AQL 2.5 vs AQL 4.0: Quality Inspection Standards Every Hardware Buyer Must Know

AQL stands for Acceptance Quality Limit. It is a statistical sampling standard, codified internationally as ISO 2859-1 (and historically as MIL-STD-105E), that defines the maximum percentage of defective units in a batch that may still be considered acceptable. It does not promise zero defects — it defines an agreed defect threshold.

AQL is applied through random sampling. An inspector does not examine every piece in a 10,000-unit shipment — they examine a statistically determined sample, count defects against the AQL threshold, and either accept or reject the entire batch. The sample size, accept number, and reject number are all derived from standard tables based on batch size and inspection level.

For B2B hardware imports, AQL is the contractual basis for rejecting a shipment that does not meet quality. Without an AQL clause in your purchase order, you have no formal grounds to refuse goods at QC — only your own subjective judgement, which is much harder to enforce internationally.

The two most commonly specified AQL levels in B2B hardware are 2.5 and 4.0. The number refers to the percentage of defective units that is considered acceptable in the batch. AQL 2.5 is stricter; AQL 4.0 is more permissive. The choice depends on the category of hardware, the price point, and the end-buyer expectation.

Every AQL inspection assigns each defect found to one of three categories. The defect category list should be defined and agreed with the manufacturer before production — ideally annexed to the purchase order. Defining the categories after inspection causes disputes; defining them in advance prevents them.

Sample size under ISO 2859-1 is driven by batch size and inspection level — General Inspection Level II is the most common default for hardware. The standard provides a lookup table: for a batch of 1,201 to 3,200 units, the sample size is 125 pieces; for 3,201 to 10,000 units, the sample size is 200 pieces; for 10,001 to 35,000 units, the sample size is 315 pieces.

Each sample size also has a specific accept and reject number at each AQL level. For example: at AQL 2.5 with a sample size of 125, the accept number is 7 and the reject number is 8. This means up to 7 major defects in the 125-piece sample is acceptable; 8 or more defects rejects the batch.

The inspector pulls samples randomly from across the batch — not just from the top of cartons. A proper AQL inspection physically opens cartons from multiple production lots to ensure the sample is genuinely representative.

AQL inspection at dispatch is the final stage — but it should not be the only stage. A proper manufacturer-side quality process runs multi-stage inspection through the production lifecycle, so that issues are caught and corrected long before final dispatch.

The AQL clause in your PO should be explicit and unambiguous. A one-line reference is not enough — the standard, the inspection level, the AQL values for each defect category, and the acceptance documentation should all be defined.

For larger orders, first-time supplier relationships, or high-value retail orders, many B2B buyers engage a third-party inspection agency to conduct the pre-shipment AQL inspection independently. The major international agencies — SGS, Bureau Veritas, Intertek, and TÜV — all have inspector networks across India, including coverage of Aligarh.

Third-party inspection cost is typically USD 250–400 per inspection day, plus inspector travel. The inspector visits the factory on the agreed dispatch date, conducts the AQL sampling, and issues a written inspection report — often within 24 hours. The report is the basis for releasing the balance payment and authorising dispatch.

For ongoing supplier relationships where trust is established and the manufacturer's in-house QC documentation has proven reliable over multiple orders, third-party inspection often becomes optional rather than mandatory.

A useful pre-shipment inspection report is structured, evidenced, and unambiguous. It should give you the data to release the balance payment confidently or to challenge the result formally.