October 01, 2026

How to Tell If Gold Is Real: 9 Tests, From the Kitchen Table to the Jeweler

By Sasha Tarasova

Filed under Hallmarks, Dating & AuthenticationGold & Precious MetalsBuying Guides

Short answer: to tell if gold is real, start with the stamp, then check wear points for a different metal showing through, then use a strong magnet. Those three home checks catch most fakes without harming the piece. They cannot prove a piece is solid gold. Only an acid test, an electronic tester or XRF analysis by a jeweler confirms the karat.

Most people asking this question are holding one of three things: something inherited, something bought online, or something a friend swears is 18k. The honest answer is that home tests are good at exposing obvious fakes and poor at confirming the real thing. Below are nine tests, ranked from the kitchen table to the jeweler's bench, with a plain verdict on each, plus the popular tests that are simply myths.

Can you tell if gold is real at home?

Partly. At home you can rule things out: a piece that sticks to a magnet, shows grey or copper-coloured metal at the edges, or carries a plating mark such as GP or HGE is not solid gold. What you cannot do at home is prove a piece is solid gold of a specific karat.

That distinction matters because the most common "fake" in circulation is not a lump of brass pretending to be gold. It is plated or gold filled jewelry that looks exactly like solid gold on the surface, sometimes for decades. A good plating passes every test that only looks at the surface. Our guide to gold plated, gold filled, vermeil and solid gold explains what each layer is and how thick it has to be.

The second complication is antique gold itself. Low-karat alloys, rose-toned copper golds and unmarked pieces are normal in antique jewelry. A ring that looks "too pink" or "too pale" or has no stamp at all can be perfectly genuine. Treat every result below as one piece of evidence, not a verdict.

The 9 tests at a glance

Here is every test worth knowing, what it can actually tell you, how far to trust it, and what it risks doing to the piece.

For completeness: fire assay, which melts a sample, is the definitive method used by refiners and assay offices. It destroys the sample, so it is for scrap and bullion, not for jewelry you want to keep.

How do you read the stamp on gold?

Look for a karat mark such as 10K, 14K or 18K, or a three-digit fineness number such as 375, 585 or 750, usually inside a ring shank, on a clasp or on a tag near the clasp. A karat is one twenty-fourth part, so 14k is 14 parts gold in 24, which is 58.5%, and is stamped 585 in the millesimal system.

A stamp is a claim, not proof. Three things to know:

• Plating marks sit next to karat marks. "14K GP", "18K HGE", "1/20 12K GF" and "RGP" all describe gold on the surface of another metal. The karat refers to the surface layer only.
• Fake stamps exist. A crisp "750" on a piece bought for very little online is not evidence. Under US law a karat stamp should be accompanied by the registered trademark of the maker or seller who stands behind it; a quality mark with no maker's mark is a reason for caution on modern pieces.
• Genuine antiques are often unmarked. American jewelry before the early twentieth century was frequently unstamped, rings that have been resized can lose their marks, and in Britain gold articles under one gram are exempt from compulsory hallmarking. An absent stamp is neutral.

A full British hallmark, with sponsor's mark, standard mark, assay office mark and often a date letter, is far stronger evidence than a bare number, because it was applied by an independent assay office after the metal was tested. Our guide to reading a hallmark walks through each element, and if you want to know what 375 or 916 actually signifies, the common fineness numbers are explained in our guide to 9k and 10k gold.

Does the magnet test work on gold?

The magnet test works in one direction only. Gold is not magnetic, so a piece that pulls firmly toward a strong magnet is not solid gold, or contains a steel or iron core. A piece that does not react has only proved it is not iron or steel.

Brass, copper, silver, aluminium, lead and tungsten are all non-magnetic too, which is why this test cannot confirm anything on its own. It is still worth doing, because a lot of cheap plated jewelry is built on steel.

How to do it properly

• Use a neodymium magnet (sold as rare-earth magnets in hardware stores). Fridge magnets are too weak to show a mild pull.
• Test the body of the piece, not the clasp. Many genuine gold clasps contain a small steel spring, which will twitch toward the magnet and mean nothing about the chain.
• Watch for a weak drag rather than a snap. Some white gold alloys containing nickel, and some older solders, can show a very faint response. A strong, definite pull is the result that matters.

What does wear tell you about real gold?

Wear is the most useful free test for older pieces. Solid gold is the same colour all the way through, so when it wears it simply gets thinner and softer at the edges. Plated gold wears through to a different metal underneath.

