How the AFQT score is calculated
The AFQT is computed from the standard scores of four ASVAB subtests: Arithmetic Reasoning, Mathematics Knowledge, Word Knowledge and Paragraph Comprehension. Word Knowledge and Paragraph Comprehension are first combined into a verbal score called VE, and the result is reported as a percentile from 1 to 99 against a 1997 reference group of 18 to 23 year olds.
Which subtests go into the AFQT?
Four subtests make up the AFQT: Arithmetic Reasoning (AR), Mathematics Knowledge (MK), Paragraph Comprehension (PC) and Word Knowledge (WK). The official scores page says AFQT scores "are computed using the Standard Scores from four ASVAB subtests" and then names those four (Understanding ASVAB Scores). The applicant FAQ says the same thing in plain language: scores from the four tests "are combined to compute your score" on the AFQT (official FAQ).
The other subtests (General Science, Electronics Information, Auto and Shop Information, Mechanical Comprehension and Assembling Objects) do not count toward the AFQT. They still matter, because the services use them in the composite scores that decide which jobs you can train for. See ASVAB line scores for that side of scoring.
Two subtests are math (AR and MK) and two are verbal (WK and PC). The Air Force describes the AFQT as "not a separate test" but a score derived from four subtests (AirForce.com ASVAB page). You never sit a separate AFQT section; your AFQT comes out of the same ASVAB session.
What is a standard score, and why does the AFQT start from it?
A standard score places your result on a fixed scale so that different test forms mean the same thing. On the ASVAB, each subtest is scored on a scale where "the mean is set to 50 and the standard deviation is set to 10" (ASVAB Scoring page). A standard score of 60 is one standard deviation above the mean, and a 40 is one below.
The scale is not a count of right answers. The same official page explains that scoring uses an Item Response Theory model called the three-parameter logistic (3PL) model. The model takes into account each question's difficulty, how well it separates stronger and weaker examinees, and the chance of guessing it. Your final ability estimate is then converted to a standard score through a process called equating, so scores "have the same meaning regardless of which item pool or test form the examinee receives" (ASVAB Scoring).
That is why the AFQT starts from standard scores rather than raw totals. Two people who answer the same number of questions correctly on the adaptive test can end up with different scores, because they saw questions of different difficulty. For more on the scale itself, read ASVAB standard scores explained.
How is the verbal score (VE) built from WK and PC?
Word Knowledge and Paragraph Comprehension are combined into one verbal score before they are used. The official military jobs page defines it this way: "VE is a verbal composite formed from an optimally weighted composite of unrounded WK and PC standard scores" (Military Jobs page, footnote 1).
Three details in that sentence matter. The word weighted means WK and PC are not simply averaged; each gets a weight, and the official page does not publish the weights. Unrounded means the calculation uses exact standard scores, so the whole numbers printed on a score report are not the precise inputs. And composite means VE is a combination of two subtests, not a third test you take.
The Air Force site gives a simpler description for applicants: PC and WK "are added together to form the Verbal Expression (VE) score," and on its MAGE page it says they "are combined first to determine a scaled Verbal Expression (VE) score" (AirForce.com ASVAB page). Read together, the two official sources agree that WK and PC are merged into a scaled verbal score; the DoD testing site gives the more exact wording.
How do AR, MK and VE combine into the AFQT?
The official sources state which scores go in: the AFQT is computed from the standard scores of AR, MK, PC and WK, with WK and PC entering through VE. The Air Force site lists the AFQT pieces as the VE score, the Mathematics Knowledge score and the arithmetic score (AirForce.com).
None of the official pages we checked publishes the exact weighting that turns VE, AR and MK into the AFQT, or the norm table that turns the result into a percentile. You may see a formula on prep sites. We do not repeat it here because we could not confirm it on an official page today, and because even with a formula you would still need the official norm table to get a percentile.
What you can take from the official wording is enough for planning:
- Both math subtests count.
- Both verbal subtests count, through VE.
- A weak score in one area pulls the composite down, so balanced preparation across all four is the safe plan.
If your verbal scores are much lower than your math scores, remember that WK and PC reach the AFQT only through VE. Improving either one moves VE, and VE feeds the AFQT. The Word Knowledge and Paragraph Comprehension pages have practice for both.
How does the composite become a percentile from 1 to 99?
The composite is compared with a reference group through a norm table. The official page says AFQT scores "are reported as percentiles between 1-99" and that the percentile shows "the percentage of examinees in a reference group" who scored at or below that score (Understanding ASVAB Scores).
