How to Tell If Gold Is Real: The Marks, the Tests, and What Each Test Misses

To tell if gold is real, read the marks first, then run the checks that cannot damage the piece, then have a professional test anything that still matters. A mark tells you what the maker claims, not what the metal is. A magnet, a gram scale, or a pair of calipers can rule some fakes out, but no home test proves a piece is gold. Each professional test has a blind spot as well: acid and X-ray tests read the surface, and a gold-covered core of a dense metal can pass both.

Key takeaways

• Read marks as claims. A karat mark should sit beside the maker’s trademark, and the law lets actual fineness fall up to 3 parts per thousand short of the mark.
• A magnet, a gram scale, and calipers can rule fakes out without harming the piece. None of them proves a piece is gold.
• Weight and size cannot catch a tungsten core, because tungsten’s density is within about 0.1 percent of gold’s.
• Professional tests have limits too. Acid and X-ray fluorescence read the surface, and the LBMA states that no single detection method offers complete certainty.

Read the marks first

Karat is parts of gold out of 24. 24K is pure gold, and 18K is 18 parts gold and 6 parts other metal, as the FTC’s jewelry guidance explains. By the same arithmetic, 10K is about 41.7 percent gold, 14K about 58.3 percent, 18K 75 percent, and 22K about 91.7 percent. A three-digit number such as 750 states fineness in parts per thousand, so 750 is the same as 18K.

The FTC Jewelry Guides (16 CFR 23.3) set what the words mean. The word gold on an alloy must be preceded by its karat, and a plated item must say it is plated. Gold Filled means a mechanically applied layer of gold alloy, at least 10 karat, that makes up at least 1/20 of the metal’s weight. Gold Plate, Gold Overlay, and Rolled Gold Plate (RGP) need at least 1/40. FTC consumer guidance describes gold electroplate as a layer of at least 10 karat gold at least .175 microns thick. Vermeil is gold over sterling silver. Solid gold means the item is not hollow, and it still needs a karat mark.

A karat mark should come with the maker’s trademark or name under 15 U.S.C. 297(b). The FTC’s advice is that a karat mark with no trademark is a reason not to buy. The law also allows some slack: a gold article’s actual fineness may fall short of its mark by no more than 3 parts per thousand, not counting solder (15 U.S.C. 295).

Coins and bars state weight and fineness instead of karat. A replica coin made or imported for sale in the United States must carry the word COPY on one face (16 CFR 304.6). A counterfeiter ignores that rule, and making a counterfeit gold coin or bar is a federal crime (18 U.S.C. 485). A mark is a claim, not proof.

How to test gold at home without damaging it

Each check in this section can rule a piece out. None of them can prove a piece is gold.

The magnet check

Gold is diamagnetic, so a magnet does not pull it, and copper behaves the same way. If a magnet pulls a piece firmly, the piece contains a magnetic metal and is not solid gold. So the answer to “does real gold stick to a magnet” is no. A piece that passes has proved nothing, though, because many base metals do not respond to a magnet either. Test the body of the piece, not the clasp: a clasp or spring can contain steel while the rest of the piece is gold.

Weight and size of coins

Coins have published specifications, so a gram scale and calipers can catch a crude fake. The 1-ounce American Gold Eagle is 32.7 mm in diameter, weighs 33.931 grams, and contains one troy ounce of fine gold, under the law that authorizes it (31 U.S.C. 5112). That works out to about 91.7 percent gold (31.1035 ÷ 33.931). For any other coin, use the issuing mint’s published specification.

Density

NIST lists pure gold at 19.32 grams per cubic centimeter, lead at 11.35, and copper at 8.96. From those figures, a 1-troy-ounce bar of pure gold (31.1035 grams) takes up about 1.61 cubic centimeters. The same volume weighs about 18.27 grams in lead and about 14.42 grams in copper, so a gold-colored bar of either metal fails a scale check.

Tungsten is the exception. NIST lists it at 19.30 grams per cubic centimeter, so a pure tungsten bar the same size would weigh about 31.07 grams, only 0.03 grams less than gold. A tungsten core barely changes a bar’s weight, and weight and size cannot catch it. Density checks also suit pure gold better than jewelry, because a karat alloy is lighter by design. In GIA’s comparison of four testing methods, hydrostatic weighing, the lab version of a density check, was the least suitable for karat-gold jewelry.

Wear and color

Gold plating wears away with use, so base metal showing at edges, clasps, or high points means the piece is plated. Pure gold does not react with oxygen, so it does not tarnish. Karat alloys contain other metals. Treat discoloration as a reason for a professional test, not a verdict.

