different. Others have a family member, partner, or friend who was recently diagnosed and are wondering whether some of the same patterns apply to them. A smaller group are clinicians or students who want to see how the instrument works before using it professionally. None of these reasons is more or less valid than another; the AQ-10 was built to be a fast, low-stakes first step for exactly this range of situations.

Notes on answering the questions

Answer for how you usually are, not for your best day or your most put-together version of yourself. It can help to picture a typical week rather than a recent unusual event, whether that event went unusually well or unusually badly. If a question describes something you can technically do but only with noticeable effort, such as reading someone's expression by consciously analyzing it rather than reading it automatically, answer based on how effortful it actually feels, not on whether you can eventually produce the right response.

This matters especially if you have spent years masking. Masking can make a question like "I find it easy to work out what someone is thinking or feeling" harder to answer honestly, because the outward result, appearing to understand, may look successful even when it took considerable conscious effort to get there. Where that applies to you, try to answer based on the effort and internal experience behind the behavior rather than the polished result other people see.

What the AQ-10 Is

The AQ-10 was published in 2012 by Carrie Allison, Bonnie Auyeung, and Simon Baron-Cohen at the Autism Research Centre, University of Cambridge, in a paper titled "Toward brief 'red flags' for autism screening" in the Journal of the American Academy of Child and Adolescent Psychiatry. The 10 items are drawn directly from the 50-item Autism-Spectrum Quotient, originally published by Baron-Cohen and colleagues in 2001. Rather than build a new instrument, the authors identified the 10 questions from the AQ-50 that carried the most weight in distinguishing autistic from non-autistic respondents, so the AQ-10 could be used where a full 50-question form was not practical, such as a short GP appointment. Respondents answer each item on an agree or disagree scale, and the AQ-10 remains the version referenced in NICE guidance for adult autism referral in the UK.

The AQ-10 covers four broad areas in its 10 questions: social awareness (reading situations and other people's intentions), conversational patterns (turn-taking and reading between the lines), imagination and flexibility (preference for facts over pretend scenarios, need for routine), and attention (strong focus on detail alongside difficulty switching tasks). Each area is represented by only one or two questions, which is what makes the AQ-10 fast, and also why it cannot capture the same texture as a longer instrument built around the same domains.

How We Score It

Our version presents each of the 10 items on a four-point scale: Definitely Agree, Slightly Agree, Slightly Disagree, and Definitely Disagree. Four of the questions are scored in the "agree" direction, meaning choosing Definitely Agree or Slightly Agree earns a point. The other six are scored in the "disagree" direction, meaning choosing Slightly Disagree or Definitely Disagree earns a point. Every other answer earns nothing. Your total is a whole number between 0 and 10, and our tool marks a score of 6 or above as meeting the threshold. This is the same threshold recommended in the original 2012 paper, so on the AQ-10 our implementation and the published research agree.

What the Original Study Found

Allison, Auyeung, and Baron-Cohen built the AQ-10 specifically to work as a rapid screen ahead of a possible referral, not as a stand-alone diagnostic step. Their paper recommends a score of 6 or above as the point at which further assessment is warranted, and that recommendation has since been adopted into NICE clinical guidance for identifying adults who may need a full autism evaluation. The instrument was built by selecting the 10 highest-weighted questions out of the original 50, so it trades some of the AQ-50's detail for speed.

The paper's authors were explicit that the AQ-10 was meant to sit at the front of a referral pathway, not to replace it. A high score signals that the fuller picture, gathered through a proper assessment, is worth pursuing; it was never intended to function as a pass or fail result on its own, and NICE guidance treats it the same way, as a prompt to refer rather than a diagnostic checkpoint.

What It Cannot Tell You

• It is self-report. Your score depends on how accurately you can observe and describe your own patterns, which is harder for people who have spent years masking those patterns from themselves as well as others.
• It is not a diagnosis. No score, high or low, can substitute for a clinical evaluation.
• It does not take a developmental history. A formal diagnosis looks at how traits have shown up since childhood and how much they affect daily functioning; 10 questions answered today cannot capture that.
• It does not rule out or confirm other conditions. Anxiety, ADHD, OCD, and several other conditions can produce similar answers on some of these items.
• Ten questions is a narrow slice of a much larger picture. The AQ-10 was built for speed, and speed comes at the cost of depth compared with the AQ-50 or RAADS-R.
• It cannot separate a lifelong pattern from a temporary one. Someone going through a stressful period may answer several items in the autistic-trait direction without that reflecting how they have generally functioned across their life; a proper evaluation looks specifically at how long traits have been present.

