AQL Calculator
Enter your order size and get the exact number of garments to inspect — and how many critical, major and minor defects mean pass or fail — straight from the ANSI Z1.4 / ISO 2859-1 standard.
Total pieces in the order (one lot)
AQL for each defect class
Pass Fail- Sample
- 80
- Accept
- ≤ 0
- Reject
- ≥ 1
Zero tolerance: one found anywhere in the sample rejects the order.
- Sample
- 80
- Accept
- ≤ 5
- Reject
- ≥ 6
AQL 2.5: major defects, apparel standard.
- Sample
- 80
- Accept
- ≤ 7
- Reject
- ≥ 8
AQL 4.0: minor defects, apparel standard.
Pull 80 pieces at random, across cartons, sizes and colours, and check each one. Count critical, major and minor defects separately. The order passes only if every class is at or below its accept number; it fails (rework or re-inspect) as soon as any class reaches its reject number.
This is a normal inspection plan, single sampling, as in ISO 2859-1 and ANSI/ASQ Z1.4. The standards’ switching rules are not modelled here: after 2 rejected lots in 5 or fewer in a row they move you to tightened inspection, and a steady run of passes can earn reduced inspection.
Based on ANSI/ASQ Z1.4 · ISO 2859-1 (tables from MIL-STD-105E), single sampling, normal inspection.
Quick answer: how to work out an AQL sample size
Find your order quantity in Table I at your inspection level to get a code letter, then read that letter’s sample size and accept and reject numbers in Table II-A. Example: 2,000 pieces at Level II is code K, so inspect 125 garments and pass on 7 or fewer major defects (AQL 2.5) and 10 or fewer minor ones (AQL 4.0).
What this tells you
You almost never inspect a whole clothing order — you inspect a random sample and decide from that. The AQL (acceptance quality limit) is the worst average defect rate you agree to tolerate, and together with your order size and inspection level it sets the pass and fail numbers above: check the sample, count the defects, and accept or reject the batch against them. It’s the same method professional inspection firms use, drawn from the American standard ANSI/ASQ Z1.4 and its international counterpart, ISO 2859-1.
The apparel standard: 2.5 and 4.0
For clothing, inspections almost always run at General Inspection Level II with two limits on the same sample:
- AQL 2.5 for major defects — faults a customer would notice and return: broken seams, wrong measurements, holes, bad stitching, visible stains.
- AQL 4.0 for minor defects — small imperfections that don’t affect wear or sale, like a stray thread. More of these are tolerated.
- Critical defects — anything unsafe (a broken needle left in a garment) is usually set to zero: none allowed, ever.
The calculator above sets all three at once, with critical defects not allowed, major at 2.5 and minor at 4.0. Change any class’s AQL in its own row to match your purchase order.
Defect classes in garments, with examples
An AQL only means something once both sides agree which class a fault belongs to. These are the usual classes for knitwear and cut-and-sew clothing. Write your own list for each product, with photos of borderline cases, and attach it to the purchase order.
| Class | Usual limit | Examples in clothing |
|---|---|---|
| Critical | Not allowed (0) | A broken needle or other metal left in the garment; sharp points or edges on trims; cords or small parts that break children’s-wear safety rules; a missing care or flammability label where the law requires one. |
| Major | AQL 2.5 | Open or broken seams; holes; wrong colour or a shade off the approved lab dip; measurements out of tolerance; a misplaced or cracked print; visible stains. |
| Minor | AQL 4.0 | Loose or untrimmed threads; small marks that clean off; slightly uneven stitching on an inside seam; a crooked hangtag. |
On our floor every garment is judged against the sealed pre-production sample, so a “wrong shade” or “out of tolerance” call is a comparison, not an opinion. See how inspection runs on our floor below, and our guide to quality control and AQL for the checkpoints before final inspection.
