Valuing Jewelry Accurately

Is Your Gold Really Gold? 5 Tests to Verify Karat

By Michael Tanguma, Founder & CEO of Heirfolio. Reviewed by Diana Cruz, GIA Graduate Gemologist. Updated May 25, 2026.

TL;DR. Five tests can verify the karat of a gold piece, ranging from free (the magnet test) to professional (X-ray fluorescence). The home tests rule out fakes; the professional tests give you the exact karat. Use them in order, stop at the first definitive answer, and never sell a piece without knowing its karat — the verification typically swings the offer by 20–60%.

The failure mode is the assumption.

A piece is stamped "14K" so the seller assumes it's 14k gold. The buyer's testing equipment says 12.4k. The offer comes in 12% below what the seller computed. Or the opposite: a piece with no stamp is assumed to be costume jewelry and tossed; the same piece, tested, turns out to be 18k Italian gold worth $2,800.

Both errors are common. Both are preventable with 30 minutes and at most $25 of testing equipment.

This article walks through the five tests in order of cost and definitiveness, what each can and can't tell you, and the cases where you should skip straight to professional testing.

→ Document each piece with a verified karat reading — we run XRF on every item received

Why does karat verification matter?

Three reasons.

Reason 1: The dollar swing is large.

A 20-gram chain stamped "14K" sells for roughly $1,140 at melt (at illustrative spot of $97/g pure gold). The same 20-gram chain, if it turns out to be 10k, sells for $812 — a 29% reduction. If it turns out to be 18k, it sells for $1,455 — a 28% increase. The stamp is a starting point, not a guarantee.

Apparent karat	Actual karat	Difference in payout
14k stamped	10k actual	-29%
14k stamped	12k actual	-14%
14k stamped	14k actual	0% (baseline)
14k stamped	18k actual	+28%
Unstamped (assumed scrap)	18k actual	from ~$0 to full value

Reason 2: Some pieces aren't gold at all.

Plated pieces, gold-filled pieces, and outright fakes are common in inherited collections. A heavily plated brass chain looks identical to a 14k gold chain in a photo. A buyer will pay melt for the chain; a seller who doesn't verify will not know whether the offer was fair until later.

Reason 3: The buyer's test is the only number that matters at sale.

Whatever you believe about your piece, the buyer's XRF or assay reading is what the offer is based on. If you've tested in advance and the readings disagree, you have grounds to challenge. If you haven't, you have nothing.

What do the karat markings actually mean?

Before testing, know what you're testing. Gold purity markings:

Stamp	What it means	Pure gold content
24K, 999, .999, 999.9	24-karat gold (effectively pure)	99.9%
22K, 916, .916	22-karat gold	91.7%
18K, 750, .750	18-karat gold	75.0%
14K, 585, .585	14-karat gold	58.5%
12K, 500, .500	12-karat gold (uncommon in modern jewelry)	50.0%
10K, 417, .417	10-karat gold (minimum karat to be sold as "gold" in the U.S.)	41.7%
9K, 375, .375	9-karat gold (common in U.K., not U.S.)	37.5%
GF	Gold-filled (thick layer of gold mechanically bonded to base metal)	Variable, typically 5–10% gold by weight
GP, EP, RGP	Gold-plated (thin electroplated layer over base metal)	Trace amount of gold
HGE	Heavy gold electroplate	Trace amount of gold
Vermeil	Sterling silver base with gold plating	Trace gold over 92.5% silver

Stamps to be cautious of:

Bare numbers (no K or fraction) — some legitimate, some not. "585" alone usually means 14k; "750" alone usually means 18k. But cross-check with another test.
"14KP" — plumb gold, exactly 14k by weight (a quality assertion, not a different karat).
Country marks — Italy ("Italy" or "ITALY" plus a number), U.K. (lion + crown + numerical hallmark), India (BIS hallmark), Russia (Cyrillic markings). These are typically reliable when present.