None of this makes the AQ-10 useless; it makes it exactly what it was built to be, a fast triage step rather than a verdict. Treat the number it produces as a prompt for further reading or a conversation with someone you trust, not as a label to carry away on its own or repeat to others as if it were a diagnosis.

Who the AQ-10 Is For

The AQ-10 was designed for adults and is most useful as a first step: something you take in a few minutes to decide whether a longer look is worthwhile. It suits people who want a quick gut-check before committing to an 80-question instrument like the RAADS-R, and it is the same short form clinicians use before deciding on a referral. It is not designed for young children; if you are screening a toddler, see the M-CHAT-R-based screening for kids instead.

It also tends to suit certain situations better than others. Someone who has never considered the possibility of autism before and just wants a rough read on whether the idea is worth exploring at all will get more value from the AQ-10 than from wading through a much longer form on a first visit. The same is true for someone preparing to bring up autism with a doctor for the first time; a short, concrete score can be an easier opening than a vague description of years of experiences. On the other hand, if you have already spent time reading about autism and recognize many of your own patterns in the descriptions on this page, you may find that 10 questions feel too thin to capture what you already suspect, in which case moving straight to the AQ-50 or RAADS-R may be more satisfying and more informative.

Parents, partners, and friends sometimes also take the AQ-10 on behalf of someone else out of curiosity. We would caution against treating a second-hand score as meaningful. The questions ask about internal experiences, such as how effortful conversation feels or how a person interprets someone else's intentions, that an outside observer usually cannot answer with any confidence. The AQ-10 works best as a genuinely first-person self-report.

How accurate is the AQ-10?

The honest answer is that "accurate" is not really the right word for what a 10-item screener does. The AQ-10 was designed by Allison, Auyeung, and Baron-Cohen (2012) to make referral decisions faster and more consistent, not to stand in for a diagnosis. It was built by picking the 10 items from the AQ-50 that separated autistic from non-autistic respondents most reliably in that study's sample, and NICE guidance in the UK recommends it for exactly that purpose: deciding whether an adult should be referred for a full autism assessment. We do not publish sensitivity or specificity percentages for our implementation of the AQ-10, because those figures come from the original clinical study's specific sample and would misrepresent what any individual person's 2-minute result can promise them.

A short instrument trades depth for speed by design. Ten questions cannot capture the full texture of a lifelong pattern the way an 80-item instrument or a multi-hour clinical interview can, and the AQ-10's authors never claimed otherwise. What it can do reliably is give you a fast, structured read on whether the topic is worth pursuing further, which is exactly the narrow job it was built for.

What a "Meets" or "Below" Result Means

A score of 6 or above meets our threshold, which mirrors the referral threshold in the original paper. That result means your answers land in the range where autistic traits are commonly reported, and a next step, such as the longer AQ-50, the RAADS-R, or a conversation with a professional, is worth considering. A score below 6 means your answers did not land in that range on this occasion. It does not rule autism out, especially if you recognize yourself in the descriptions on this page but answered in a way that reflects how you present rather than how you feel internally. Either way, see how we explain results for more detail on what each band does and does not mean, and read our method for the full picture of how this screening was built.

Recommended next steps

Whatever your score, treat it as a starting point rather than a conclusion. If your score meets our threshold, a reasonable next step is one of the longer instruments on this site, either the AQ-50 for a domain-by-domain picture or the RAADS-R if you want the most detailed self-assessment available here, followed by a conversation with a professional if you want a formal evaluation or access to accommodations. If your score is below the threshold but the descriptions on this page still feel familiar, the same next step applies. A single short screening is not built to be the last word on the question, and it is common for people who mask well to score lower than their lived experience would suggest.

In either case, it can help to write down which specific questions felt most true for you before you move on, since that gives you something concrete to bring into a longer screening or a conversation with a doctor, rather than trying to reconstruct the feeling from memory later. See how we explain results for more detail on what each band across every screening on this site does and does not mean.

Frequently Asked Questions