Is My Gold Real? 8 Ways to Check (and What Each One Proves)

Start With the Hallmark

Before any test, look for a stamp. Almost all commercially sold gold carries a purity mark somewhere discreet: the inside of a ring shank, the clasp or a small tag on a chain, the edge of a pendant bail, the reverse of a brooch. You will usually need a 10x loupe or your phone's macro mode to read it.

The number tells you the claimed purity: 999 is 24K, 916 is 22K, 750 is 18K, 585 is 14K and 375 is 9K. Older or non-metric pieces may say 22K, 18K or 14K directly. Our hallmark decoder covers the full list, including country assay marks.

A hallmark is a claim, not a proof: stamps can be forged, and plated items are sometimes marked to look solid. But an absent hallmark on a piece sold as gold is a strong warning sign, and a legible one narrows the question enormously.

The 8 Checks, Compared

Check	Cost	What it proves	What it misses
Hallmark inspection	Free	The claimed purity and often the maker and country	Forged stamps; plated items stamped as solid
Photo / AI identification	Free	What the item most likely is, plus what the stamp means	Cannot measure metal content
Magnet test	Free	Rules out iron and steel cores	Brass, copper and lead are also non-magnetic
Weight vs size	Free	Gold is unusually heavy, so fakes often feel light	Tungsten fakes match gold's density closely
Ceramic streak test	Under $10	Real gold leaves a gold streak; plating leaves black	Scratches the piece; not karat-specific
Specific gravity test	Needs a scale	Density, which correlates strongly with karat	Fails on hollow pieces and gemstone settings
Acid test	$15–25 kit	Karat band, reliably, in about a minute	Leaves a small mark; only tests the surface layer
XRF assay at a jeweler	Often free	Actual elemental composition, non-destructively	Reads the surface, so heavy plating can still fool it

1. Read the Stamp

Free, instant, and it answers the question most of the time. Photograph the mark with your phone's macro lens and zoom in: tiny stamps are far easier to read on a screen than through a loupe. Watch for markings that indicate the piece is not solid gold: GP (gold plated), GF or 1/20 12K GF (gold filled), HGE (heavy gold electroplate), RGP (rolled gold plate) and vermeil (gold over sterling silver).

2. Run a Photo Through an Identifier

An AI identifier is the fastest way to work out what you are actually holding: a bullion coin versus a jewelry piece versus a natural nugget, and what the markings on it mean. It is the right tool for orientation, not for verdicts: it reads appearance, and appearance is only ever a strong hint about metal content. See how to photograph gold for identification to get the best possible result.

3. The Magnet Test

Gold is not magnetic. Hold a strong neodymium magnet near the piece: if it jumps or clings, there is ferrous metal inside and it is not solid gold. A fridge magnet is too weak to be meaningful.

The limit is important: brass, bronze, copper, aluminium and lead are all non-magnetic too. Passing the magnet test only tells you the fake isn't steel.

4. Weigh It Against Its Size

Gold is remarkably dense: pure gold has a standard density of 19.32 g/cm³, against roughly 8.5 for brass and 7.8 for steel. A solid gold item feels startlingly heavy for its size. If a chunky-looking "gold" ring feels like costume jewelry, trust that instinct.

The classic exception is tungsten, at 19.25 g/cm³, almost identical to gold. Tungsten-cored fakes are the main reason large bars and coins deserve a proper test rather than a hand-weigh.

5. The Ceramic Streak Test

Drag the piece firmly across an unglazed ceramic tile (the underside of a bathroom tile works). Real gold leaves a gold-colored streak. Plated or base metal leaves a black or grey streak. It costs almost nothing but it does scratch the item, so use an inconspicuous spot, or skip it entirely on anything with sentimental or antique value.

6. Specific Gravity

Weigh the item dry, then weigh it suspended in water. Divide the dry weight by the weight loss in water and you get its density, which maps closely onto karat: roughly 19.3 for 24K, 17.7–17.8 for 22K, 15.2–15.9 for 18K and 12.9–14.6 for 14K.

It is genuinely accurate for solid, single-metal items. It is useless on hollow chains, pieces with gemstones, or anything with solder joints of a different alloy.

Tip: Use the app's calculator alongside a specific-gravity result. Once you have a defensible karat figure and a weight, you can work out the melt value in seconds. See how to calculate what your gold is worth.

7. The Acid Test

A testing kit gives you a scratch stone and bottles of acid calibrated to different karats. Rub the item on the stone to leave a metal streak, apply the acid for the karat you suspect, and watch: if the streak survives, the piece is at least that karat. If it dissolves or discolors, it is lower.

It is the standard test in pawn shops and scrap buyers for good reason: cheap, fast and reliable to within a karat band. Two caveats: it marks the piece slightly, and it only reads the surface, so thick plating over base metal can pass an acid test on the outside. Serious buyers file a small notch first to test beneath the surface.

8. XRF at a Jeweler

X-ray fluorescence guns read the actual elemental composition without damaging the item, and give you percentages for gold, silver, copper, zinc and nickel. Most jewelers and gold buyers have one and will scan a piece for free or for a nominal fee, hoping you will sell to them.

This is as close to certainty as you will get outside a laboratory. Its one blind spot is depth: XRF reads the outer layer, so a thickly plated fake can still register high. For high-value bars, combine XRF with dimensions, weight and an ultrasound or ping test.