An unstamped piece is not necessarily fake — many older or imported pieces are unmarked. But it raises the question, and only testing can answer it.

The 5 tests, in order

Run these in order of cost. Stop at the first definitive answer. The early tests are cheap and rule out fakes; the later tests confirm the exact karat.

Test 1: The visual inspection (free, 5 minutes)

Before any chemistry, look.

What to look for:

Hallmark or karat stamp. On a clasp, the inside of a ring band, the back of a pendant. Use a 10x loupe or a magnifying glass.
Color. Pure gold (24k) is a deep, warm yellow with no shine. 18k yellow gold is similar but slightly cooler. 14k is a paler yellow. 10k is paler still and slightly greenish. White gold and rose gold are alloyed differently and require a separate check.
Wear patterns. Plated pieces wear at the high-contact points — clasps, ring shanks, chain clasps — revealing a different-colored base metal underneath. Solid gold wears uniformly.
Tarnish or discoloration. Pure gold doesn't tarnish. Any green, black, or brown discoloration suggests either a non-gold base metal or a heavy plating that's worn through.
Construction quality. Soldered joints, prongs, settings — real gold work is typically clean and precise. Sloppy construction is a yellow flag.

What it tells you: A piece with a clear hallmark, uniform color, no wear-through, and no tarnish is probably solid gold of the stamped karat. A piece with no hallmark, color variation, wear-through, or tarnish needs testing.

What it can't tell you: The exact karat. Visual inspection is a screen, not a verdict.

Test 2: The magnet test (free, 30 seconds)

Gold is not magnetic. Hold a strong magnet (a rare-earth or neodymium magnet, not a refrigerator magnet) close to the piece. If the piece sticks or visibly pulls toward the magnet, the core is not gold.

What it tells you: Rules out steel jewelry plated to look gold. Definitive on this question.

What it can't tell you: Whether the piece is plated gold over a non-magnetic base metal (copper, brass, nickel). A piece that passes the magnet test could still be plated over copper. Magnet test is necessary but not sufficient.

Cost: $0 if you have a magnet, $5 for a neodymium magnet on Amazon.

Test 3: The acid test ($15 home kit, 5 minutes per piece)

A small gold testing kit from Amazon or eBay ($15–$25) includes bottled acids calibrated to different karats (10k, 14k, 18k, 22k), a small testing stone (a piece of dark slate), and a brief instruction sheet.

Method:

Make a small scratch on the piece in an inconspicuous spot (the inside of a ring band, the back of a clasp). The scratch should be a few millimeters long and just deep enough to expose fresh metal.
Rub the scratched area against the testing stone to leave a thin line of metal residue.
Apply a drop of the test acid (start with 14k acid for an unknown piece) to the residue line.
Observe.

What the result means:

The line stays. The piece is at least the karat of the acid you used. If you used 14k acid and the line stayed, the piece is 14k or higher. Move to 18k acid to confirm.
The line fades or partially dissolves. The piece is below the karat of the acid. If you used 14k acid and the line faded, the piece is 12k or lower. Move to 10k acid.
The line dissolves immediately. The piece is not gold, or is heavily plated with a non-gold base metal.

What it tells you: The approximate karat, within one karat-step (10k, 12k, 14k, 18k, 22k).

What it can't tell you: Sub-karat precision (e.g., whether a piece is exactly 14k or 13.5k). Cannot distinguish between solid and heavily-plated gold of the same karat at the surface. Can damage pieces with delicate finishes or stones.

Cost: $15–$25 for the kit, good for 100+ tests.

Cautions:

The acids are corrosive. Use gloves and eye protection.
Test in a well-ventilated area.
The scratch is permanent. Test in an inconspicuous spot.
Do not acid-test pieces with delicate stones (pearls, opals, emeralds) or with patinated finishes that the test will mar.

