Stainless Steel Bolt Torque Chart: A2-70 & A4-80 (M3–M24)

Grab a carbon-steel torque chart and use the same numbers on a stainless bolt because the size matches, and you are working from the wrong chart. Stainless is not a drop-in for Class 8.8 or 10.9 — it is weaker in a different way, and it fights back on the way in.

Stainless fasteners on this site are almost always called out as A2-70 or A4-80 — the property classes defined by ISO 3506-1. The letter and the number mean two completely different things, and mixing them up is the single most common stainless fastener mistake.

A2 and A4 are corrosion groups, not strength grades

ISO 3506-1 splits the label into two independent pieces. The letter-digit (A2, A4) is the chemical composition subgroup, and it is entirely about corrosion resistance: A2 is the standard 18/8 chromium-nickel alloy (no molybdenum) — close to AISI 304 — good for general indoor and mild-outdoor use. A4 adds 2–3% molybdenum — close to AISI 316 — which raises the pitting-resistance equivalent number from roughly 18–19 to 24–28 and is why A4 is the standard choice for marine, coastal and chloride-heavy service.

The number after the dash (50, 70, 80) is the property class, and it is entirely about how much the wire was cold-worked before it became a fastener. Austenitic stainless has no martensitic transformation, so it cannot be hardened by quench-and-temper the way an 8.8 or 10.9 bolt is. Instead the manufacturer strain-hardens it — cold heading, cold drawing and thread rolling all add dislocation density and, in A2 grades especially, convert some austenite to harder martensite. More cold work gives a higher property class. A2-80 and A4-80 are the same steel as A2-70/A4-70, just worked harder and left with less ductility in reserve.

The strength gap: why the 8.8 number does not transfer

Compare the published minima side by side and the problem is obvious:

A note on the 8.8 row. Those are the ISO 898-1 minima for bolts up to 16 mm. Above 16 mm the standard rates class 8.8 higher — 660 MPa yield and 830 MPa tensile — so for the M20 and M24 rows in the torque chart below, those are the carbon-steel figures to compare against. The nearest US grade is SAE J429 Grade 5 (92 ksi yield, 120 ksi tensile for 1/4–1 in.), and it is a neighbour, not an equivalent: the two are separate specifications and are not interchangeable. None of these strength numbers is a torque — torque follows from diameter, pitch, friction and the clamp-load target, which is what the chart below works out. Source: ISO 898-1 property-class table as reproduced at Wikipedia, Screw (mechanical classifications).

An M10 A2-70 bolt yields at 450 MPa. A same-size Class 8.8 bolt yields at 640 MPa — 42% higher. A4-80 closes some of the gap at 600 MPa but still does not reach Class 8.8, let alone Class 10.9’s 940 MPa. If you take a torque value calculated to reach 70% of Class 8.8’s proof load and apply it to a same-size A2-70 bolt, you are not landing anywhere near 70% of its proof load — you are much closer to its yield point, with none of the margin the number implied.

Why stainless seizes: galling

Carbon steel bolts are protected, almost by accident, by their own corrosion: a thin layer of iron oxide sits between the threads and acts as a mild lubricant. Stainless steel’s whole reason for existing — a hard, passive chromium-oxide film only 2–5 nanometres thick — removes that accident. Under the extreme contact pressure at the thread flanks during tightening, that film ruptures, exposes bare metal on both surfaces, and the two surfaces cold-weld to each other. Keep turning and the weld tears, leaving a rougher, hotter, higher-friction surface behind. Within as few as two or three tightening cycles a dry stainless-on-stainless joint can seize hard enough to twist a bolt in half rather than back out. That is galling, and it is the reason stainless torque numbers scatter so much more than steel numbers: you are not measuring one consistent friction coefficient, you are measuring how close the joint got to cold-welding before it stopped turning.

It is also why a torque wrench spends so little of its reading on actual clamp force even in the best case. Of the torque you apply, roughly 40–50% is consumed by thread friction and 35–45% by friction under the head or nut — only about 10–15% becomes useful preload. On stainless, dry, that split gets worse, not better, because galling raises friction as the joint tightens rather than holding it steady.

• Highest risk: stainless bolt into a stainless nut of the same or similar grade — identical hardness on both sides is exactly the condition that cold-welds most easily.
• Lower risk: stainless bolt into a plated-steel or bronze nut (check galvanic compatibility for the environment first).
• Always: slow down. Impact drivers generate exactly the friction heat that starts the gall. Run stainless in by hand or on a low-speed setting, especially for the last turn.
• Never reuse a stainless fastener that has been tightened to near its proof load — microscopic galling damage from the first install makes the second install far more likely to seize.