Look with a 10x loupe, or your phone camera zoomed in, at the places that take the most friction:

• the underside of a ring shank, where it rubs on the palm and on surfaces
• the edges of a clasp, jump rings and the links nearest the clasp
• the high points of any raised decoration
• the back of earrings and pendant bails

Grey, silvery, coppery or dark patches with a sharp edge between them and the gold are the tell. So is green corrosion in crevices, which points to a brass or copper base. Flaking or bubbling surface is another sign of plating.

Be careful with two false alarms. Antique rose and red golds contain enough copper to look pinkish throughout; that is the alloy, not base metal. And white gold is usually plated with rhodium, a bright white metal, which wears to reveal the slightly yellow or greyish white gold beneath. That is normal maintenance, not fakery.

How does the density test work?

The density test compares how heavy the metal is for its volume, and gold is exceptionally dense. Pure gold has a specific gravity of about 19.3, meaning it weighs 19.3 times as much as the same volume of water. Brass and copper sit around 8.4 to 9, steel around 7.8.

Step by step

1. Confirm the piece is solid, stone-free and not hollow. Stones, enamel, glue, hollow links and lockets all distort the result. Do not immerse pieces with pearls, opals, organic materials or glued-in stones.
2. Use a digital scale reading to 0.01g. A kitchen scale reading to 1g is useless for jewelry.
3. Weigh the piece in air and write the figure down.
4. Put a small cup of water on the scale and zero it.
5. Hang the piece from a thin thread so it is fully submerged but touches neither the sides nor the bottom of the cup. Read the scale. That figure is the weight of water displaced.
6. Divide the weight in air by the displaced weight. The result is the specific gravity.

Approximate specific gravity ranges, which vary with the exact alloy:

A reading of 8.5 on a chain stamped 18K is a clear fail. A reading of 13.5 on something stamped 14K is consistent. On a two-gram ring, a single hundredth of a gram changes the answer, so treat small pieces as indicative only. White golds containing palladium can read a little higher than yellow gold of the same karat.

Which home gold tests are myths?

The bite test, the float test, the skin test and the flame test are unreliable, and two of them damage jewelry. They survive because they are easy, not because they work.

• The bite test. Pure gold is soft enough to dent with teeth, which is where the image of prospectors biting coins comes from. Jewelry gold is an alloy and is much harder; biting it tells you nothing except that you have marked a ring and possibly chipped a tooth. Lead is softer than gold and dents beautifully.
• The float test. Real gold sinks. So does every other metal used in jewelry. The test is meaningless.
• The skin test. The idea is that fake gold turns skin green and real gold does not. Copper-rich metals do cause green marks, but so can genuine 9k, 10k and 14k gold on some skin, especially with sweat, lotion or chlorine. Black marks on skin are often caused by makeup abrading the metal, real gold included.
• The flame test. Holding a lighter to jewelry risks cracking stones, burning enamel, weakening solder joints and discolouring the surface. Do not do it.
• The ceramic streak test. Dragging the piece across unglazed porcelain leaves a streak whose colour is supposed to reveal the metal. It does remove metal, and a plated piece leaves a gold streak until the plating is gone. We do not recommend it.
• The vinegar test. Real gold does not react to vinegar, but neither does a sound plated surface, so a pass means little. Vinegar can etch pearls, coral, opal and some old solders. Skip it on anything set with stones.

What will a jeweler do to test gold?

A jeweler will usually combine an electronic tester or XRF with an acid test when there is doubt. Each takes minutes.

The acid test

The piece is rubbed on a black stone called a touchstone to leave a thin streak of metal. Drops of testing acid, calibrated for different karats and based on nitric acid, and for high karats on aqua regia, are applied to the streak. Whether the streak dissolves or survives indicates the karat. It is fast, cheap and accurate, but only if the metal tested is the metal underneath. For that reason a careful tester files a tiny notch in a hidden spot first, so that a plated surface is not mistaken for solid gold. That notch is minor but real, and the acids are hazardous. This is a professional test, not a home kit project, and an acid kit in untrained hands is how people etch stones and burn skin.

The electronic tester

Electronic testers measure how the metal responds electrically at a contact point, usually through a small drop of gel, and display an estimated karat. They leave no mark. Their weakness is plating: a heavy gold layer can read as gold because the tester only sees the surface.

XRF analysis

X-ray fluorescence (XRF) bombards a small area with X-rays and reads the characteristic energy each element gives off, reporting the percentage of gold, silver, copper, zinc, nickel and so on. It is non-destructive and precise, and it is what assay offices and many refiners use to screen metal. It measures only a thin surface layer, so a thick plating can still mislead it, which is why assayers who need certainty scrape or sample beneath the surface.

What if the "gold" is actually something else?

Many pieces that fail home tests are not modern fakes. They are honest period materials that were never meant to be solid gold.

