Reviews of nine electronic gold testers




This article reviews several electronic gold testers, including how they compare with using chemical (acid) testers.

AuRACLE AGT3 by GemOro


* Price * Ease-of-use * Accuracy *
Scroll down to see how they compare, in detail

Mizar M24 £159.00 including VAT. Has eight lights: 9ct, 10ct, 12ct, 14ct, 16ct, 20ct, 22ct and 24ct. Accuracy is plus or minus one light, e.g. if it reads 14ct it might be 12ct or it might be 16ct. Fiddly to use; uses a strong acid as a contact fluid; will only test small items.

GemOro AuRACLE AGT1 DELUXE has a row of 32 LED lights from 6ct to 24ct, so you can judge which carat is lighting up (the reading 'drifts' up an down) - but it is a matter of judgment and does take time to learn to interpret the readings.  From 9ct to 12ct it's usually accurate to within 1ct; at 10ct to 14ct it's OK plus or minus 1ct; at 14ct to 24ct, OK plus or minus about 2ct.

KEE.  This wins on being the easiest to use, for consistent reading, and for accuracy. You should (with very careful calibration) get a good reading all the way up to 24ct. Scroll down to see a review.

If the accuracy of the electronic testers listed above isn't good enough for you, use acids instead. Boxed sets of acids cost between £60.00 and £90.00, the best sets test for all carats of gold plus silver, platinum, Palladium, and steel. 

An analogy. You can get printers with individual ink cartridges, one for each colour, when one colour runs out you replace that cartridge; you can also get all-in-one ink cartridges, when one colour runs out you have to replace the entire cartridge.

The probe of an electronic gold tester, about £50.00, is the equivalent of an all-in-one ink cartridge (one probe for all carats); bottles of acid, about £12.50, are the equivalent of individual ink cartridges, a different bottle for each carat/metal.


This is the best of the MIZAR models.


Plus or minus one light, e.g. if it reads 14ct it might be 12ct or it might be 16ct but not this accurate above 18ct. OK if you want to know if an item is "low carat" or, possibly, "high carat" - otherwise, fairly useless.


- very cheap
- very simple, a light for each carat
- Mizar make a few models, this is by far the best of them


- not accurate.
- the acid contact fluid will dissolve the sensor if it is not cleaned meticulously after every use
- you need good eyesight and a steady hand to place the test-item into the tiny well that you fill with acid
- you can only test tiny items
- no carrying case, no mains power option, no way of calibrating it.


Not the best, but certainly the cheapest. 



Click here to buy. 
(or click here if you already have one, and need a replacement probe)


You should (with careful calibration) get a reading to the nearest 1ct all the way up to 24ct.


  • good price for a tester that works well
  • the easiest to use. You still have to read the instruction manual but apart from calibrating (and cleaning the sample) it's self-explanatory, you won't have to keep referring to the instruction manual to find out about menus and error messages
  • fastest response of all the electronic gold testers
  • it's more accurate than the other models at testing high-carat (e.g. for 20ct to 24ct 'Asian' gold)
  • good size for carrying around (supplied with a pouch that takes the tester and all the accessories)
  • extra set of leads for testing large items


  • no option for a mains power supply, it's battery-only (though the battery should last many months)
  • very large items of jewellery won't fit on the test plate, you have to use the extra set of leads (though I've never known anyone to need to use them)
  • not reliable for testing platinum, the best you can say of a "platinum" reading is, 'It might be platinum or Palladium".


If you think anything with a digital display must be more accurate than an old-fashioned mechanical meter - think again. There is no connection between the type of display (digital or analogue) and accuracy. A digital display will give you a reading to the nearest 1ct, but it is not necessarily accurate. 


This is definitely my favourite, we've been selling this since 2018, it is the most accurate of the electronic gold testers, nobody has complained that they don't understand how to use it, and it is the most reliable (very few break down in the first year).  

Download the instruction manual.


Update: although we discontinued these a few years ago we still sell the replacement probes. The AGT1 testers are no longer for sale in the UK due to problems with electrical compliance. 

Very accurate from 9ct to about 12ct (usually correct to within 1ct); reasonably accurate for 12ct to about 14ct (plus or minus 1ct); somewhat approximate between 14ct and 24ct (plus or minus about 2ct).


  • remarkably cheap for an electronic gold tester that actually works.
  • very responsive once you've 'got the knack' of how to interpret the readings


  • not easy to use. It works well once you've got the 'knack' for how the lights shoot up, then slowly go down, and usually settle on a reading...but not everyone has the patience to spend time 'getting the knack'. Shop staff who don't get proper training and are under pressure in a busy shop - they will have no chance of getting a meaningful reading.
  • each time you switch it on you must calibrate against the 14ct calibration sample and sometimes you need to do this two or three times before it will work
  • it helps if you 'charge' the probe by carrying out a few 'dummy' tests before getting a reliable reading.


