How to Tell If Gold Is Real: The Home Tests That Work, and the One Fake They Miss

Every four years, Olympic champions bite their gold medals for the photographers. The Paris 2024 gold medal weighed 529 grams, and six of those grams were gold, plated over silver. So the most famous gold test in the world is performed, on live television, on a piece of silver.

Biting is out, then. No single home test proves gold is real, but four cheap ones catch most fakes: read the stamp, look where the piece rubs, try a magnet, and, for coins and bars, check the weight, size and density. Gold is denser than almost anything a faker can afford, which makes it unusually easy to check. The one fake those tests miss is tungsten, and it gets its own section below.

Think of it as buying a used car. The sticker on the window says what the seller would like you to believe, while the rust under the paint, the worn tires and the odometer say what actually happened to it. Your ring is about to get a test drive.

What do the stamps on gold jewelry mean?

The stamp says how much of the piece is gold. 10K, 14K and 18K mean 10, 14 or 18 parts gold out of 24; the European-style numbers 417, 585 and 750 say the same thing in thousandths. Stamps that mean the piece is not solid gold include GP (gold plated), GEP and HGE (gold electroplate and heavy gold electroplate), RGP (rolled gold plate) and GF (gold filled). The scrap gold calculator has the full list, with what each grade is worth per gram, and the guide to gold and silver stamps explains every mark and who checks it.

Read it with a magnifying glass, usually on the inside of a ring band or on the tag beside a chain’s clasp. Then treat it exactly like the sticker on the car window: a useful starting point, written by the person who wants your money. A counterfeiter’s stamping tool can say 14K as easily as a jeweler’s can, and it costs about as much as a rubber stamp that says “Approved.”

A missing stamp doesn’t mean fake, either. Older pieces, handmade pieces and anything resized or repaired often lost the stamp or never had one. It means the next tests carry more weight.

In the UK the stamp is called a hallmark, and it is the law. UK jewelers write purity in carats, “9ct” or “18ct”, which mean the same as 9K and 18K. Under the Hallmarking Act 1973, gold weighing 1 gram or more can’t be sold as gold in the UK unless an assay office has tested and marked it, so a genuine UK hallmark is checked by someone other than the seller. Pieces made before 1950 have an exception if their age and fineness can be shown. The government’s hallmarking guidance notes show what a full mark looks like. Hallmarks can be faked too, but a forger now has to copy an assay office’s mark, not just a number. In the US, Canada and Australia the stamp is the maker’s own claim.

Test 1: Where does the color wear off?

Plating is a coat of paint, and paint wears through where things rub. Look at the clasp, the edges of chain links, the underside of a ring where it meets the table, and the back of an earring. If a different color shows through there, usually a coppery pink or a silvery gray, the gold is a coat and the rest of the piece is something else.

Solid gold can scratch and dent, but it’s the same color all the way through. A fifty-year-old wedding band worn every day is still gold where it’s thinnest. That is the rust-under-the-paint check, and it’s free.

Two cautions. A green mark on your skin usually means copper, and 10K and 14K gold contain plenty of copper, so a green finger proves very little. And gold vermeil, which is sterling silver plated with gold, will show silver at the wear points while still being worth something for the silver.

Test 2: Is gold magnetic?

No. Gold isn’t attracted to a magnet in any way you can feel. Hold a strong rare-earth magnet, the small kind sold for hardware projects, against the piece. If it jumps to the magnet, there is iron or steel inside, and the gold is a coat at best.

Test the body of the piece, not the clasp. Many solid gold chains use a small steel spring inside the clasp, so a magnet that grabs the clasp has found the spring and nothing else.

And a piece that ignores the magnet hasn’t passed anything much. Brass, copper, aluminum and silver don’t stick either, and they are what most plated jewelry is made of. The magnet is the tire kick: it catches a car with no tires, and that’s about all it catches.

Test 3: How much should a gold coin weigh?

For coins, this is where the testing gets serious, because the mints publish exactly what they made. Weigh the coin on a scale that reads to a hundredth of a gram, measure its width and thickness with digital calipers, and compare.

The same figures, with their sources, sit inside Is My Coin Fake?, which checks your readings for you and covers 20 coins. The Gold Eagle and Krugerrand weigh more than an ounce because they are 22K: the copper and silver are extra, and each still holds one troy ounce of gold.

The trick is to check weight and size together, and gold makes it easier than silver does. Brass is less than half as dense as Gold Eagle alloy. A brass fake made to the right weight would take up more than twice the space, so it would be visibly too thick or too wide. A brass fake made to the right size comes out at less than half the weight. Either way, it can’t pass both checks. A fake that gets one number right is like a car with a perfect paint job and an odometer reading 400,000 miles.

Worn old coins, like a sovereign from 1900, can read a little light from a century of pockets. A reading that is off means keep testing, not panic.

Test 4: How do you do the water test for gold?

This is the specific gravity test, and it’s the odometer: the one reading that is hard to roll back, because density belongs to the metal itself. You need the same scale, a cup of water and a piece of thread.