Test 4: The electronic gold tester ($60–$200, 30 seconds per piece)

A more modern version of the acid test. Touches a small probe to the piece, reads electrical conductivity, and displays the estimated karat on a screen.

What it tells you: The approximate karat for plain solid-gold pieces, typically accurate within one karat-step.

What it can't tell you: The exact karat for plated or layered pieces (the tester reads the surface, which can be misleading). Cannot reliably test pieces with stones (the prongs and settings confuse the reading). Cannot test very thin pieces.

Cost: $60–$200 for a tester (Kee Gold Tester, Gold Tester Pro, GXL-24 Pro), no consumables.

Best use: When you have a moderate-to-large collection to test and the acid kit's per-test friction would be annoying.

Test 5: X-ray fluorescence (XRF) (professional, $0–$50 per piece)

The gold standard. An XRF machine bombards the piece with low-energy X-rays and reads the spectral fluorescence to determine the exact metallic composition non-destructively.

What it tells you: The exact karat, the alloy composition (copper, silver, palladium, nickel, zinc proportions), and the surface vs. bulk reading (which can detect plating).

What it can't tell you: Anything beneath the X-ray penetration depth (~5–50 micrometers depending on the alloy). A heavily-plated piece can read as the plating composition rather than the base metal. Most modern XRF machines compensate for this with multi-energy reads, but it's still a known limitation.

Cost: Most jewelers and pawn shops have an XRF machine and will run a test for free if you're considering selling to them. Some independent labs charge $20–$50 per piece. Heirfolio runs XRF on every piece received for sale, no charge.

Best use: Final verification before any sale, and the reference standard against which a buyer's offer can be evaluated.

When should you skip straight to professional testing?

Six cases:

Inherited pieces with no stamp or unknown provenance — the cost-benefit of buying acid kits to test a single mystery piece doesn't make sense.
Pieces with significant stones — acid and electronic tests can damage settings or stones; XRF doesn't.
Pieces with a possible high karat (22k, 24k) — the home tests are calibrated for the more common karats; high-karat pieces deserve a more precise read.
Pieces likely worth over $5,000 — the precision of XRF is worth it for any meaningful sale.
Pieces being sold to a buyer who hasn't done their own test — you want documentation of karat before the sale conversation begins.
Pieces being evaluated for insurance or estate filings — formal documentation requires a credentialed test.

How to weigh a piece accurately

A separate but related question. Karat × weight × spot price = melt value. The weight matters as much as the karat.

For most pieces: a small kitchen scale that reads to 0.1 grams ($15 on Amazon) is accurate enough. For very small pieces (a single earring, a small charm) a jewelry scale that reads to 0.01 grams ($20–$40) is better.

Tips:

Zero the scale before each measurement.
Weigh on a hard, flat surface — soft surfaces produce variable readings.
Weigh each piece individually if you want piece-by-piece data; weigh the whole collection together for a quick aggregate.
Stones add weight. If a piece has stones, weigh it, then have the stones' weight estimated by a jeweler or estimated visually (a 1-carat diamond weighs 0.2 grams). Subtract the stone weight from the gold weight when computing melt value.

See How Much Is 14k Gold Worth Per Gram Today? for the full weight-to-value math.

→ Get a real karat verification before you sell — no shipping required to start

What about gold-filled and gold-plated pieces?

A distinction worth getting right.

Designation	What it is	Gold content	Resale value
Solid gold (10k–24k)	Gold alloyed throughout	41.7%–99.9% by weight	At or near melt
Gold-filled (GF)	Thick gold layer mechanically bonded (typically 1/20 or 5% of weight)	5–10% by weight	10–20% of solid-gold melt
Heavy gold electroplate (HGE)	Thicker electroplated layer	Trace	Near zero
Gold-plated (GP, EP)	Thin electroplated layer (typically 0.5–2.5 microns)	Trace	Near zero
Vermeil	Sterling silver base with gold plating	Trace gold over 92.5% silver	Silver melt value only
Rolled gold plate (RGP)	Older term for gold-filled, sometimes thinner	5% by weight typical	5–15% of solid-gold melt

For sellers: gold-filled pieces have some recoverable gold value but are not worth treating like solid gold. Plated pieces have effectively no metal recovery value — the resale value, if any, is in the design or the brand, not the metal.