Accept and reject numbers at Level II
One sample is checked against both limits. Count the major defects and compare them with the AQL 2.5 columns; count the minor defects separately and compare them with the AQL 4.0 columns. The order passes only if it passes both. A common misreading is that AQL 2.5 allows 2.5% of the order to be faulty. It does not: at 1,000 pieces the plan accepts up to 5 majors in a sample of 80, and the AQL describes the supplier’s long-run average, not a percentage of one lot.
| Order size | Sample | Majors (AQL 2.5) | Minors (AQL 4.0) |
|---|---|---|---|
| 91 – 150 pieces | 20 | accept 1, reject 2 | accept 2, reject 3 |
| 151 – 280 pieces | 32 | accept 2, reject 3 | accept 3, reject 4 |
| 281 – 500 pieces | 50 | accept 3, reject 4 | accept 5, reject 6 |
| 501 – 1,200 pieces | 80 | accept 5, reject 6 | accept 7, reject 8 |
| 1,201 – 3,200 pieces | 125 | accept 7, reject 8 | accept 10, reject 11 |
| 3,201 – 10,000 pieces | 200 | accept 10, reject 11 | accept 14, reject 15 |
| 10,001 – 35,000 pieces | 315 | accept 14, reject 15 | accept 21, reject 22 |
How much protection a plan gives: the chance a faulty order passes
Every sampling plan has an operating characteristic: the chance it accepts a lot, given the share of garments in that lot that really are faulty. The table works that out for the major-defect plans at AQL 2.5, Level II, for four common order sizes. A lot running at the AQL itself passes almost every time, and that is by design: the plan is built not to reject a supplier who is working to the level you agreed.
The surprise is in the middle. An order in which 1 garment in 20 has a major fault, twice the AQL, still passes between 58% and 79% of the time. Larger samples separate good lots from bad ones more sharply: with 8% of garments faulty, the 80-piece plan for a 1,000-piece order passes 38% of the time, and the 200-piece plan for a 5,000-piece order passes 7%.
| Order size (Level II plan) | 1% faulty | 2.5% faulty | 5% faulty | 8% faulty | 10% faulty |
|---|---|---|---|---|---|
| 281 – 500 pieces (sample 50, accept 3) | over 99% | 96% | 76% | 43% | 25% |
| 501 – 1,200 pieces (sample 80, accept 5) | over 99% | 98% | 79% | 38% | 18% |
| 1,201 – 3,200 pieces (sample 125, accept 7) | over 99% | 99% | 71% | 21% | 6% |
| 3,201 – 10,000 pieces (sample 200, accept 10) | over 99% | 99% | 58% | 7% | under 1% |
So use the plan for what it does well, and add the checks it cannot replace. For a first order, or a launch you cannot move, choose Level III: at 1,000 pieces it raises the sample to 125 with accept 7, and the chance of passing a lot with 8% faulty garments falls from 38% to 21%. Keep the in-line checks, which find a systematic fault while it is still a few hundred pieces. And judge a supplier over several lots, not one: the standard moves to tightened inspection when 2 of 5 or fewer consecutive lots are rejected on first inspection (MIL-STD-105E, clause 4.7.1), which the calculator above does not model.
How the calculator works, and what it leaves out
- Tables. The plans come from Table I and Table II-A of MIL-STD-105E, which carry the same values as ANSI/ASQ Z1.4 and ISO 2859-1 for single sampling under normal inspection. A test script checks every Level II plan, for all eleven AQLs, and every code letter against the printed tables.
- Arrows. Where the table shows an arrow instead of numbers, the calculator follows it to the first plan in that direction and uses that plan’s sample size, as the standard requires.
- Small orders. If a plan’s sample is as large as the order, every piece is inspected.
- Three classes, one sample. Arrows can give each class a different sample size. You pull the largest once; each class keeps its own accept and reject numbers and is counted within its own sample, the first pieces you check.
- “Not allowed” means accept 0, reject 1 on the whole sample. That is a buyer’s zero-tolerance rule, not a plan from the tables.
- Not included: tightened and reduced inspection and the switching rules between them, double and multiple sampling, and lots that mix styles or production runs. Your order quantity should be one lot.