1. Weigh the piece dry and write it down.
2. Set a cup of water on the scale and press tare, so it reads zero.
3. Hang the piece on the thread so it’s completely underwater, touching neither the bottom nor the sides. The scale now shows the weight of the water it pushed aside.
4. Divide the dry weight by that reading. The answer is the specific gravity.

Gold is denser than every common fake metal except one

Density, grams per cubic centimeter (the same number as specific gravity).

Pure gold reads about 19.3, 18K about 15.6, 14K about 13.1 and 10K about 11.6. Exact figures vary a little with the alloy, so treat each as a neighborhood rather than an address. Brass reads about 8.4, which is why it fails so loudly.

Try it on a Gold Eagle. It weighs 33.931 grams and should push aside about 1.91 grams of water, for a reading near 17.8. A brass fake at the same weight would push aside about 4.04 grams, more than double, for a reading of 8.4. That is an odometer that has been rolled back with a hammer.

Into the weeds: why this is hard on jewelry

Small pieces are the problem. A 5-gram 14K ring pushes aside only about 0.38 grams of water, and a 10K ring of the same weight about 0.43. A kitchen scale that reads to a hundredth of a gram can’t reliably tell those apart, though it will still catch brass at about 0.60. For small pieces, weigh several of one karat together on the thread, or use a jeweler’s scale that reads to a thousandth. Take out stones first if you can, since a diamond or a glass stone drags the reading down. Tap the piece once underwater to knock off air bubbles, which make it read light.

You may have seen the “float test” recommended instead: drop the piece in water and see whether it sinks. Every metal sinks. A brass key sinks. The float test catches plastic, and if your gold ring is plastic you probably knew already.

The fake a kitchen can’t catch: tungsten

Tungsten is a gray metal used for light-bulb filaments and the tips of drill bits, and it has a density of 19.25. Gold’s is 19.3. No other affordable metal comes within a mile of that, which is why tungsten is the favorite core of the serious gold counterfeiter.

In September 2012 a Manhattan jewelry merchant named Ibrahim Fadl drilled into gold bars he had bought from a regular source. Inside was gray tungsten. Someone had hollowed out real bars, filled them, sealed them back up and sold them with genuine-looking serial numbers and papers. A local TV station reported that a 10-ounce bar bought for almost $18,000 was worth about $3,600. “Sick to my stomach,” he told NBC News, “but I thank god we didn’t sell this to somebody.”

Here is how close the match is. A one-ounce Gold Maple Leaf, at 31.11 grams of pure gold, pushes aside 1.610 grams of water. A tungsten copy of the same weight pushes aside 1.616 grams. The difference is six thousandths of a gram, less than a single grain of rice weighs, and no kitchen scale on earth will notice it. For 22K coins it’s worse: because coin gold is lighter than pure gold, a counterfeiter can blend tungsten with cheaper metals to hit the weight, the size and the density all at once.

So for tungsten you need equipment that reads something other than weight. Tungsten conducts electricity differently from gold and rings differently when struck, and that is what dealers test. Their handheld electronic verifiers read a coin’s electrical properties beneath the surface, and ultrasound can see a foreign core inside a bar. An X-ray fluorescence analyzer, the gun-shaped scanner many gold buyers use, reads only the outer layer of metal, so a thick gold skin over tungsten can fool it. That is why a drilled-and-filled bar, real gold outside and tungsten inside, is the hardest fake of all.

Which gold tests should you skip?

• The bite test. Pure 24K gold is soft enough to dent with your teeth, but so is lead, and 14K jewelry is too hard to mark. You learn almost nothing and you may chip a tooth. See also: Olympic champions, biting silver.
• The lighter test. Real gold doesn’t blacken in a flame, and some plated base metals do. But a flame can crack stones, melt solder and discolor a finish, and thick plating passes anyway.
• The vinegar test. Some base metals change color in vinegar and gold doesn’t. Plated jewelry has a gold surface, so it usually passes, which is the one thing you wanted to catch.
• The ceramic streak test and acid kits. Both work by scraping metal off. Acid testing on a touchstone is what many jewelers do, and it is fine for scrap, but a scratch is a bad idea on a coin or a piece you might sell as jewelry.

When should you have gold tested professionally?

When the piece is worth more than a few hundred dollars, when it’s a bar, or when your home tests disagree. Many jewelers, coin shops and gold buyers will test a piece, often for free, usually with an acid kit, an electronic tester or an X-ray analyzer, and anyone buying gold from you will test it before paying. For coins worth well above their metal, a grading service checks authenticity as part of the grade.

When you’re buying rather than selling, the best defense is the source: a mint, its authorized dealers or an established shop with a reputation to lose and a door you can walk back through next year. After the 2012 bars turned up, one gold expert told NBC News that the people who get hit are not the bigger dealers.

Isn’t most gold jewelry real anyway?

Probably, if it came from a jewelry store and carries a stamp. Most of what turns up in a jewelry box is exactly what it says, and the usual surprise is the reverse: a chain someone assumed was plated turns out to be 14K. These tests take ten minutes and cost nothing, and they matter most at the two moments money changes hands, when you buy from a stranger and when you sell. The buyer is going to test it either way. It’s worth knowing first.

