FL22 is a Mazda-specific long-life antifreeze fluid, specifically designed to last 10 years and 120K miles.
Any cheap replacements from the Walmart’ shelf will not be an equivalent in the protection performance and longevity. FL22 will not be cheap.
I would rather trust the manufacturer, who specifically designates FL22 for this long life application and instructs non-FL22 to be replaced much more often.
About $30/gal
Would Beck/Arnley be the ones make it for Mazda? One is Green - other Blue
https://www.pepboys.com/beck-arnley-blue-extended-life-premium-antifreeze-coolant/product/1853564?quantity=1
Unlikely you will get the answer to this question.
$30 per gallon is not that bad, I would go with OEM fluid if you want to keep the car long term, any other “long-life” fluid if not.
The blue Pentosin Pentofrost A3 is $13 per gallon from Rock Auto (it looks like the blue says it’s with Type FL22, the green says it’s without Type FL22, so if it were me I’d purchase the blue)
Did u factor in shipping?
Good observation - Blue is the one with FL22
That would be for you to find out as only you know where you live.
They are a marketing/distribution company, not a manufacturer. They’ve been known for decades as a parts supplier for imported cars, especially.
Edit- @VOLVO_V70 already said what I was about to, shipping wouldn’t matter since we don’t live in the same area 
Much like spark there is a lot of nuances here too. Mazda’s FL22 is Green, whereas RockAuto’s is Blue!
Better buy direct from mazda - it only needs one gallon. saving $10 for a 10/5 year maintenance is not needed.
Spark - get the exact NGK
Thanks folks - long but great discussion.
BTW, At 85k, instead of removing the tranny hose, the dealer guy removed coolant hose. They then topped it up. I want to try the @Tester’s suggestion first. Then want to check the Youtube as to how to do it on Mazda
Any one of these. They only test the freeze point but the housing is clear so you can see the coolant.
https://www.autozone.com/heating-and-cooling/antifreeze-coolant-tester
https://www.amazon.com/Prestone-AF-1420-Antifreeze-Coolant-Tester/dp/B000BOA9RY.
I have never heard of the multimeter test that @Tester linked to but I’m going to try it. I have found that as the additives get weaker or depleted, the coolant turns cloudy, but that is because the additives have actually depleated to the point that the metal is already corroding and causing the cloudiness. If the multimeter test catches that sooner, then I think it would be better.
Edit: let me add the following. The initial cloudiness is aluminum. Aluminum is pretty high on the activity scale so it will be the first to start corroding. If your coolant looks like muddy water, that’s iron and you have a serious problem. With just a little cloudiness, I would only do a drain and refill. I don’t like flushing a cooling system but if it is muddy, a flush would be in order.
That method is also good for the bake fluid water presence testing.
You have to have a decent multimeter, the one with MegaOhm range, ideally spanning into 10-20 MegaOhms.
DVOM–galvanic reaction test: The majority of today’s braking systems use a combination of dissimilar metals. Manufacturers use aluminum in pistons and housings, and steel in brake lines and some wheel cylinders. When moisture mixes with brake fluid, a galvanic reaction (corrosion) can occur. The higher the moisture content in brake fluid, the higher the galvanic reaction and the greater the erosion/corrosion it causes. The DVOM–galvanic reaction test uses a DVOM to measure the voltage created by the galvanic reaction due to the level of moisture in the fluid
You measure for voltage.
Just like with the coolant.
Tester
What voltage should you get for Brake fluid?
Should the engine needs to be warm?
Here’s an article on doing the coolant galvanic reaction test.
The brake fluid test should read no more than .3V, according to 2 sources I found.
OK, I’ve been doing some research on this and the ritual above is pretty much the consensus of the various sources I have found. 200 mV seems to be the upper limit for HOAT/OAT coolants.