Print the plan, share it or use it in Excel
Print / Save PDF turns the result into a one-page inspection sheet: the plan for each class, a tally grid with the reject number marked, a defect log, and lines for the inspector, the date and the supplier. Copy link gives an address with your inputs in it, so a colleague or a third-party inspector opens exactly the same plan. The Excel version holds the same tables with live formulas: type the quantity, pick the level and the three AQLs from drop-downs, then enter the defects found and it shows accept or reject. It has no macros and opens in Excel, Google Sheets and LibreOffice. All three are free, with no sign-up.
How inspection runs on our floor
Quality control at Collective Studio is led by Walter Meyer, who has 38 years in the trade. Knitwear is checked while it is sewn, not only at the end: in-line checks catch a systematic fault at 200 pieces rather than 5,000. Final inspection to an agreed AQL is part of the price. At 150 pieces every unit is inspected against the sealed pre-production sample; at bulk volumes we apply the AQL sampling you specify in the purchase order, and the balance is paid against that inspection report before anything ships.

Sample size at a glance (Level II, AQL 2.5)
- 51 – 90 pieces
- inspect 20
- 91 – 150 pieces
- inspect 20
- 151 – 280 pieces
- inspect 32
- 281 – 500 pieces
- inspect 50
- 501 – 1,200 pieces
- inspect 80
- 1,201 – 3,200 pieces
- inspect 125
- 3,201 – 10,000 pieces
- inspect 200
- 10,001 – 35,000 pieces
- inspect 315
Full sample size table (S-1 to S-4, Levels I–III)
Every lot size, at every inspection level in Table I, with its code letter and the sample size Table II-A prints beside that letter. At some AQLs an arrow moves the plan to a larger or smaller sample; the calculator above follows those arrows for you, and the second table shows where they are.
| Lot size | S-1 | S-2 | S-3 | S-4 | I | II | III |
|---|---|---|---|---|---|---|---|
| 2 – 8 | 2 A | 2 A | 2 A | 2 A | 2 A | 2 A | 3 B |
| 9 – 15 | 2 A | 2 A | 2 A | 2 A | 2 A | 3 B | 5 C |
| 16 – 25 | 2 A | 2 A | 3 B | 3 B | 3 B | 5 C | 8 D |
| 26 – 50 | 2 A | 3 B | 3 B | 5 C | 5 C | 8 D | 13 E |
| 51 – 90 | 3 B | 3 B | 5 C | 5 C | 5 C | 13 E | 20 F |
| 91 – 150 | 3 B | 3 B | 5 C | 8 D | 8 D | 20 F | 32 G |
| 151 – 280 | 3 B | 5 C | 8 D | 13 E | 13 E | 32 G | 50 H |
| 281 – 500 | 3 B | 5 C | 8 D | 13 E | 20 F | 50 H | 80 J |
| 501 – 1,200 | 5 C | 5 C | 13 E | 20 F | 32 G | 80 J | 125 K |
| 1,201 – 3,200 | 5 C | 8 D | 13 E | 32 G | 50 H | 125 K | 200 L |
| 3,201 – 10,000 | 5 C | 8 D | 20 F | 32 G | 80 J | 200 L | 315 M |
| 10,001 – 35,000 | 5 C | 8 D | 20 F | 50 H | 125 K | 315 M | 500 N |
| 35,001 – 150,000 | 8 D | 13 E | 32 G | 80 J | 200 L | 500 N | 800 P |
| 150,001 – 500,000 | 8 D | 13 E | 32 G | 80 J | 315 M | 800 P | 1250 Q |
| 500,001+ | 8 D | 13 E | 50 H | 125 K | 500 N | 1250 Q | 2000 R |
Table II-A: accept and reject numbers, normal inspection
| Code | Sample | 0.065 | 0.10 | 0.15 | 0.25 | 0.40 | 0.65 | 1.0 | 1.5 | 2.5 | 4.0 | 6.5 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| A | 2 | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | 0 / 1 |
| B | 3 | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | 0 / 1 | ↑ |
| C | 5 | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | 0 / 1 | ↑ | ↓ |
| D | 8 | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | 0 / 1 | ↑ | ↓ | 1 / 2 |
| E | 13 | ↓ | ↓ | ↓ | ↓ | ↓ | ↓ | 0 / 1 | ↑ | ↓ | 1 / 2 | 2 / 3 |
| F | 20 | ↓ | ↓ | ↓ | ↓ | ↓ | 0 / 1 | ↑ | ↓ | 1 / 2 | 2 / 3 | 3 / 4 |
| G | 32 | ↓ | ↓ | ↓ | ↓ | 0 / 1 | ↑ | ↓ | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 |