The torque math: T = K · F · d

The relationship between the torque you apply and the clamp force you actually get is the same formula the site’s carbon-steel chart uses — only the numbers going into it change:

The number that actually swings the chart is K. On plain dry steel it runs about 0.19–0.20. Dry stainless runs 0.30–0.45 — call it 0.30 for a fastener that has not started to gall yet, which is already 50% higher than plated steel before anything goes wrong. Anti-seize or a PTFE paste brings it back down to about 0.12–0.15. We use K = 0.30 dry and K = 0.15 with anti-seize below — both from published stainless-specific K-factor tables, not the general-purpose steel value.

Two things to read off that table. First, anti-seize is not a rounding error here: it roughly halves the required torque for the same clamp force, because K roughly halves. That is a much bigger swing than the 25% reduction quoted for a standard plated bolt going from dry to oiled — because dry stainless friction is unusually high, not just “unlubricated.” Second, A4-80’s column runs about a third above A2-70’s at every size, in exact proportion to its higher R_p0.2 (600 vs 450 MPa) — same formula, same K, just a stronger bolt.

We checked the A2-70 anti-seize column above against a published stainless torque table (K ≈ 0.14, roughly 73% utilisation) at six sizes — M6, M8, M10, M12, M16 and M20 — and landed within 4% at every one. The residual gap is exactly what you would expect: our K is a touch higher (0.15 vs 0.14) and our utilisation a touch lower (70% vs ≈73%), and the two nearly cancel.

Lubrication is not optional on stainless

On a carbon-steel bolt, anti-seize is a convenience — it makes disassembly easier and buys some corrosion protection. On stainless it is closer to a requirement: it is the difference between a joint that comes apart in ten years and one that shears off in the attempt. PTFE-based paste (μ ≈ 0.05–0.10) gives the lowest and most consistent friction for torque-critical work; copper or nickel-graphite anti-seize is the common shop choice and is what most torque tables — including the one above — are calculated around. Full product guidance, including which compound to use where, is in the anti-seize guide: copper vs. nickel.

A2 or A4 — which one do you actually need

If the fastener never sees salt spray, road salt, a swimming pool, or a coastal atmosphere, A2-70 is the default and the cheaper choice. If it does, spend the extra on A4 — the strength scheme is identical, so moving from A2 to A4 buys corrosion resistance without changing how you read the torque chart. More on head markings, and how these compare to carbon-steel classes generally, is in bolt grades, heat treatment and hydrogen embrittlement.

Try it in your garage

1. Check the head marking. A genuine A2/A4 fastener is stamped with the class (e.g. “A2-70” or just “70” with a maker’s mark) — unmarked stainless hardware is common but unverifiable to a specific property class.
2. Magnet test, as a hint only. Heavily cold-worked A2 (304) picks up some magnetism; A4 (316) stays closer to non-magnetic even at high cold work. A magnet cannot confirm property class or composition on its own — treat it as a screening hint, not proof.
3. Lubricate before you torque. Coat the threads and the bearing face, not just the threads — a dry bearing face changes the torque split even with lubricated threads.

Some links below are Amazon affiliate links — as an Amazon Associate we may earn a small commission at no extra cost to you.

Frequently asked questions

Keep going

Bottom line: a stainless bolt is not a shinier version of the carbon-steel bolt in the same hole. It is weaker at the same property-class number, it fights you on the way in unless you lubricate it, and the only chart that gets both of those right is one built for A2/A4, not 8.8/10.9. When you need the exact torque for a real fastener, the wrench & socket finder and our vehicle database answer in one search.

Sources

• SSBolts.com: A2-70 / A4-80 strain-hardened austenitic property classes — R_p0.2/R_m minima, cold-work mechanism, PREN and A2 vs A4 composition.
• SSBolts.com: lubricants and anti-seize compounds for stainless steel — the galling mechanism, the ISO 16047 K = T/(F×d) framework, and the K-factor / friction-coefficient tables for dry vs. lubricated stainless.
• TorqueSpec.org: stainless steel bolt torque specifications — the published A2-70 anti-seize torque table used to validate the computed column, and a second, independent stainless K-factor range.
• Big Bolt Nut: ISO 3506-1 bolt grades explained — A2 = 304 / A4 = 316 confirmation and marine-use guidance.
• WN Global: stainless steel, ISO 3506-1 — A2/A4 corrosion-resistance framing and marine/offshore use case.
• ISO 898-1 — the Class 8.8 / 10.9 minima (already published on this site’s metric bolt torque chart) and the tensile stress-area formula used for every A_s value above.
• ISO 3506-1 — the A2/A4 property-class scheme itself.