Do not buy this if you have to ask, "Is it easy to use". Do not buy this if it will be used by a succession of shop staff who don't have time to learn. But it's OK if you are prepared to persevere until you 'get the knack. My colleague prefers this to the up-market version (AGT3) because he can see how it is behaving rather than trusting a computer readout.

(click here if you already have one, and need a replacement probe)

AuRACLE AGT3 by GemOro

Although we stopped selling these a few years ago, we have an account with GemOro and can get them to special order, ask for latest price and delivery time. 


  • Instead of a row of LED lights (as with GemOro's earlier models) it has a digital display. It is not just having a digital display that makes the difference, it's the way they have computerised the on-display instructions. It tells you when to calibrate; when testing it tells you to wait a few seconds and it tells you when the reading has stabilised and the test is complete; it tells you if the test has failed and you need to try again. All of this removes the element of 'judgment' that you need with the earlier models (judging when the row of LED lights have stabilised and you have a valid reading).
  • Easy calibration, you press one button and follow the instructions, easier than having to turn a knob (I suspect that people who have been brought up pressing buttons don't know what to do when presented with a knob).
  • There are two modes of testing. A 'quick' test will indicate the approximate carat (e.g. 10ct [because it's American], 14ct, 18ct or 22ct); a detailed test will give a reading to the nearest 1ct. (i.e. it gives a reading to the nearest 1ct, that does not mean it is accurate to the nearest 1ct).
  • The readings are very clear, no interpretation is necessary, when buying gold you can show the seller the display as 'proof' (though there are limitations, see below).


  • The 'prompts' on the screen are very good reminders, so that you don't forget what to do, but you do have to read the instruction manual carefully. The instruction manual is beautifully illustrated but very badly written, you will have to read it several times.
  • The system often crashes or the display goes corrupt (it shows bands of lines or goes blank) - you have to restart it a few times to get it to work. I had dozens of email exchanges with GemOro about this and they say it's due to the "environmental conditions" in the UK, it is not something that can be fixed.
  • In Enhanced Mode the readings you get are very precise, to the nearest 1ct. Giving a reading to the nearest 1ct is not the same as being accurate to the nearest 1ct. At low carats the reading is usually accurate (correct) to plus or minus 1ct. At high carats the reading is usually accurate (correct) to plus or minus 2.5ct. If you are buying scrap from the general public in the UK (nearly all items are 9ct, occasionally 14ct or 18ct) this is not a problem. However, it is not suitable for testing "Asian" (high-carat) gold.
  • It is wise to take two or three readings to see if they are all the same, and if they are not, just work on an average. Getting a different reading each time is not a 'fault' and is not covered by the guarantee.
  • Theoretically it tests platinum, but this is not reliable, the reading sometimes indicates platinum and sometimes indicates not gold. And Palladium usually  reads 'platinum'.
  • Large. Carrying case measures 300 X 150 X 75mm (12 X 10 X 3 inches), not something you can slip into a bag.
  • Guarantee is for six months only. That means if there is a problem we must receive it in our office within 6 months. The suppliers are not helpful, if we receive it in 6 months + 1 day, it's out of guarantee (they expected you to throw it away and buy a new one).
  • Unlike our regular products, no support is available with this special order item.


How do they compare for accuracy?

Most scrap-buying in the UK involves 9ct gold, occasionally 14ct or 18ct, rarely 22ct. However, there is also a market for 'Asian gold' and traders need to know if an item is 20ct, 21ct, 22ct or 23ct. Most of these electronic simply aren't accurate enough to do this, the only one I would recommend is the Kee tester (if you calibrate it very carefully it is accurate to the nearest 1ct). Other than the Kee tester, if you want to test high carat gold, use acids, they are more accurate. 

M24: don't even think that you can get a meaningful reading on high carat gold. It might be OK if all you want to know is, "It's gold, probably a low carat" or "It's gold, it might be about 14ct but we don't really know" or "It's gold, it might be high-carat".

AGT1: quite accurate for testing low carat (e.g. 9ct to 14ct); for high carat it is accurate to plus-or-minus about 2ct. 'Reading' the lights is a matter of judgment.

AGT3: quite accurate for testing low carat (e.g. 9ct to 14ct) but for high carat it is only accurate to about plus-or-minus about 2ct (even though the display will always give reading to the nearest 1ct).

Kee: the most accurate for all carats (including high carat) providing you calibrate it very carefully.