| H | 50 | ↓ | ↓ | ↓ | 0 / 1 | ↑ | ↓ | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 |
| J | 80 | ↓ | ↓ | 0 / 1 | ↑ | ↓ | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 |
| K | 125 | ↓ | 0 / 1 | ↑ | ↓ | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 | 14 / 15 |
| L | 200 | 0 / 1 | ↑ | ↓ | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 | 14 / 15 | 21 / 22 |
| M | 315 | ↑ | ↓ | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 | 14 / 15 | 21 / 22 | ↑ |
| N | 500 | ↓ | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 | 14 / 15 | 21 / 22 | ↑ | ↑ |
| P | 800 | 1 / 2 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 | 14 / 15 | 21 / 22 | ↑ | ↑ | ↑ |
| Q | 1250 | 2 / 3 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 | 14 / 15 | 21 / 22 | ↑ | ↑ | ↑ | ↑ |
| R | 2000 | 3 / 4 | 5 / 6 | 7 / 8 | 10 / 11 | 14 / 15 | 21 / 22 | ↑ | ↑ | ↑ | ↑ | ↑ |
Want the reasoning behind these numbers and how a real inspection runs? Read our guide to quality control & AQL for bulk apparel. Planning a production run and want it inspected properly? Start a project with us — every order ships with photos at each stage and inspection built in.
Sources
- MIL-STD-105E, Sampling Procedures and Tables for Inspection by Attributes (10 May 1989), US Department of Defense ASSIST database: clauses 3.15, 3.17, 4.2, 4.3, 4.7.1 and Table II-A
- NIST/SEMATECH e-Handbook of Statistical Methods, 6.2.3.2 Choosing a Sampling Plan with a given OC Curve (binomial probability of acceptance)
What clients say
Best supplier! On time, reliable and perfect craft.
Collective Studios under the direction of Mr Jamish and his daughter Homaira represent the future of RTW production in Bangladesh. A thorough approach to detail governed by a genuine standard of quality. Masters of and partners in the process!
Frequently asked questions
What is AQL in clothing manufacturing?
AQL stands for acceptance quality limit. The older US military standard behind the ISO 2859-1 tables, MIL-STD-105E, defines it as the quality level that, over a continuing series of lots, is the limit of a satisfactory process average: in practice, the highest average defect rate, in percent, that you are prepared to accept from a supplier over a run of orders. It is not the number of defects allowed in one sample; that is the acceptance number, which the tables set from three inputs: your order size, the inspection level and the AQL itself. In practice, an inspector pulls a set number of garments at random, checks each against the approved sample and the tech pack, counts the defects, and accepts or rejects the whole order against those two numbers. It exists because checking every piece of a 5,000-piece order is slow and costly, while checking a handful proves nothing. Sampling to a published standard gives both sides a rule they agreed before anyone looked at a garment.
What does AQL 2.5 mean?
It means the sampling plan is built for a supplier whose long-run average is no worse than 2.5 defects per hundred units. It does not mean 2.5% of your order may be faulty, and the difference matters. At 1,000 pieces, Level II, AQL 2.5 gives a sample of 80 with accept 5 and reject 6: find five major defects in those 80 and the order passes; find six and it fails. Five in 80 is 6.25% of the sample, which surprises anyone who reads 2.5 as a hard ceiling. The plan tolerates sampling noise on a single lot and is tightened over a run of lots, which is why ISO 2859-1 has switching rules that move to tightened inspection when lots keep failing. Apparel uses 2.5 for major defects by convention. It is a buyer’s choice written into the purchase order, not a legal limit.