Home tests that do not settle the question

Popular home tests include a flame, vinegar, lemon juice, biting, makeup or ceramic streaks, and ice. None of the federal rules, lab studies, or industry reports cited on this page treats any of them as a reliable test.

• Flame or lighter. Heat can damage a genuine piece.
• Vinegar or lemon juice. Gold resists most acids, but a plated piece shows only its gold surface to the liquid, so a pass says little about what is underneath.
• Biting. Biting can damage a genuine piece, and any mark it leaves is in the surface.
• Makeup or ceramic streaks. A streak comes from the surface, which is the part a plating controls.
• Ice. None of the sources cited here supports it as a test.

The pattern is the same across the list. These tests read the surface or the weight, and the LBMA describes counterfeits built to match both.

How professionals test gold, and what each method misses

When real money depends on the answer, a professional test is the next step. Each method has a blind spot, so ask the tester which method it uses and whether it marks the piece.

Test	What it checks	What it can miss	Damages the piece?
Acid or touchstone	A streak rubbed onto a stone is tested with acids. Gold does not dissolve in nitric acid alone, which dissolves silver and base metals.	The streak comes from the surface, so a thick plating can pass.	Yes. The rubbing removes a little metal.
X-ray fluorescence (XRF)	The elements at the surface, read without marking the piece.	A piece plated with the same element, such as gold over lower-karat gold.	No.
Electronic tester	An electrical reading of the metal (capacitive decay).	GIA’s 1992 comparison found it useful only within limits.	Ask the tester.
Hydrostatic weighing (density)	The average density of the whole piece.	An average hides layers and cores. GIA rated it least suitable for karat-gold jewelry.	No.

GIA calls XRF “one of the most reliable ways to analyze metal composition in the jewelry industry”. The same 2025 lab note describes its weak point. An item submitted as 14K turned out to be a 9K rose gold base under a 13.5K yellow gold layer. Depending on its thickness, plating like that can produce false XRF results, and GIA treats a microscope check for discoloration or plating damage as essential.

GIA’s 1992 comparison of density, touchstone, an electronic tester, and XRF found that none of the four reliably determined gold content within the 3-parts-per-thousand legal tolerance. For bars, the LBMA reaches a similar conclusion: no single detection method offers complete certainty, particularly against well-executed fakes.

Counterfeit gold bars and coins

Counterfeit gold bars are a documented problem. The LBMA reported in The Alchemist, issue 117 that more than 1,000 counterfeit gold bars, worth over $50 million, were found in vaults worldwide in 2019. They carried forged refinery stamps, matched expected assay and weight specifications, and some tested at 99.98 percent purity.

The LBMA also describes the methods. Counterfeiters use dense cores to keep the weight while cutting the gold content. They copy designs with dimensionally accurate molds and replicate branded assay cards, tamper-evident seals included. Even combined, detection tools cannot always catch a counterfeit made from real gold that carries fraudulent markings or paperwork.

The density figures above explain the core. A tungsten gold bar is gold on the outside and tungsten inside. It can pass the scale and the calipers because tungsten’s density is within about 0.1 percent of gold’s, even though most of the metal is not gold.

• Keep purchase records for anything you buy.
• Before you buy, ask the seller how it authenticates and what happens if an item fails a test. An answer in writing can be checked later.
• Treat a sealed assay card as part of the claim, not proof. The LBMA reports that counterfeiters replicate the cards and the seals.
• Know where fakes circulate. The LBMA reports that counterfeiters sell through online platforms and informal dealer networks.

One request should end any conversation. No government agent asks anyone to buy gold bars or hand them to a courier, and the FTC states that such a request is a scam. Fraud built around gold IRAs follows related scripts; see gold IRA scams.

What to do with a piece you are unsure about

Free checks come first and paid tests last.

1. Read every mark with a magnifier and write each one down.
2. For a coin, weigh and measure it against the issuing mint’s specification.
3. Try a magnet on the body of the piece, not the clasp.
4. For anything worth more than its melt value to you, ask a jeweler, dealer, or assay lab to test it. Ask which method it uses and whether the method marks the piece.
5. Get the result in writing.
6. If you plan to sell, compare offers as a percentage of melt value. The gold price per gram calculator and the gold bar and coin value calculator give that baseline, and how to sell gold covers the comparison.

Related research

• Gold price per gram: melt value by weight and karat.
• Gold bar and coin value: melt value of common bars and coins.
• How to sell gold: where to sell, how offers are set, and the tax.
• Gold IRA scams: fraud patterns that use gold as the hook.
• IRA-eligible gold: which bullion products an IRA can hold.
• Gold IRA glossary: terms used across this site.