If you want to test more accurately than listed above, then there is no point in buying any of the low-cost electronic testers (listed above), either use acids or splash out on a £12,000.00 XRF tester.

How do they compare for size of the machine?

The models that connect to a phone are the smallest [now discontinued] but not quite as small as the photographs suggest, you also have the connecting leads and a box of electronics. After that, the M24 is the smallest (there is no carrying case)followed by the Kee (in its carrying pouch). The Gemoro (AGT models) are larger and come in huge carrying cases.

How do they compare for the size of items you can test?

The M24 has a tiny 'well' (indentation) which you fill with acid, you must 'dip' the item into this, that limits you to testing very small items. The Gemoro (AGT) testers have a good-size test plate which will take all but the largest items of jewellery. The Kee has a smaller test plate (large enough to test most jewellery) and also comes with a second set of connecting leads for testing items that don't fit onto the test plate.

XRF testers (also known as "x-ray guns") cost between £12,000.00 for a 'basic' handheld model, to £50,000.00 for a good quality desk model. They display a chemical breakdown of the metal, listing not only the percentage of gold but also the percentage of all the other metals.  An example: an XRF tester might tell you that the metal contains 44.3% gold, 14.03% silver, 0.6% platinum and 0.05% Palladium. Our electronic testers will tell you it's about 9ct (37.5% gold).  We do not sell XRF testers.

How do they compare for speed of testing?

AGT1: slightly faster than acids.

AGT3: when set to enhanced mode is about the same speed as acids, though there's not a lot in it, and it depends on the carat. The quick mode is much faster than acids but isn't accurate.

KEE: faster than acids.

XRF: has quick and enhanced modes. On enhanced, it is slower than both AGTs and acids (the videos always show them in quick mode!)

How do they compare for safety?

AGT and KEE testers use a probe containing a harmless salt solution, the M24 tester uses acid.

XRF testers use x-rays, when used properly they present no risk. However, we see them used by dealers who hold items in their hands whilst testing, and so give themselves a dose of radiation several times per day. There is a high chance that this will give them cancer. Needless to say, they should never be pointed at a customer!

Acids must be handled with care. Providing it isn't spilt on skin (simple solution: wear gloves) the level of exposure becomes nil.

How do they compare for consumables? (prices include VAT)

AGT pen-probes and Kee pen-probe cost about £50.00. They are like felt pens, they contain a harmless salt solution. Both manufacturers say they last "up to" a few thousand tests. AGT pen-probes are not guaranteed, that means if it stops working after a few weeks (or a few days) you must buy a new one. Kee pen-probes are guaranteed for one month. A survey of our customers indicates that KEE pen-probes last between 3 and 18 months (anywhere from a few dozen to few thousand tests). If you can get them to last several months, it's good value; if you find they only last a few weeks then it works out expensive. We don't know why there is such a variation, our best guess is that customers don't replace the caps firmly enough and they dry out. 

A replacement bottle for a Mizar M24 costs £29.50 and lasts about 50 tests, very expensive.

Bottle of acid costs about £15.00 and lasts for 100 to 150 tests, very good value.

An analogy would be that of ink cartridges. A pen-probe is like an all-colours-in-1 cartridge, once the cartridge is 'used up' you can't print (like a pen-probe being used up) - as opposed to a cartridge for each colour (like an acid for each group of carats) where you can just replace the one that is used up.  For a busy shop this could be the difference between spending £50.00 per month or £15.00 per month on gold-testing.

XRF is different. When an XRF tube wears out, a new one will cost from £2000.00. Plus a service, plus any repairs that are needed, plus VAT...assume £3000.00.

How do they compare for service and repairs?

KEE TESTERS. We give a one year guarantee. During that year (and providing we are certain that there is a fault) we will replace it. The suppliers do not offer a repair service, though we can carry out basic checks (battery, connecting leads, calibration) and we also know how to reprogram the microchip.

We find that GEMORO (AGT1 and AGT3 TESTERS) will replace any that are faulty in the first 14 days IF they are returned in absolutely 'new' condition but not if there are any signs of use.  Their guarantee is very strictly six months. Any returns to the USA take 2 to 3 months, They do not offer a repair service, if your Gemmoro breaks down one day after six months, you are expected to throw it away and buy a new one (in the US this is expected of low-value items that cost only a few hundred pounds). 

The M24. It's very simple - not worth repairing. Also be aware that the contact fluid is a strong acid, you must clean the 'well' thoroughly after every test, otherwise the acid dissolves it.

A CHEMICAL TESTER cannot 'break down', you simply buy a new bottle of acid.