What AQL should I use for clothing?
The common apparel convention is AQL 2.5 for major defects and 4.0 for minor defects, both at General Inspection Level II, with critical defects (anything unsafe, such as a needle fragment or a sharp trim) set to zero. Brands selling premium product, or with a history of returns, often tighten majors to 1.5, which at 1,000 pieces moves the acceptance number from 5 down to 3 on the same 80-piece sample. A stricter number than you can enforce is false comfort: if nobody writes down what a major defect is for your garment, the limit means little. Put the limits in the purchase order next to a defect list: open seams, measurements outside tolerance, shade against the approved lab dip, print faults, stains. Minor defects are things like a loose thread end that a trim removes. With us, the AQL you specify in the PO is the one the final inspection uses.
How many garments should be inspected?
It depends on the order size and inspection level, not on a fixed percentage. At General Inspection Level II, 1,000 pieces needs a random sample of 80, 3,000 needs 125, 5,000 needs 200 and 20,000 needs 315. So the share you check falls as the order grows: 8% of a 1,000-piece order, but only about 1.6% of 20,000. That is intended, because a sample drawn at random from a large lot gives the standard’s confidence without opening thousands of garments. The pieces must be drawn at random across cartons, sizes and colours, not lifted from the top of the first box, or the result describes that box rather than the order. For very small orders the table can call for a sample as large as the order itself, and then every piece is checked. On our floor a 150-piece order is inspected piece by piece rather than sampled.
What is the difference between major and minor defects?
A major defect is one a customer would notice and could reasonably return the garment for: a seam that opens, a hole, a measurement outside the agreed tolerance, a visible stain, a print that is misplaced or cracked, a colour that does not match the approved lab dip. A minor defect is a small imperfection that does not affect wear or sale, such as an untrimmed thread end or a slightly uneven stitch on an inside seam. A critical defect is anything that could hurt someone or break a product-safety rule, such as a broken needle left in a garment or a drawcord on children’s wear; the usual limit for those is zero. The categories only work when they are written down for your product. A defect list agreed before production, with photos of borderline cases, prevents the argument that otherwise happens at inspection over whether a fault is major or minor.
Which inspection level should I choose?
General Inspection Level II is the default in both ANSI/ASQ Z1.4 and ISO 2859-1, and the right choice for almost every clothing order. Level I uses a smaller sample, about 40% of Level II’s at common order sizes, and suits a supplier you already trust on a repeat style, where the cost of inspection matters. Level III uses a larger sample and suits first orders, a new factory, or product where a failure would be expensive, such as a launch you cannot move. At 5,000 pieces and AQL 4.0 the sample is 80 at Level I, 200 at Level II and 315 at Level III. The standards also have special levels, S-1 to S-4, for destructive or costly tests such as wash testing, and the calculator covers those too. Whatever you choose, keep it the same across inspections of the same product, or the results cannot be compared.
How do you read an AQL chart?
An AQL chart is two tables used in sequence. First, find your order size in the lot-size column of Table I and read across to your inspection level; that gives a code letter. At Level II, 1,000 pieces is code J. Second, find that letter in Table II-A for normal inspection. Its row gives the sample size, 80 for J, and the column for your AQL gives two numbers, Ac (accept) and Re (reject). At AQL 2.5, row J reads Ac 5, Re 6. Some cells hold an arrow instead of numbers: follow it to the first plan in that direction and use that plan’s sample size as well, which is where most hand-read charts go wrong. If the sample is equal to or larger than the order, inspect every piece. The calculator above follows exactly these steps using the MIL-STD-105E tables, which carry the same values as Z1.4 and ISO 2859-1.
What happens if an order fails AQL inspection?
A failed result means the lot, as presented, is rejected. It does not mean every garment is bad. The first step is finding the cause: one operator’s seam fault, a roll of fabric with a shade problem, or a measurement drifting after a machine setting changed. The affected goods are then sorted piece by piece and the faulty ones repaired or replaced, and the lot is inspected again as a fresh sample, at the same terms or tighter, never looser. Commercially, this is why payment terms matter. With us, the balance is paid against the final inspection report and before the goods ship, so a failed inspection holds the shipment and the money in Narayanganj rather than arriving as a problem in your warehouse. In-line checks during sewing exist to stop it reaching that point: they catch a systematic fault at 200 pieces rather than 5,000.