• Pinchbeck is a copper and zinc alloy devised in the early eighteenth century by the London clockmaker Christopher Pinchbeck to imitate gold. Georgian pinchbeck is collectible in its own right; our Georgian jewelry guide covers the period.
• Gold filled and rolled gold were respectable, widely sold materials in the late nineteenth and twentieth centuries, especially for watch cases, lockets and chains.

None of these is a fraud unless someone sells it as solid gold. That is the question to ask of the seller, and our note on what a dealer should disclose sets out what a straight seller will tell you without being asked.

A practical checklist for any piece of gold

1. Find the marks with a loupe. Note karat or fineness numbers, maker's marks, and any GP, GF, HGE, RGP or fraction.
2. Inspect the wear points for a second metal, green corrosion or flaking.
3. Try a neodymium magnet on the body of the piece, away from the clasp spring.
4. If the piece is solid and stone-free, run a density test on an accurate scale.
5. Skip the bite, float, flame, vinegar and scratch tests. They risk the piece and prove little.
6. If anything is uncertain or the piece matters, take it to a jeweler for an electronic or XRF test, and agree beforehand whether an acid test and file mark are acceptable.
7. Keep the result in writing, with the date and who tested it. It is useful for insurance and appraisal later.

When should you see a professional?

See a jeweler or appraiser whenever money is about to change hands, whenever a piece is set with stones you cannot remove, and whenever a home test gives a mixed result. A stamped 14K piece that is not magnetic but shows odd wear is exactly the case where a two-minute XRF reading saves weeks of guessing.

If you are buying rather than checking something you already own, the burden should sit with the seller. A reputable dealer states the metal and karat in the listing, stands behind it with a return policy, and does not mind you having it tested. Our guide to buying antique jewelry online covers the questions to ask before you pay, and if a piece has no marks at all, our guide to dating a ring without a hallmark explains what construction can tell you instead.

For anything valuable, heavy or signed, a proper appraisal is worth its fee. A signed piece such as a vintage Tiffany & Co. 18k bracelet carries maker's marks that have their own authentication questions, and final authentication of a signed piece belongs to the house or a qualified specialist.

Questions people actually ask

Does real gold stick to a magnet?

No. Gold is not magnetic, so solid gold jewelry should not pull toward a strong magnet. A firm pull means the piece contains iron, nickel or steel. But passing the magnet test does not prove a piece is gold, because brass, copper, silver and many other metals are not magnetic either. Test the body of the piece, not the clasp, which may contain a steel spring.

Is gold real if it has no stamp?

It can be. Many antique pieces were never stamped, marks are often lost when a ring is resized, and in Britain gold items under one gram are exempt from compulsory hallmarking. An unmarked piece needs testing rather than assumption. A jeweler can confirm the karat in minutes with an electronic tester, an XRF reading or an acid test.

What is the most accurate way to test gold at home?

A specific gravity test on an accurate scale is the most informative home test, because it measures how dense the metal is. It only works on solid pieces without stones, hollow parts or glue, and small pieces give unreliable results. Combine it with a stamp check, a wear inspection and a magnet test, then confirm anything important with a jeweler.

Does the vinegar test work for gold?

Not well. Real gold does not react to vinegar, but neither does the surface of a well-plated piece, so a pass tells you very little. Vinegar can also etch pearls, opal, coral and some older solders. It is one of the least useful home tests and is not worth the risk on anything set with stones or of sentimental value.

Can real gold turn your skin green?

Yes, occasionally. Lower karat golds such as 9k, 10k and 14k contain copper and silver, which can react with sweat, lotion, chlorine or some skin chemistry and leave a green or dark mark. It is more common with plated brass or copper jewelry, but a green finger on its own does not prove a piece is fake.

What does 14K GF or 18K GP mean?

These are not solid gold marks. GF means gold filled, a layer of karat gold bonded to a base metal core. GP means gold plated, a much thinner electroplated layer. The karat figure describes only the gold layer on the surface. A piece stamped 14K GF is gold filled, not 14 karat gold throughout.

Will a jeweler's acid test damage my jewelry?

Slightly. To avoid being fooled by plating, a careful tester files a tiny notch in an inconspicuous spot and tests the metal from there, and the acids can harm stones if handled carelessly. On antique or signed pieces, ask for an electronic or XRF test first. Both are non-destructive and usually sufficient.

Every piece we list is described by metal and karat, and you can browse the full range of antique and vintage gold jewelry with that in mind. If you are looking for a particular piece in solid gold and want someone else to do the checking, our sourcing service runs the search and the testing for you. More on karats, alloys and colours sits in our gold topic archive.