- The sticker price and the per-gallon cost rank filters in opposite orders. The most expensive cartridge set in this guide costs the same per gallon as the cheapest single pitcher filter, because it lasts fifty times longer.
- ZeroWater’s cost is set by your water, not by the product. Its own published table shows capacity falling from 40+ gallons to under 8 as input TDS rises, a five-fold swing in running cost for the identical filter.
- RO wastewater is not a rounding error. At the EPA’s stated typical efficiency, a household producing 1,000 gallons a year sends about 5,667 gallons down the drain, which in an expensive city costs roughly as much as the replacement filters.
Why does the advertised price tell you nothing?
Because it is a price for a cartridge, and what you are actually buying is a number of gallons.
Two filters at the same price can differ by a factor of fifty in how much water they treat before they need replacing. The pitcher filter on the supermarket shelf looks like the budget option because it costs eight dollars. It is the budget option in the same way that a taxi is cheaper than a car.
The calculation that matters is simple: cartridge price divided by rated gallons. Here is what that does to the ranking.
What does each filter actually cost per gallon?
| Filter | Cartridge price | Rated gallons | Interval | Cost per gallon |
|---|---|---|---|---|
| Brita Original pitcher (6-pack) | $5.00 each | 40 | 2 months | $0.125 |
| Brita Elite pitcher (2-pack) | $16.50 each | 120 | 6 months | $0.138 |
| Aquasana AQ-5200 under-sink | $73.99 | 500 | 6 months | $0.148 |
| Brita Faucet (2-pack) | $14.99 each | 100 | 4 months | $0.150 |
| Brita Elite pitcher (single) | $19.99 | 120 | 6 months | $0.167 |
| Brita Faucet (single) | $17.99 | 100 | 4 months | $0.180 |
| Brita Original pitcher (single) | $7.99 | 40 | 2 months | $0.200 |
| Clearly Filtered 3-stage under-sink | $409.50 | 2,000 | 9 to 15 months | $0.205 |
| PUR PLUS pitcher (3-pack) | $9.54 each | 40 | 2 months | $0.238 |
| ZeroWater (4-pack), soft water | $15.50 each | 40 | varies | $0.388 |
| ZeroWater (4-pack), typical US water | $15.50 each | 25 | varies | $0.620 |
| ZeroWater (4-pack), hard water | $15.50 each | 8 | varies | $1.938 |
Look at what happened to the two extremes of price.
The $409.50 cartridge set costs 20.5 cents a gallon. The $7.99 cartridge costs 20.0 cents a gallon. They are the same price to run. One of them just asks for the money all at once.
And an under-sink system is not the expensive option it appears to be. Aquasana’s cartridge, at $73.99 for 500 gallons, comes to 14.8 cents, which beats every single-cartridge pitcher price in the table and sits within a penny and a half of buying Brita in bulk.
Why does ZeroWater cost so much more to run?
Because it is doing something different, and because its life depends on your water rather than on the filter.
ZeroWater uses ion exchange resin to strip dissolved solids down to a meter reading of zero. That is a genuinely different job from a Brita, which reduces chlorine taste and, in the Elite version, some health contaminants. Removing everything exhausts the resin quickly, and the more there is in your water, the quicker it goes.
Here is the company’s own published guidance:
| Your water’s TDS | Gallons you get | Cost per gallon |
|---|---|---|
| 2 to 50 | 40 or more | $0.39 |
| 51 to 200 (described by ZeroWater as most US water) | 25 to 40 | about $0.48 |
| 201 to 300 | 15 to 25 | about $0.78 |
| 301 to 400 | 8 to 15 | about $1.35 |
| 400 or more | 8 or fewer | $1.94 to $2.50 |
The same box of filters costs one household 39 cents a gallon and another household five times that. Nothing about the product changes. Only the postcode does.
If you live somewhere with hard water and you have seen ZeroWater recommended without this caveat, that recommendation was incomplete. Check your TDS before buying, not after.