Do small clothing orders need an AQL inspection?
Not necessarily. The tables still give a sampling plan for a small order, so it can be sampled, but it does not have to be. At Level II and AQL 2.5, a 150-piece order needs 20 pieces checked with accept 1 and reject 2, and any order of 25 pieces or fewer calls for a sample of 5, which becomes the whole order at 5 pieces or fewer. For a first drop of 150 to 300 pieces, checking every garment is often simpler and more useful than sampling, because each return costs a small brand proportionally more. That is how we work: at 150 pieces every unit is inspected against the sealed pre-production sample rather than a sampled percentage, and at bulk volumes we apply the AQL sampling you specify in the purchase order. If you hire a third-party inspector instead, the calculator gives you the sample size and accept and reject numbers to put in their brief.
What AQL should I use for critical defects?
For clothing, the usual answer is none: critical defects are not allowed, so a single one found anywhere in the sample rejects the order. A critical defect is one that could hurt the wearer or break a product-safety rule, such as a broken needle or other metal left in a garment, a sharp edge on a trim, a drawcord in the hood of children’s wear, or a missing care or flammability label where the law requires one. Some buyers write a very low AQL such as 0.065 or 0.10 instead of zero. That is still a sampling plan, which can accept one or more criticals in a large sample, and it often forces a bigger sample than the other classes need: at 500 pieces, Level II, AQL 0.10 calls for 125 pieces and accepts none, where the major plan at 2.5 needs only 50. Zero tolerance is simpler to write into a purchase order and to enforce.
When should I use special levels S-1 to S-4?
Use a special level when each check is slow, expensive or destroys the garment, and a small sample is acceptable. In clothing, that means checks such as measuring every point on the size chart, a fit check on a form, a seam-strength pull, or a wash test for shrinkage and colour change. The main visual inspection should stay at General Level II. The special levels give much smaller samples: at 3,000 pieces and AQL 2.5, S-1 and S-2 call for 5 pieces, S-3 for 20 and S-4 for 32, against 125 at Level II. The trade-off is weaker protection, because a small sample can pass a lot that has a real problem. One practical pattern is two plans on the same order: Level II for workmanship and appearance, and S-2 or S-3 for measurements. Pick the special level in the calculator and it gives the plan for each defect class, ready to print.
Should each colour or size be inspected as a separate lot?
The standard says a lot should, as far as practicable, consist of one type, grade, class, size and composition of product, made under essentially the same conditions and at essentially the same time (MIL-STD-105E, clause 4.3). In clothing, colour is the line that matters most, because each colourway is dyed in its own batch at the mill and a shade fault lives in one dye lot, not across the whole order. So inspect each colourway as its own lot when you can. That means more garments checked: three colours of 1,000 pieces each need three samples of 80 at Level II, where one lot of 3,000 would need 125. Read strictly, the clause separates sizes too. Keeping all sizes of one style and colour in one lot is a reasonable choice if the sample is spread across every size in proportion, so a grading fault in one size can still be found.
If one garment has several defects, do I count them once or separately?
That depends on which counting method your purchase order names, and the standard allows both (MIL-STD-105E, clause 4.2). Percent defective counts garments: a piece with three faults is one defective unit, classed by its most serious fault, so a garment with one major and two minor defects counts as one major defective and nothing else (clauses 3.15 and 3.17). Defects per hundred units counts every fault, so the same garment adds one major and two minors to the tally. The tables give the same accept and reject numbers either way, which is why the choice matters. Take a sample of 80 with accept 5 for majors: three garments with two major faults each is 3 defectives and a pass, but 6 defects and a fail. Neither method is wrong. Write the one you want into the purchase order and the inspector’s brief, so every inspection fills in the tally sheet the same way.