Gold fakery is as old as gold, and modern plating is good enough to fool the eye. The good news: gold’s physical properties — density, conductivity, chemical inertness — are nearly impossible to counterfeit all at once. Here are seven tests, ordered from ten-second checks to lab-grade certainty. Use the early ones to rule out, the later ones to confirm.
1. Read the stamp
Karat or millesimal marks (14K, 585, 750) are where to start — decode them with the karat chart. Marks like GF, GP, HGE or 1/20 mean plated or filled: near-zero gold. But stamps can lie, so a mark is a claim, not proof.
2. The magnet test
Gold is non-magnetic. Use a strong neodymium magnet: a pull means steel or iron under the surface — fake. No pull proves little (brass and copper pass too), so this is purely an elimination test. Watch for items where only the clasp is magnetic; clasps are often steel-sprung even on genuine pieces.
3. The ceramic streak test
Drag the item firmly across an unglazed ceramic tile. Real gold leaves a gold streak; pyrite and most fakes leave a black or greenish one. It leaves a faint scratch on the item, so use an inconspicuous spot.
4. Weight and dimensions (for coins and bars)
Bullion is minted to tight specs. Weigh to 0.01 g and check diameter and thickness with calipers against the published numbers (our bullion comparison tablehas weights). A “Gold Eagle” that weighs 31.1 g instead of 33.93 g is busted in ten seconds. Counterfeits that get the weight right are usually too thick or too wide, because no cheap metal matches gold’s density.
5. The density test
The classic. Weigh the item dry, then suspended in water; dry weight ÷ (dry − wet) = specific gravity. Pure gold reads about 19.3, 22K around 17.8, 14K roughly 12.9–14.5 depending on alloy. Brass (8.5) and lead (11.3) fail badly. The one metal that passes — tungsten, 19.25 — is why thick bars deserve test #7.
6. Acid test kits
Jeweler’s kits (~$15) include nitric-based solutions for each karat. Rub the item on the kit’s touchstone and apply acid: the streak dissolves if the metal is below that karat, holds if at or above. Accurate to roughly ±2 karats and the standard pawn-shop method. Wear gloves, work ventilated, and expect a tiny mark on the piece.
7. Electronic and XRF testing
For certainty without damage: XRF analyzers read elemental composition through the surface in seconds (most bullion dealers will scan an item free or cheap), and precious-metal verifiers like the Sigma Metalytics use electrical conductivity to check the metal inside a coin or small bar — the practical defense against tungsten cores, alongside ultrasonic thickness gauges for large bars.
Once it tests real
Weigh it, note the karat, and get the number that matters: melt value. The scrap gold calculator handles jewelry by karat, and the US and world gold coin calculators handle coins — all at the live spot price. Then read what buyers actually pay against melt before you sell.
Frequently asked questions
Will real gold stick to a magnet?
No — gold is not magnetic. A strong pull means iron or steel underneath. But passing the magnet test proves little, since brass, copper, lead and tungsten are also non-magnetic. Use it only to rule items out.
Can fake gold pass a density test?
One metal can: tungsten, whose density (19.25) is almost identical to gold's (19.3). Tungsten-core fakes exist mainly as bars and thick coins. Dimension checks, ping tests and ultrasound or XRF at a dealer catch them.
Does the bite test work?
It is a myth from gold-rush movies. Pure gold is soft enough to mark, but so is lead — and you would be biting jewelry-grade alloy anyway. Skip it; a ceramic streak test is just as fast and actually means something.