There is also an inconsistency in ZeroWater’s own materials worth flagging. Its IAPMO-certified data sheet states a rated service life of 15 gallons, which sits below the 25 to 40 gallon band the company describes as typical, and one retailer Q&A gives 20 to 25. Three different official-looking numbers for the same filter.
What does reverse osmosis really cost?
This is where the cartridge price stops being the whole bill, and it is the part most guides skip entirely.
RO systems send water down the drain to produce clean water. How much depends on the system, and the industry’s own numbers are not the ones to use.
The EPA states the reality plainly:
Typical POU RO systems operate at efficiencies between 10 and 20 percent, meaning approximately four to nine gallons down the drain for every gallon produced.
Now put a price on that. A household producing 1,000 gallons of filtered water a year, using Boston’s 2026 combined water and sewer rate of $0.0173 a gallon:
| System efficiency | Waste ratio | Wasted per year | Cost of the waste |
|---|---|---|---|
| 30% (EPA WaterSense labelled) | 2.33 to 1 | 2,330 gallons | $40 |
| 20% (good end of typical) | 4 to 1 | 4,000 gallons | $69 |
| 15% (EPA’s stated average) | 5.67 to 1 | 5,670 gallons | $98 |
| 10% (poor end of typical) | 9 to 1 | 9,000 gallons | $156 |
For context, Waterdrop states its G3P800 costs about $155 a year in replacement filters. At the EPA’s average efficiency in an expensive city, the wasted water costs roughly as much as the filters do. The advertised running cost is about half the real running cost.
This varies enormously by where you live, and that is the honest caveat. Memphis charges about $0.0032 a gallon for water, against San Francisco’s combined rate of roughly $0.0459. The same wasted 5,670 gallons costs about $18 in Memphis and about $260 in San Francisco. In a cheap-water city, wastewater genuinely is a minor line. In an expensive one it is half your filtration budget.
Why can you not trust a published drain ratio?
Because the number is not standardised, and two manufacturers using the same figure can mean opposite things.
This is not a subtle point. Consider two systems, both advertising “3:1”:
| Manufacturer | States “3:1” meaning | Actual recovery |
|---|---|---|
| Waterdrop G3P800 | 3 pure to 1 drain | 75%, good |
| APEC (undercounter, per its FAQ) | 3 waste to 1 pure | 25%, poor |
A shopper comparing those two identical-looking numbers would reach exactly the wrong conclusion. Waterdrop states its ratios pure-first and spells it out in the copy. APEC and iSpring state theirs waste-first.
The EPA has documented this as a consumer problem in its own words:
EPA received feedback indicating that the distinction between the two terms was unclear and caused unnecessary confusion to consumers.
And it goes further, explaining why the more flattering of the two industry terms is misleading:
Consumers might see that a system has a 50 percent recovery rating and purchase the system thinking it sends just one gallon of water down the drain for every gallon of water it produces. However, if the system has a storage tank, inefficiencies due to back pressure from the storage tank will make the system significantly less efficient.
The NSF committee responsible for the standard has established a task group to delete the term “recovery rating” from NSF/ANSI 58 entirely. The EPA’s own labelling template standardises the direction as waste-to-treated, written as Y:1 with the waste first. Under that convention, Waterdrop’s “3:1” would have to be printed as “0.33:1”.
Do manufacturers agree with themselves?
Frequently not, and the pattern of the disagreement is consistent: the marketing page carries the favourable end of the range.
- iSpring publishes a flat “2:1” drain ratio on its support FAQ. Its own user manual for the same systems states “2:1 to 3:1”. The website shows only the better end of the manual’s range.
- APEC states 3:1 for undercounter systems on its FAQ, while its service documentation says the waste ratio “should be 4:1 to 5:1” and warns that a lower ratio means higher TDS in the output. That is the difference between 25 percent and 17 to 20 percent recovery, from one company.
- AquaTru advertises a 75 percent efficiency rate on its FAQ, while its technology page describes 80 percent of the water being flushed as waste with 20 percent passing through. Both can be true only through a recirculation loop that reprocesses reject water, which neither page explains.