For buyers: a piece stamped "GF" or "GP" is not solid gold. Any offer that treats it as solid gold is either confused or fraudulent.

Common counterfeits and how to spot them

A few specific patterns:

Pattern 1: Brass or copper with heavy gold plating.

The piece looks gold, feels heavy, and may even have a fake karat stamp. The magnet test passes (brass and copper are non-magnetic). The acid test fails — the plating dissolves, exposing the base metal.

Pattern 2: Tungsten core with gold plating.

Used in counterfeit bars and coins, occasionally in jewelry. Tungsten has roughly the same density as gold, so the piece feels "right" by weight. Detection requires XRF or destructive testing (drilling a small hole and examining the interior).

Pattern 3: Mismarked stamps.

A real piece with a stamp that says "18K" but is actually 14k. Detection requires acid or XRF testing — visual inspection cannot distinguish.

Pattern 4: "Mexican gold" and other regional misnomers.

Some 1970s-era costume jewelry was sold as "Mexican gold" and similar terms, with no actual karat verification. Acid or XRF testing reveals the truth — often plated brass or a very low karat (5k or 6k) that isn't even legally "gold" under U.S. labeling rules.

Pattern 5: Replica branded pieces.

Counterfeit Cartier, Tiffany, Van Cleef, and other designer pieces are increasingly common. The metal may be real gold of the right karat (sometimes), but the piece is not authentic to the brand and is worth only the metal value, not the brand premium. Authentication requires brand-specific expertise — the brand's own service centers, or specialized authenticators.

How Heirfolio approaches karat verification

Briefly.

Every piece received for sale through Heirfolio is tested by XRF on intake. The XRF reading is recorded in the piece's file, along with the photographs, the weight, and the date. The offer is computed using the XRF karat, not the stamped karat — if they disagree, the XRF wins.

For documentation-only services (pieces you're not selling, just adding to your inventory), Heirfolio's pipeline uses a combination of stamp recognition (AI), photo-based assessment (AI), and optional gemologist review (human) to estimate karat without requiring the piece to leave the house. For pieces over $1,500 in estimated value, an in-person XRF reading is recommended before any sale.

Free tier covers documenting up to five items. Vault tier ($29/month) includes XRF reading on intake for any piece sold through the platform.

→ Document every piece once, with the karat reading attached forever

Frequently asked questions

How can I tell if my gold is real at home?

Start with the magnet test (free, 30 seconds). Gold is not magnetic; if a strong magnet pulls the piece, the core is not gold. If the magnet test passes, run an acid test with a $15 home kit. The acid will either dissolve the test residue (revealing a lower karat or a non-gold metal) or leave it intact (confirming the karat). For final certainty, take the piece to a jeweler or pawn shop with an XRF machine — most will test for free if you're considering selling to them.

What's the most accurate way to test gold?

X-ray fluorescence (XRF) is the gold standard for non-destructive testing. The machine reads the exact metallic composition and reports the karat plus the alloy proportions. Most jewelers, pawn shops, and refineries have XRF machines and will test for free if you're considering a transaction with them. Independent labs charge $20–$50 per piece. The acid test is the most accurate home method, accurate within one karat-step.

Does the karat stamp on jewelry always match the actual karat?

Not always. Mis-stamped pieces are common, especially in older or imported jewelry. The mismatch can go either way — a piece stamped 14k that's actually 12k (under-karat) or a piece stamped 14k that's actually 18k (under-stamped, often Italian or older European work). Verification by acid or XRF testing is the only way to know. The dollar swing from a mismatch can be 20–60% in either direction, which is why verification before sale is high-leverage.

