The ten ASVAB subtests: questions and time limits
The computer version (CAT-ASVAB) given at a MEPS or MET site, in the order it is given: 135 scored questions in 197 minutes without tryout questions. This practice test uses these lengths and limits. The four marked AFQT make your AFQT score; all ten feed the scores the services use for jobs. The paper version, given at a few MET sites, is longer: 225 questions in 149 minutes.
| Subtest | Questions | Time limit |
|---|---|---|
| General Science | 15 | 12 min |
| Arithmetic ReasoningAFQT | 15 | 55 min |
| Word KnowledgeAFQT | 15 | 9 min |
| Paragraph ComprehensionAFQT | 10 | 27 min |
| Mathematics KnowledgeAFQT | 15 | 31 min |
| Electronics Information | 15 | 10 min |
| Auto Information | 10 | 7 min |
| Shop Information | 10 | 6 min |
| Mechanical Comprehension | 15 | 22 min |
| Assembling Objects | 15 | 18 min |
Source: officialasvab.com, What to Expect, read October 3, 2026. Tryout questions, which do not count, are added to two to four subtests with extra time.
What Mechanical Comprehension covers
Levers, pulleys, gears, ramps, wheel and axle, screws, hydraulics, springs, structures, materials, motion, fluids and pressure. Every diagram is drawn from the numbers in its question, and every calculation is checked by a script.
It does not count toward your AFQT. It feeds the line scores and composites each service uses to match you to jobs: see how the Army, Navy, Air Force and Marine Corps use it, or take the full ASVAB.
How the AFQT is scored, and why a practice test cannot score it
The AFQT is not a separate test. It is computed from four of the ASVAB's ten subtests. DoD's technical bulletin gives the formula as AR + MK + 2 × VE, where VE is a verbal score that combines Word Knowledge and Paragraph Comprehension. That sum is converted to a percentile from 1 to 99 against a national sample of about 6,000 people aged 18 to 23 who took the test in 1997. An AFQT of 50 means you scored as well as or better than half of that sample.
Each subtest score behind it comes from a statistical model that weighs how hard each question was, not from a count of right answers, and the computer test picks harder or easier questions as you go. A practice form cannot reproduce that. So this page gives you what it can measure honestly, your results by subtest and topic, and, if you ask for it after the full practice test, a range where a result like yours would most likely fall.
How our range is built. Each subtest's percent correct is placed against the published averages for the official paper ASVAB forms 23 to 27 as applicants took them (DoD Technical Bulletin No. 4, Tables 6.1 and 6.2, put on the 1997 scale with Segall's Table 3.9), assuming our versions are about as hard as an official form. Word Knowledge and Paragraph Comprehension are combined into VE with the official weights, the four are added with the official formula, and the total is read in the official percentile table (Segall 2004, Table 2.5). The range then covers three sources of error: a 55-question test, questions that have never been calibrated on real test-takers, and the real test's own measurement error. It is an 80 percent interval under those assumptions, never checked against real scores, so in the middle of the scale it spans about 20 points either way.
Sources: officialasvab.com, Understanding your scores; ASVAB Technical Bulletin No. 4, section 2.2.1; Segall (2004), Development and Evaluation of the 1997 ASVAB Score Scale.
AFQT categories and service minimums
| Category | AFQT percentile |
|---|---|
| I | 93 to 99 |
| II | 65 to 92 |
| IIIA | 50 to 64 |
| IIIB | 31 to 49 |
| IVA | 21 to 30 |
| IVB | 16 to 20 |
| IVC | 10 to 15 |
| V | 1 to 9 |
Minimums the services publish on their own sites (October 3, 2026): Army 31; Air Force 31 for high school seniors or graduates and 50 with a GED; Marine Corps 31, or 50 with a GED or nontraditional diploma; Coast Guard 32. The Navy and Space Force recruiting pages give no single number. Under DoDI 1145.01, no one in Category V may enlist, and no more than 4 percent of a year's recruits may come from Category IV. Each job then sets its own line scores from the full ASVAB: see every Army, Navy, Air Force and Marine Corps job and its minimum.
How to study so it sticks
- Test yourself, do not just reread. Pulling an answer from memory is one of the two methods rated most useful in a 2013 review of ten study techniques (Dunlosky and colleagues, Psychological Science in the Public Interest). That is what a practice test is for.
- Space it out. The same review rates spreading practice over several days as highly useful. Redo your missed questions two or three days later, then take a new version.
- Mix topics. Practicing different kinds of problems together feels harder, and in a 2020 classroom study (Rohrer and colleagues) students who did so scored higher on a later test than those who practiced one kind at a time.
- Learn word parts. Many words on Word Knowledge share Latin and Greek roots. One root can unlock several words you have never seen.
- Practice math by hand. No calculator on test day, so none here. Write every step and check your answer against the question before moving on.
- For the technical subtests, learn how things work. Most questions ask why a part or rule does what it does, so a rule you understand (more rope strands, less pull) carries over to questions you have never seen.
- Sleep. A short review and a normal night's sleep will do more than a late cram.
The ASVAB testing program endorses no preparation method beyond school coursework in math, English and science; these are general study findings, not official advice.
Free official practice, and study guides
The testing program names three legitimate ways to prepare: school coursework in math, English and science; its free sample questions at officialasvab.com and asvabprogram.com; and the Army's free March2Success, which asks you to register and states that it shares nothing with recruiters unless you are already in the Army's Future Soldier Program. It also warns that no one selling preparation has real test questions, and that guaranteed scores are false claims.