The reference group is "a sample of 18 to 23 year old youth" who took the ASVAB in a national norming study in 1997. The norming page adds that the current norms were implemented in 2004, using a nationally representative sample of about 6,000 youths aged 18 to 23 from the Profile of American Youth 1997 study (Norming Information). A 2026 Office of People Analytics note says recent evaluations show the 1997 reference group and 2004 norms "are still representative" (OPA executive note, footnote 1).
So the last step is a lookup, not arithmetic you can repeat at home. Your composite is placed against the 1997 sample, and the percentile is the share of that sample who scored at or below you.
What does an AFQT percentile actually tell you?
It tells you where you stand against the reference group, and nothing about the share of questions you got right. The FAQ example: an AFQT of 62 means "you scored as well as or better than 62%" of the 1997 sample (official FAQ). The Air Force makes the same point about its scores: they are "not based on the number of questions you answered correctly" (AirForce.com).
Example: Your score report shows an AFQT of 62. What does it mean?
- The AFQT is a percentile, not a percent correct.
- A percentile of 62 means 62 percent of the 1997 reference group scored at or below your score.
- Check the category table on the AFQT categories page: 62 is inside the 50 to 64 range.
Answer: You scored as well as or better than 62 percent of the 1997 sample of 18 to 23 year olds, which places you in Category IIIA.
Keep the two ideas apart when you read practice results too. A practice set where you got 62 percent right says nothing about your AFQT percentile, because the AFQT compares you with people, not with a question count.
Why can't you calculate your own AFQT from practice tests?
Because three ingredients are missing from any practice test: the item statistics that produce a standard score, the VE weights, and the official norm table. The scoring page describes an adaptive process that starts every examinee at an ability estimate of 0.0 and updates it after each answer with a Bayesian procedure (ASVAB Scoring). Practice tests do not reproduce that.
The official AFQT Predictor Test (APT) shows how hard prediction is. It is a 20-item adaptive test that a recruiter can issue, and the official page still warns that its score "is not a guarantee" of your ASVAB result (APT page). Only the recruiter who registered you can see the APT score.
AptiStep reports accuracy on practice sets for the four AFQT subjects. It does not estimate a percentile, and you should not read a practice percentage as one. Use practice results to find weak skills, then follow a plan such as the study plan by subtest.
Does the test version change how the AFQT is calculated?
No. The same four subtests and the same reporting scale apply to the computer and paper versions. The FAQ says scores "are statistically linked" across the two methods, so you "would be expected to receive a similar score" either way (official FAQ). The researcher page explains that equating studies are run for every CAT-ASVAB item pool and every paper form (ASVAB Scoring).
What does differ is how unfinished questions are handled. On paper, unanswered items are treated as incorrect. On the CAT-ASVAB, a penalty is applied that equals the score you would get by guessing at random on the unfinished items (ASVAB Scoring). See CAT-ASVAB vs paper for strategy.
How should this change the way you study?
Study all four AFQT subjects, and spend extra time where your practice shows the most errors. Arithmetic Reasoning and Mathematics Knowledge each count directly. Word Knowledge and Paragraph Comprehension count through VE. No calculator is allowed on any version, so practice the hand methods in the math workout.
- Take a short diagnostic in each of the four subjects on the practice test page.
- Rank the four by error rate, not by how you feel about them.
- Work the weakest first, then rotate so none of the four goes untouched for more than a few days.
- Review every mistake using the method in how to review mistakes.
Then confirm with your recruiter which minimum and which line scores apply to the jobs you want. The minimum scores by branch page lists what each branch publishes.
Practice questions
Work each one on paper first, then open the explanation. These are original AptiStep items, not real ASVAB questions.
Which set of ASVAB subtests is used to compute the AFQT?
- General Science, Arithmetic Reasoning, Word Knowledge and Mechanical Comprehension
- Arithmetic Reasoning, Mathematics Knowledge, Word Knowledge and Paragraph Comprehension
- Word Knowledge, Paragraph Comprehension, Electronics Information and Mathematics Knowledge
- All ten ASVAB subtests, equally weighted
Show answer and explanation
Answer: B. Arithmetic Reasoning, Mathematics Knowledge, Word Knowledge and Paragraph Comprehension
The official scores page names four subtests for the AFQT: Arithmetic Reasoning, Mathematics Knowledge, Paragraph Comprehension and Word Knowledge. Science and technical subtests are used for job composites, not for the AFQT.