Which Test Should You Actually Use?

Work down this list and stop when you have the confidence the decision needs.

Curiosity, an inherited box of jewelry: stamp, then a photo identification, then a magnet. Free, non-destructive, takes two minutes.
Buying secondhand from a stranger: add weight against size and a streak test. If the seller refuses a streak test on an unglazed tile, that is your answer.
Selling for scrap: get an acid test or XRF before you accept an offer, so you walk in knowing the karat rather than taking the buyer's word for it.
Insuring, probate, or anything over a few hundred dollars: XRF, and for bullion add dimensions and a ping test. Then get a written appraisal.

The pattern is simple: free tests rule things out quickly, and only destructive or laboratory tests rule things in conclusively.

Testing Awkward Items

Several very common items defeat the standard tests, and knowing which is which saves you from a false conclusion.

Hollow chains and tube bangles

Specific gravity fails completely here, because the trapped air makes the piece read far lighter than its karat suggests. A genuine hollow 22K bangle can produce a density figure that looks like brass. Use the stamp and an acid or XRF test instead, and treat any density result as meaningless.

Pieces with stones

Gemstones, pearls and enamel all have their own densities, so they corrupt both the weight-against-size check and specific gravity. For valuation, dealers weigh the piece, estimate and subtract the stone weight, and price only the net metal. Note that the stones themselves may be worth more than the gold, so never scrap a set piece without a separate stone appraisal.

Watches

Watch cases are the single most common place to find heavy gold plating passed off as solid. Check the case back, the lugs and the clasp separately: it is entirely normal for a solid gold case to sit on a plated bracelet, or the reverse. Marks like "20 microns" or "plaqué or G20" indicate plating thickness, not solid gold.

Coins

Never file, streak or acid-test a coin before checking its numismatic value. A scratch can wipe out a collector premium many times larger than the metal is worth. Coins have published diameters, thicknesses and weights, so measure first and test only if the measurements are already suspicious.

How Fakes Are Actually Built

Knowing the construction tells you which test will catch it.

Plated base metal. A micron-thin gold layer over brass or copper. Caught by wear-through at the edges, weight, and a filed acid test. Passes a surface acid test.
Gold-filled. A bonded layer, typically 5% of total weight, marked GF or 1/20 12K GF. Legitimately contains recoverable gold, but is not solid gold and is worth a fraction of it.
Vermeil. Gold over sterling silver. The US Federal Trade Commission's jewelry guides (16 CFR Part 23) reserve the term for a sterling silver base carrying a gold coating of at least 10 karat and a minimum thickness of 2.5 microns. The core has real value, so it feels reasonably heavy and is not magnetic. Caught by the stamp (925) and by acid.
Tungsten core. The serious one. Tungsten's density of 19.25 is almost identical to gold's 19.3, so it defeats weight and specific gravity alike. Used mainly in large bars and coins, where the payoff justifies the effort. Caught by ultrasound, conductivity testing, a ping test, or drilling.
Forged hallmarks. A real stamp on a fake piece. Genuine marks are usually crisp and struck to an even depth; forged ones are often shallow, uneven, or use a font that does not match the claimed assay office. A stamp is a claim, never a proof.

Red Flags Worth Taking Seriously

No hallmark at all on a piece being sold to you as gold.
Green or black skin marks where the metal touches you: a copper alloy reacting with sweat, typical of plated or base-metal jewelry.
Wear-through at high-friction points (the edges of a ring shank, the back of a pendant), showing a duller, whiter metal underneath.
A price well below spot. Nobody sells real gold for a fraction of its melt value.
Bubbles, seams or a mold line on a "bullion" bar. Genuine minted bars have crisp, clean edges.

The Honest Bottom Line

Stack the cheap checks first: read the hallmark, run a photo through the app, try a magnet, and feel the weight. That combination will resolve the majority of everyday questions and costs you nothing. If the answer still matters, because you are buying, selling, insuring or inheriting, spend the extra step on an acid test or a free XRF scan at a jeweler. No app, including ours, replaces that final step.

Start With a Photo, Free

Gold Identifier names the item, reads the hallmark and shows an estimated karat range in seconds. Free on iPhone and Android.

Frequently Asked Questions

Can an app tell if gold is real?

An app can tell you what the item most likely is and what its hallmark means, which resolves a lot of everyday uncertainty. It cannot measure metal content, because a photograph carries no information about what is under the surface. Use it as a fast first pass, then confirm anything valuable with an acid test or a professional XRF assay.

Does real gold stick to a magnet?

No. Gold is not magnetic, so a strong magnet should not attract it. But passing the magnet test does not prove a piece is gold: brass, copper, lead and aluminium are all non-magnetic too. It only rules out iron and steel.

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

A specific-gravity test on a solid, single-metal item is the most accurate home method, because density correlates closely with karat. An acid testing kit is the most practical, giving you a karat band in about a minute for around $20.

Will a jeweler test my gold for free?

Most jewelers and gold buyers will run an XRF scan or acid test at no charge, since they would like to buy the piece. Getting two independent opinions before selling is normal and worth doing.

What does it mean if my gold turns my skin green?

Green marks come from copper in the alloy reacting with sweat and skin acids. It is common in low-karat and plated jewelry. Solid high-karat gold rarely causes it, so persistent green marks suggest a base-metal or heavily alloyed piece.