How to Test Gold at Home: Methods From Free to $25

Every Method at a Glance

Learning how to test gold at home is really learning a funnel: cheap, fast checks eliminate obvious fakes, and each step up in cost buys certainty. Here's the menu before we walk through it dish by dish:

No single home method is conclusive alone. A piece that passes the magnet, density, and acid tests together, with a coherent hallmark, is very probably what it claims to be. A piece that fails any one of them deserves suspicion regardless of what the stamp says.

Free Check #1: Read the Marks First

Before touching any test, find the stamp. Check clasps, the inside of ring shanks, earring posts, and the tag near a chain's clasp. You're looking for karat marks (10K–24K), fineness numbers (375, 417, 585, 750, 916/917, 999), or — just as usefully — the marks that tell you to stop hoping: GF, GP, GEP, HGE, RGP (plated or filled), and 925 or "Sterling" on gold-colored pieces (that's gold-plated silver). The complete decoder ring lives in our guide to gold hallmarks.

Two honest caveats. First, a missing mark proves nothing — older American pieces and much handmade jewelry were never stamped, and worn marks polish away. Second, a present mark also proves nothing by itself, because counterfeiters own stamping tools too. The hallmark sets your hypothesis; the physical tests below check it. Think of the stamp as the defendant's testimony: worth hearing, requires corroboration.

Free Check #2: The Magnet Test

Gold is not magnetic. Neither, for the record, are silver, copper, brass, zinc, or lead — but steel and iron emphatically are, and they're the skeleton of a huge share of cheap fakes, fashion jewelry, and "gold" chains sold from car trunks.

Use the strongest magnet you can find — a neodymium magnet from a hardware store works far better than a fridge magnet. Hold the piece by a thread or let it lie on a table, and bring the magnet close. A real gold piece will not respond at all. A sharp jump or stick means steel or iron under the surface: fake, or at best gold-plated steel.

Know the test's two blind spots. A weak pull at the clasp alone is common and innocent — clasps often contain steel springs even on genuine gold chains, so test the body of the piece, not the hardware. And a piece that ignores the magnet has only cleared one hurdle: brass, copper, and tungsten fakes pass the magnet test all day. Passing means "keep testing," never "it's gold."

Free Check #3: Density, the Honest Number

Gold is famously dense: 19.32 g/cm³ pure. Alloys dilute that — 14K typically lands around 13–14 g/cm³ depending on the recipe — but everything in the gold family is still far denser than brass (~8.5), steel (~7.8), or pot metal. Density is hard to fake cheaply, which makes this the best free test for solid pieces.

The home version needs a scale that reads 0.1 g and a glass of water:

1. Weigh the piece dry. Say a ring reads 9.66 g.
2. Fill a glass, place it on the scale, and tare to zero.
3. Suspend the piece by a thread so it's fully underwater without touching the glass. The scale now shows the volume of displaced water in grams — effectively the piece's volume in cm³. Say it reads 0.70.
4. Divide: 9.66 ÷ 0.70 = 13.8 g/cm³. That's consistent with 14K. A brass ring of the same size would have read closer to 8.5.

The folk "float test" — real gold sinks fast — is just this principle minus the math; nearly all metal sinks, so floating or slow-sinking only exposes hollow or plastic-cored junk. Run the full division when the answer matters. Caveats: gemstones, hollow construction, and watch movements wreck the measurement, so use density only on solid, stone-free metal.

Free Check #4: The Ceramic Streak

Drag the piece firmly across an unglazed ceramic tile — the rough bisque back of a bathroom tile works, and tile shops will often hand you a broken sample for nothing. Real gold leaves a gold-colored streak. Plated base metal typically leaves a gray or black streak as the thin gold skin gives way and the underlying metal writes its confession.

This is the most destructive of the free tests: you are literally sanding off a microscopic amount of the piece, and it can leave a fine scratch. Use an inconspicuous spot — the inside of a bangle, the back of a pendant — and a short stroke. Don't streak-test anything with collector or antique value; for heirlooms, skip straight to a professional scan instead. As with the magnet, interpret asymmetrically: a gray streak is strong evidence of fake; a gold streak is encouraging but not conclusive, since thick plating can briefly leave gold color too.