Which tester is the 'least destructive'?

Every tester (whether you spend £15.00 or £12,000.00) only tests the surface. This means you must file the surface with a fine needle-file (supplied with each tester) in a place where it won't show, to remove any surface plating. There are no exceptions. If you are not prepared to file surface then you cannot test jewellery with an electronic or acid tester, you cannot buy scrap, you cannot test anything, you cannot deal in precious metals.  

Acid (including the contact fluid of the M24) leaves a stain, this can be polished off. The AGT and KEE testers can leave a very faint circle of salt which wipes off easily, or a very slight stain if you repeatedly test the same spot (it too can be polished off).  The only time this is ever a problem is with brand new items polished to a mirror-finish, unless you are a professional polisher you won't be able to restore the mirror-finish. XRF leaves no mark (though you still have to file the item).

How do they compare for testing silver, platinum and Palladium?

Mizer and AGT testers can only test for gold. KEE testers can also test for platinum (with care). Acids can also test for silver, steel, platinum and Palladium. XRF can test for several more metals including copper, tin and zinc.

How do they compare for calibration?

Every gold tester has to calibrated. Calibration is the matching of the electronics to an actual sample. 

If you haven't used a bottle of acid for some time (they get weaker over the weeks and months) you must check it against known samples (e.g. any item with a UK hallmark).

The AGT testers need to be checked against a 14ct sample, you press some buttons, the electronics calibrate the machine to the sample. You will need to do this whenever you turn the machine on. This is frustrating when you are testing dozens of items for a customer and it keeps telling you to calibrate. Sometimes the calibration procedure fails and you have to do it again...and again...and maybe a  fourth or fifth time.

The KEE tester must be calibrated against a sample of 18ct, you test the item whilst turning a knob until it reads exactly 18ct (and to test 20ct to 24ct accurately you must also calibrate it against a 22ct sample too) - turning a knob is very simple. You do not have to do this every time but it's a good idea to do it more often as the fluid in the pen-probe gets low.

The M24 tester cannot be calibrated, you just have to hope the readings are more or less correct and don't 'drift' over time.

An XRF tester must be sent back to the supplier if it needs calibrating. If that's all that is required, it will only cost a few hundred pounds, but a repair can easily cost £2000.00 or £3000.00. 


You can also buy a replacement probes.

Please scroll up to find out how the GemOro AuRACLE testers (all AGT models) compare with other gold testers. Please note, this section is for information only, these models are no longer available.

Not all AGT1s are the same.

AGT1 (original model)

Quite OK once you've got the 'feel' for how the lights shoot up, then keep going up, then slowly go down, and usually settle on a reading...but many people try for a few minutes then give up. Shop staff who don't get proper training and don't have time to 'get the feel' and are under pressure in a busy shop - they will have no chance of getting a meaningful reading.

Each time you switch it on you must calibrate against the 14ct calibration sample, sometimes it doesn't work and you have to calibrate it 2 or 3 times, then you must 'charge' the probe by carrying out a few 'dummy' tests before getting a reliable reading, and all of this takes a couple of minutes. Every few days (if you're unlucky) or every few weeks (if you're lucky) it has a nervous breakdown and won't work at all, it is not 'broken', you can fix it by calibrating it 6 or 7 times in succession, it's not difficult to do but it is frustrating, especially if you have a queue of customers. Not very accurate at testing above 18ct.

AGT1 Plus

Compared with the AGT1 (original) above, calibration usually works first time, it has nervous breakdowns less often, it 'feels' nicer to use (more responsive). You still have to 'charge' the probe by carrying out a few 'dummy' tests before getting a reliable reading. It's not as frustrating to use as the original AGT1 but there will still be times when calibration doesn't work the first time and you have to calibrate it 3 or 4 or 5 times in succession. You still need to spend time 'getting the feel' of the readings.

AGT1 Deluxe

The full name given by the manufacturer's is AGT1 PLUS DELUXE

The supplier says that the electronics were re-designed and you don't have to 'charge' the probe before use...but they also say that you should charge the probe anyway. We couldn't find any difference between this and the earlier models as regards operation.

CONCLUSION: these are not, "Simple to use." However, if you are happy with their idiosyncrasies and are prepared to spend time and effort 'getting the feel' - then the AGT1s are quite OK. 


This consisted of a probe, a testing plate, a box of electronics (not shown in the manufacturers pictures), connecting leads and a large fitted case. You connected it to your phone. It was discontinued in 2018. Then they discontinued the app. - that meant if you updated the operating system on your phone or got a new phone, you could no longer use it.  Although this is all 'ancient history' it's worth bearing in mind in case you are offered a secondhand one.