What does 585 mean on gold jewelry?

It's the European decimal notation for 14k gold — 585 parts per thousand of pure gold, or 58.5%. Equivalent stamps in other notations: 14K, .585, 14KT. Similar marks: 750 = 18k (75% pure), 916 = 22k (91.7% pure), 999 = 24k (99.9% pure). The decimal notation is more common on European and Middle Eastern pieces; the karat notation (14K, 18K) is more common on U.S. pieces.

Will a magnet test tell me if my gold is real?

It will tell you if the core metal is non-magnetic, which is a necessary condition for solid gold. But many non-gold metals (copper, brass, aluminum) are also non-magnetic, so a piece that passes the magnet test could still be plated gold over a non-magnetic base metal. The magnet test rules out one specific kind of fake (steel jewelry plated to look gold) but doesn't confirm the piece is solid gold. Pair it with an acid test or XRF for actual verification.

Can a jeweler test my gold for free?

Often yes, especially if you're considering selling to them or having work done. Most jewelers have an XRF machine and a few minutes to run a test. Some will charge a small fee ($10–$25) if you're not transacting with them. Pawn shops also typically have XRF machines and will test pieces they're considering buying. The test is non-destructive — the piece is unchanged by the testing.

What's the difference between gold-filled and gold-plated?

Gold-filled (GF) pieces have a substantial layer of gold mechanically bonded to a base metal, typically 5% of the piece's weight in gold. Gold-plated (GP) pieces have a much thinner layer of gold electroplated onto a base metal, typically only a few micrometers thick. Gold-filled has some recoverable metal value (10–20% of solid gold melt value); gold-plated has effectively none. Both are properly stamped — GF, GP, EP, RGP, HGE — and should not be confused with solid gold.

How accurate are home gold testing kits?

Home acid testing kits are accurate within one karat-step (e.g., a piece will correctly test as "14k or higher" or "10k to 14k range") when used correctly. They cannot give sub-karat precision (e.g., distinguishing 14k from 13.5k). They cannot reliably test pieces with stones or delicate finishes. For pieces worth over $1,500, an XRF reading from a jeweler or refiner is meaningfully more precise — and typically free.

Can fake gold pass an XRF test?

Rarely, but it can happen with tungsten-core counterfeits where the gold plating is thick enough to mask the tungsten in the X-ray penetration depth. Modern XRF machines compensate with multi-energy reads, but the limitation is real. For pieces where counterfeiting is a serious concern (large bars, high-value coins, suspicious provenance), additional testing methods (ultrasonic, electrical conductivity, density measurement) can confirm. For ordinary jewelry, XRF is reliable.

Should I test my gold before selling?

Yes, in nearly every case. The karat reading swings the offer by 20–60% in either direction. Verification before the sale conversation begins gives you the data to evaluate any offer and to challenge any reading you disagree with. The home tests are fast and cheap; the professional tests are usually free if a jeweler is involved. Selling a piece without knowing its karat is the single highest-frequency way to leave money on the table in any gold sale.

What to do next

If you have a piece you're considering selling: run the magnet test first (free, 30 seconds). If it passes, run an acid test ($15 kit) or take it to a jeweler for XRF (typically free).

If you have a collection of inherited pieces: invest in a basic acid kit ($15) and a digital scale ($15). Twenty minutes per piece gets you the karat and weight, which is enough to compute melt value and evaluate any offer.

If you have any piece worth more than $1,500: skip the home tests for that piece and go straight to XRF. Most jewelers will run the test for free if you're considering selling to them; Heirfolio runs XRF on every piece received for sale.

If you want every piece in your collection documented with verified karat readings, kept in one private inventory: that's what Heirfolio's Heir Protocol is for. The documentation lives once and is exportable to insurance, to estate documents, or to a buyer's offer comparison.

The stamp is a starting point. The test is the verdict. Don't accept an offer until you know what you have.