- Home Master publishes 4:1 for its standard systems and 1:1 for the permeate-pump versions, which is a 75 percent reduction in waste. Its marketing describes this as “up to 80 percent”, slightly ahead of what its own two spec figures support.
Notably, neither the APEC ROES-50 nor the iSpring RCC7 publishes any drain ratio on its product page at all. You have to go into service documentation or a manual PDF to find it.
Is a cheap filter as good as an expensive one?
For running cost, often yes. For what is actually removed, that is a different question with a different answer, and it is the one that should drive your choice.
Cost per gallon tells you what a filter costs to operate. It says nothing about what it removes. A pitcher certified only for chlorine taste and an under-sink system certified for lead, cysts and PFAS can sit at the same cents-per-gallon while doing completely different jobs.
There is also a certification detail that bears directly on running past the rated life. NSF’s testing margins differ by standard:
- NSF/ANSI 42, the aesthetic standard, tests to 100 percent of rated capacity. No margin at all.
- NSF/ANSI 53, the health standard, tests to 200 percent without a filter-change indicator, or 120 percent with one.
So a taste-and-odour pitcher run past its rating is operating outside anything that was tested. A health-certified filter has a real margin built in.
Reassuringly, independent research suggests early breakthrough is not the main failure mode. A 2022 Virginia Tech study found that point-of-use filters “almost always reduce consumer lead exposure, even when operated beyond their rated capacity.”
The bigger real-world problem turns out to be clogging. In water above 0.3 mg/L of iron, that study found half the units clogged at 38 to 68 percent of their rated capacity, and rural households needed two to four times more cartridges than the rating implied.
Which is the same lesson as ZeroWater’s TDS table, generalised: your water chemistry, not the number on the box, sets what you will actually spend.
What should you actually do?
- Test your water first. TDS and hardness at minimum. Without those numbers you cannot predict what any filter will cost you, and the ZeroWater table above shows why the spread is not small.
- Price the cartridge at the manufacturer’s own store. Retailer “list prices” inflated one cartridge in this guide by 1.9 times.
- Divide by rated gallons, not by months. A six-month interval means nothing until you know how much water your household actually runs through it.
- Buy multipacks if you will use them. The gap between single and multipack pricing is 20 to 30 percent on every product in the table above.
- For RO, ask for the efficiency percentage measured with the tank connected. Not the ratio, and not the recovery rating, which the standards body is currently removing.
- Then check the certified claim list, because none of this cost work matters if the filter does not remove what is actually in your water.
Frequently asked questions
What is the cheapest water filter per gallon?
Brita Original bought in a six-pack is lowest at about 12.5 cents per gallon, followed by Brita Elite in a two-pack at 13.8 cents and the Aquasana AQ-5200 under-sink cartridge at 14.8 cents. The three are close enough that what each one removes matters far more than the cost difference.
Is a Brita cheaper than an under-sink filter?
Per gallon, usually not. A single Brita Original cartridge costs about 20 cents a gallon. The Aquasana under-sink cartridge costs 14.8 cents. Even the $409.50 Clearly Filtered set works out to 20.5 cents, the same as that single pitcher filter, because it treats 2,000 gallons.
Why is ZeroWater so expensive to run?
Because its ion exchange resin removes all dissolved solids, which exhausts it quickly, and because its life depends on your water. ZeroWater’s own table shows 40 or more gallons at low TDS falling to 8 or fewer above 400 TDS. That is roughly 39 cents to about $2 per gallon for the identical product.
How much water does reverse osmosis waste?
The EPA states that typical point-of-use RO systems run at 10 to 20 percent efficiency, meaning roughly four to nine gallons to drain per gallon produced. At its stated 15 percent average, producing 1,000 gallons a year sends about 5,667 gallons down the drain.
Does RO wastewater cost enough to matter?
It depends entirely on your water rate. At EPA average efficiency, the wasted water costs about $18 a year in Memphis, about $98 in Boston and about $260 in San Francisco. In an expensive city that is roughly what the replacement filters cost, effectively doubling the running cost.