But the definitive test is a pH test. The pH needs to be between 6.5 and 9.0. Outside of that range and you have ZERO protection. Also the article above shows putting the + (red) lead into the coolant, that results in a Negative voltage. Most people use an auto-ranging multimeter these days so which lead you use really doesn’t matter but if you have an analog meter, put the + lead on the negative battery post or ground and the - (black) lead into he coolant. BTW, in one test on a Silverado, the DEXCOL tested well above 9.0 but the voltage test was 0.350 (350 mV) so I think the 400 mV limit maybe outdated for modern coolants.
It also doesn’t matter if the engine is hot or cold or if it is on or not. It is probably best done engine off so you are less likely to get a false reading from a poor ground. I’d check on and off because if you have a poor ground, it can cause electrolysis to occur even if the coolant is good.
I tested two vehicles this weekend, my daughters 2014 Camry (90k) and my 2014 Subaru (114k). Both have the factory fill, never been changed. In the Camry’s owners maintenance and warranty booklet, there is no interval for changing the coolant, out to 12 years or 144k miles, but it called for the coolant to be checked every 30k. It does not define “check”. The Subaru interval is 11 years or 110k miles.
Both were ever so slightly cloudy, like one drop of milk in a gallon of premix. Normally this meant start making plans for a coolant change. The voltage test kinda surprised me though. The Camry tested at 170mV (0.17V) which is OK for now. The Subaru tested at 200mV (0.2V) which is a little. more concerning. BTW, I tested it at operating temp with the engine on and with it off, the results were consistent. I tested the Subaru cold. I’m going to test it hot just to see if that makes a difference.
Anyway, I feel this confirms my cloudiness theory as a method of inspecting coolant, but my standards will be a little more stringent from now on. I’m going to be replacing the coolant in both vehicles this fall when the weather is a little cooler.
Edit: Tested the Subaru hot and got around 180mV (0.18V) so it appears that temp may make a small difference. Next test will be this fall when I change the coolants in each vehicle and see what fresh coolant measures.
One more edit: tested the 07 Silverado, 242k miles. Last coolant change with Prestone Dexcool was 6/18 @ 210k miles. Cold test 250mV (0.25V). Umm, grounds looked a little corroded so I think I will clean them up and test again, later. BTW, I suspect that this was a Super Storm Sandy salvage that was not disclosed when I bought it because of the amount of surface corrosion on all the aluminum parts. No body rust though.
@tester @keith
Does it matter what coolant type is?
My Mazda used FL22 - what volt should I observe?
Acura is 1999 - it uses the regular coolant - what volt should I observe?
@keith - you are saying volt drops when the car heats up - so cold voltage is below 400mV then there is no need to warm up? How about pH levels - did u check?
Hoat/Oat seem to be all pretty much alike. I found references where GM (Dexcool), Toyota, Hyundai, BMW and Nissan were tested. All specified pH 6.5 to 9.0 and 200 to 250 mV as the max voltage.
I observed only a 20mV difference between cold and hot so I would be very concerned if I read anywhere near 400mV. I am really concerned about my Silverado but first I need to clean up the ground straps to be sure that the voltage isn’t from bad grounds.
Somewhere around here I have a pH meter but haven’t found it yet. I think my wife borrowed it and it is somewhere in her gardening stuff so finding it will be like looking for a needle in a black hole. I know, mixed metaphors.
400mV is what is in the link from @Tester
https://www.familyhandyman.com/project/coolant-testing-with-a-multimeter/
So there is varying info - any definite answer?
I’m not sure how old the reference that @Tester used is but every other reference I came across specified 200-250mV as the upper limit. Most listed 200mV as the max they would allow. Also @Tester’s reference came from Handyman Magazine. We have subscribed to that magazine and it is focused on household projects, very little automotive focus.
It could be an old article that was using the old antifreezes before OAT/HOAT became the norm. BTW, I am not trying to be critical of @Tester, his link certainly opened the door to more research on the subject. So I thank him for that.