The Skin Test and Other Myths

You'll hear that real gold never discolors skin and anything that leaves a green or black mark is fake. Mostly myth. High-karat gold is indeed inert, but 10K and 14K pieces are roughly half copper, silver, nickel, and zinc by weight — and those alloy metals can absolutely react with sweat, lotion, chlorine, and individual skin chemistry. Plenty of people get a faint dark smudge from genuine 10K rings; plenty of fakes sit on skin politely for months.

Related folklore worth retiring:

• The bite test. Theatrical, ruins jewelry, and lead bites softer than gold anyway. Olympians bite medals for photographers, not assayers.
• "It's heavy, so it's real." Heft without measurement is a feeling. Tungsten and lead were chosen by counterfeiters precisely because they feel right. Do the density math instead.
• Color judgment. Polished brass under warm store lighting fools experienced eyes routinely.

Plenty of what circulates as advice on how to test gold at home is folklore of exactly this grade. If a method doesn't produce a number or a binary physical result, treat it as entertainment.

The $20 Upgrade: Acid Test Kits

An acid kit is the classic jeweler's bench test and the best accuracy-per-dollar purchase in this whole topic — typically $15–25 for a set of bottles marked 10K, 14K, 18K, and 22K, plus a basalt touchstone.

The procedure: rub the piece firmly on the stone to leave a visible metal streak a half-inch long. Apply one drop of the acid matching the claimed karat. Read at thirty seconds: if the streak stays bright and intact, the metal is at least that karat; if it fades, browns, or vanishes, it's below the claim or not gold at all. Work up or down the bottles to bracket the true karat — a streak that survives 14K acid but dissolves under 18K acid is 14K-to-17K metal, in practice 14K. For suspected plating, file a tiny notch in a hidden spot first so the acid reaches the core; a streak from the surface tests only the skin.

Safety is non-negotiable: these bottles contain nitric and hydrochloric acid. Wear nitrile gloves and eye protection, work over a disposable surface in a ventilated room, keep baking soda nearby to neutralize drips, cap bottles immediately, and store the kit away from kids — treat it like the corrosive it is, not a craft supply. Used carefully, an acid kit will correctly sort the overwhelming majority of jewelry that crosses a kitchen table — say, confirming that an unmarked 12.4 g ring is truly 14K, which at an assumed $4,400/oz spot ($82.51/g for 14K) makes it $1,023 of melt value instead of a $4 brass mystery. A local dealer paying 65–85% of melt would offer $665–$870 for that ring; knowing the number before you walk in is the entire point.

Electronic Testers: Worth It?

Electronic pen testers measure electrical conductivity through a gel or probe and display an estimated karat. Entry models start around $25; the units pawn shops actually trust run $150 and up. For a household testing one inheritance, they're hard to justify — the cheap ones are fooled by thick plating and confused by white gold, and the good ones cost more than most testing problems are worth. For someone buying estate lots monthly, a quality unit pays for itself in speed: no acid, no marks, ten seconds per piece.

If you do buy one, calibrate it on a piece of known-genuine gold first, test several spots per item, and treat readings on plated-looking or unusually light pieces as provisional. An electronic pass plus a density pass is a strong combination; an electronic reading alone is a hint with a battery.

What Home Tests Can't Catch — and the Free Fix

Honesty about the ceiling: two fakes defeat every method above. Tungsten-cored bars are the nasty one — tungsten's density (19.25 g/cm³) is within half a percent of pure gold's, so a bar with a real gold shell over a tungsten slug passes magnet, density, streak, and surface acid tests simultaneously. This is overwhelmingly a bullion-bar problem rather than a jewelry problem, which is one reason to buy bars only from reputable dealers. The second is high-quality heavy plating over silver: silver's conductivity and a thick gold skin can fool streak tests and budget electronic testers alike.

The answer to both is X-ray fluorescence. An XRF gun reads the actual elemental composition in seconds, non-destructively, and nearly every serious coin or bullion dealer owns one. Walk in as a potential seller and the scan is almost always free — the dealer wants to know what your metal is even more than you do. (For bar-shaped doubts, dealers use ultrasonic thickness gauges to find cores that XRF, a surface technique, can miss.) Treat the free XRF as the final word after your home funnel, and as a fraud check on any deal that smells wrong; the FTC's guidance on selling gold jewelry makes the same point about verifying before you sell. If a buyer refuses to test in front of you, or a seller refuses to allow testing at all, that refusal is itself a test result — see the playbook in cash for gold scams. And once a piece is confirmed, even humble 10K is real money at recent prices near $4,300–$4,500 per troy ounce.

Frequently Asked Questions