What does a 3:1 drain ratio mean?
It depends on the manufacturer, which is the problem. Waterdrop’s 3:1 means three pure to one drain, or 75 percent recovery. APEC’s means the reverse. The EPA has documented this confusion and standardises the direction as waste-first, so Waterdrop’s 3:1 becomes 0.33:1 under its labelling.
How do I convert recovery percentage to a waste ratio?
Waste to pure equals one minus recovery, divided by recovery. So 50 percent recovery is 1:1, 30 percent is 2.33:1, 20 percent is 4:1, and 15 percent is 5.67:1. Asking for the efficiency percentage avoids the direction ambiguity in ratios entirely.
Can I use a filter past its rated capacity?
It depends on the certification. NSF/ANSI 53 health claims are tested to 200 percent of rated capacity without a change indicator, or 120 percent with one, so there is margin. NSF/ANSI 42 aesthetic claims are tested only to 100 percent, meaning a taste-only filter run past its rating is outside anything tested.
Do water filters fail early in real use?
A 2022 Virginia Tech study found point-of-use filters almost always reduce lead exposure even beyond rated capacity, so early breakthrough is not the main issue. Clogging is. Above 0.3 mg/L iron, half the units tested clogged at 38 to 68 percent of capacity and rural households needed two to four times more cartridges.
Why do retailer prices differ from manufacturer prices?
Retailers often display an inflated “list price” against which a discount is shown. We found the Aquasana AQ-C2-R listed at $139.98 by major retailers while Aquasana’s own store sold it for $73.99. Using the retailer figure would have nearly doubled the calculated cost per gallon.
Is it cheaper to buy bottled water by the gallon?
Almost never. Even the most expensive option in this guide, ZeroWater in very hard water at roughly $2 a gallon, is comparable to bottled water, and every other filter here is between 12 and 24 cents. Tap water itself costs between about $0.003 and $0.046 a gallon depending on the city.
Should I buy water filters in multipacks?
If you will use them before they expire, yes. Every product checked here was 20 to 30 percent cheaper per cartridge in a multipack. Brita Original falls from 20 cents to 12.5 cents a gallon, and Brita Elite from 16.7 to 13.8 cents.
Related reading
- NSF/ANSI 42 vs 53 vs 58 vs 401: what each certification covers
- The bottled water microplastics brand lists are made up
- How to test water quality at home
- Well water testing cost by state (2026)
Sources
- Brita replacement filter FAQ and product pages, for capacities and prices on the Original, Elite and Faucet cartridges.
- Aquasana AQ-C2-R replacement cartridge, manufacturer store price and rated capacity.
- Clearly Filtered 3-stage replacement filters, manufacturer store price and rated capacity.
- ZeroWater support, filter life by input TDS, and ZeroWater’s own comparison page.
- PUR PLUS pitcher and faucet cartridge specifications from pur.com. Pricing for the pitcher three-pack from a major retailer, as PUR does not publish prices.
- EPA WaterSense specification for point-of-use reverse osmosis systems, supporting statement, November 2024, for typical efficiency ranges, the efficiency and recovery rating definitions, and the consumer confusion finding.
- EPA WaterSense performance overview for point-of-use RO systems, January 2025, for the mandated label template and waste-to-treated ratio convention.
- APEC waste water ratio guidance and FAQ; iSpring support FAQ and the iSpring residential RO user manual; Waterdrop G3P800 and G3P600 product pages; Home Master comparison table; AquaTru FAQ and technology pages.
- Purchase, J. (2022). Point-of-use filter performance for lead reduction, Virginia Tech, on operation beyond rated capacity and clogging in high-iron water.
- NSF certification test margins for NSF/ANSI 42 and 53, from an NSF legislative affairs presentation to ANSI.
- Boston Water and Sewer Commission 2026 residential rates; MLGW 2026 utility bill comparison for Memphis; San Francisco Public Utilities Commission rate schedules for FY2026-27.