According to the official definition, how is the verbal score VE formed?
- It is a separate verbal test given after Word Knowledge
- It is the higher of the WK and PC scores
- It is an optimally weighted composite of unrounded WK and PC standard scores
- It is the number of WK and PC questions answered correctly
Show answer and explanation
Answer: C. It is an optimally weighted composite of unrounded WK and PC standard scores
The officialasvab.com military jobs page defines VE as an optimally weighted composite of unrounded WK and PC standard scores. It is not a separate test, not a maximum, and not a raw count of correct answers.
Jordan has an AFQT of 45. Which statement is correct?
- Jordan answered 45 percent of the questions correctly
- Jordan scored as well as or better than 45 percent of the 1997 reference group
- Jordan is in AFQT Category II
- Jordan scored below 45 percent of all current applicants
Show answer and explanation
Answer: B. Jordan scored as well as or better than 45 percent of the 1997 reference group
An AFQT score is a percentile against the 1997 sample of 18 to 23 year olds, so 45 means as well as or better than 45 percent of that group. It is not percent correct, and 45 falls in Category IIIB (31 to 49), not Category II.
A subtest standard score of 60 means what, on the ASVAB scale?
- One standard deviation above the mean of 50
- 60 percent of questions correct
- The 60th percentile on the AFQT
- Exactly average
Show answer and explanation
Answer: A. One standard deviation above the mean of 50
ASVAB subtest standard scores have a mean of 50 and a standard deviation of 10. Subtract the mean: 60 minus 50 is 10, and 10 divided by 10 is 1, so the score is one standard deviation above the mean.
Why can a practice test not give you your real AFQT percentile?
- Because practice tests use harder questions
- Because the AFQT is only given on paper
- Because the AFQT counts General Science twice
- Because the conversion uses official item statistics, VE weights and a norm table that practice tests do not have
Show answer and explanation
Answer: D. Because the conversion uses official item statistics, VE weights and a norm table that practice tests do not have
The official scoring uses an Item Response Theory model, equating to standard scores, a weighted VE composite and the 1997 norm table. None of those are reproduced by a practice test, so a practice percentage cannot be converted to an AFQT percentile.
Taylor did very well on Mechanical Comprehension. How does that subtest affect the AFQT?
- It raises the AFQT through VE
- It replaces the lowest AFQT subtest
- It does not count toward the AFQT, though it can count in job composites
- It counts only on the paper version
Show answer and explanation
Answer: C. It does not count toward the AFQT, though it can count in job composites
Only AR, MK, WK and PC are used for the AFQT. Mechanical Comprehension appears in service composites, for example the Air Force Mechanical score, so it can matter for job qualification but not for the AFQT itself.
On the CAT-ASVAB, what happens to items you do not finish before time runs out?
- They are skipped with no effect
- A penalty is applied equal to guessing at random on the unfinished items
- They are counted as correct
- You get extra time at the end of the session
Show answer and explanation
Answer: B. A penalty is applied equal to guessing at random on the unfinished items
The official scoring page says CAT-ASVAB examinees who run out of time get a penalty on their final ability estimate, equivalent to the score they would have received by guessing at random on the unfinished items.
Frequently asked questions
Is there an official AFQT formula I can use?
The official pages we checked list the four subtests and define VE, but they do not publish the AFQT weights or the norm table. See the official scores page and the military jobs page.
Do Word Knowledge and Paragraph Comprehension count as much as math?
They count through VE, which is a weighted verbal composite. The official pages do not publish the weights, so prepare both verbal and both math subtests.
Will my AFQT be different on paper and on computer?
It should be similar. The official FAQ says scores are statistically linked across the two versions. See CAT-ASVAB vs paper.
Can AptiStep tell me my AFQT?
No. AptiStep shows practice accuracy for the four AFQT subjects. It does not estimate an AFQT percentile or eligibility.
Sources
Checked November 14, 2026. AptiStep is an independent study site; confirm anything that affects your enlistment with a recruiter.
- Official ASVAB: Understanding ASVAB Scores
- Official ASVAB: ASVAB Scoring (researchers)
- Official ASVAB: Military Jobs (Service composites)
- Official ASVAB: FAQs
- Official ASVAB: Norming Information
- Office of People Analytics: Appropriate Use of ASVAB Scores (PDF on officialasvab.com)
- AirForce.com: ASVAB
- Official ASVAB: AFQT Predictor Test (APT)