Why can you test gold at home at all?

Because gold carries its own proof. Its color all the way through, its indifference to a magnet, and above all its weight for its size are properties of the element, the same in an ancient Lydian workshop as on your kitchen counter. A stamp is a promise someone made. Density is a promise nobody can make on gold’s behalf.

Gold is the rare used car that comes with an odometer nobody can roll back. Nobody, that is, except a man with a drill and a tungsten rod, which is why the last check belongs to a dealer, and why the medal on the podium is still mostly silver.

Sources

• Paris 2024 gold medal (529 grams, six grams of gold, the rest silver): Kitco News, July 26, 2024, and APMEX.
• The 2012 tungsten-filled bars, Ibrahim Fadl, the $18,000 bar worth $3,600 and both quotes: NBC News, September 19, 2012.
• Gold Eagle alloy (91.67% gold, 3% silver, 5.33% copper), weight, width and thickness: Wikipedia, American Gold Eagle. The other coins’ specifications and their sources are listed on Is My Coin Fake?
• Specific gravity of 10K, 14K and 18K yellow gold, 24K gold and brass: Stuller. Densities of gold, tungsten, lead, silver and copper: Wikipedia element entries (tungsten, 19.254 g/cm³). The 22K coin figure (17.8) is calculated from the Gold Eagle’s alloy.
• Tungsten’s use in fake bars and coins, the diamagnetic and paramagnetic difference, and detection by acoustic and electrical testing: Wikipedia, Gold coin.
• The legal meanings of gold filled, rolled gold plate and gold electroplate: FTC Jewelry Guides, 16 CFR 23.3; vermeil: 16 CFR 23.4.
• Advice for people selling gold, including sorting by karat and getting several quotes: Better Business Bureau.
• UK hallmarking (gold of 1 gram or more must be assayed and hallmarked before it is sold as gold; the pre-1950 exception): Hallmarking Act 1973, Schedule 1; GOV.UK hallmarking guidance notes.

Calculated here: water pushed aside by a Gold Eagle (1.91 g) and a brass copy of the same weight (4.04 g); a Gold Maple Leaf (1.610 g) against a tungsten copy (1.616 g); a 5-gram 14K, 10K and brass ring (0.38, 0.43 and 0.60 g); brass at 47% of Gold Eagle alloy’s density.

Questions people actually ask

How can you tell if gold is real at home?

Use several cheap tests together. Read the stamp, check the spots that rub for a different color showing through, hold a strong magnet to it, and for coins and bars compare the weight and size with the mint's specification and do a water test for density. Each catches different fakes, and none alone proves a piece is gold.

Does real gold stick to a magnet?

No. Gold isn't attracted to a magnet in any way you can feel. If a piece jumps to a strong magnet, it has iron or steel inside. Test the body rather than the clasp, which often has a steel spring, and remember that brass and copper don't stick either.

Can you tell if gold is real with a lighter?

Not reliably. Real gold doesn't blacken in a flame and some plated base metals do, but thick plating passes, and the heat can crack stones, melt solder and spoil a finish. A magnet, the wear spots and a water test tell you more without the risk.

Does biting gold tell you if it's real?

Barely. Pure 24K gold is soft enough to dent with your teeth, but so is lead, and 10K, 14K and 18K jewelry is too hard to mark. Even Olympic gold medals are mostly silver: the Paris 2024 medal held six grams of gold over silver.

How do you tell if a gold chain is real?

Look for a karat stamp on the clasp or the tag beside it, then check the clasp and the edges of the links for another color wearing through. Hold a magnet to the chain itself, not the clasp. For a heavier chain, a water test shows whether its density matches the stamp.

Does fake gold turn your skin green?

Often, but so can real gold. The green comes from copper, and 10K and 14K gold contain plenty of copper, so a green mark proves very little on its own. Plated jewelry that has worn through to brass or copper is the usual culprit.

Can a gold piece be stamped 14K and still be fake?

Yes. A stamp is a claim, not proof, and a counterfeiter can stamp 14K as easily as a jeweler can. Treat the stamp as a starting point and confirm it with the wear check, a magnet and, for heavier pieces, a density test or a jeweler's test.

How can you tell if a gold bar is real or tungsten?

Not with a kitchen scale. Tungsten's density, 19.25, is almost exactly gold's, 19.3, so a gold-plated tungsten bar weighs right. Dealers catch it with electronic verifiers that read the metal's electrical properties, with ultrasound, or by testing how it rings. Buying from established sources is the main defense.

What is the specific gravity of 14K gold?

About 13.1 for yellow 14K, against about 19.3 for pure gold, 15.6 for 18K and 11.6 for 10K, per Stuller's jewelry reference table. Exact figures vary with the alloy. Brass, a common fake metal, reads about 8.4.

Does real gold float or sink?

It sinks, but so does every metal, so the float test only catches plastic. The useful water test is the specific gravity test: weigh the piece dry, then weigh the water it pushes aside, and divide the first number by the second.