Which water filter for hiking and camping? 16 models compared

The mountain stream runs fast, the water is cold and clear, and you still do not know whether you can drink it straight from the source. Bacteria, protozoa and viruses are too small to see¹, so the look of the water settles nothing. A hiking filter deals with some of them, but not all, and shop descriptions rarely say plainly which ones.

Shopping for a filter usually means wading through dozens of listings that promise the same thing in different words. I did that legwork so you do not have to. I went through more than a hundred listings on Amazon.pl, because Amazon is one of the world’s largest online stores. I read 22 product pages in full and checked the claims of most models against the makers’ websites. Sixteen models made the cut, and you should find one that suits you among them. I have not tested any of them myself, so for every claim I say where it comes from: the product listing, the maker’s website or a test report.

Packaging is full of numbers like 99.9999%, micrometres and acronyms such as EPA or NSF. Some of them rest on a public lab report, some on the name of a standard alone, and the rest on nothing. So before you get to the models, read this short guide. It explains how a filter differs from a purifier, what pore size and “log 6” mean, what no filter will remove, and when you still need to boil the water.

The comparison covers straws, bottle filters, filter bottles, gravity bags and pumps, split into three price bands. For each model I write what the maker claims, on what basis, and what the filter will not do.

Which water filter should you choose and what should you look for?

Before you compare grams and litres per minute, answer one question first: where will your water come from? A mountain stream, a lake next to a village and your tap at home when the mains water has to be boiled before drinking are three different situations, and each one calls for something different.

• Viruses. Many hiking filters are microfilters, which stop bacteria and protozoa but do not remove all viruses². If you cannot rule viruses out, you need a purifier, or a filter plus a disinfectant³. More in the section on filters and purifiers.
• How you drink. With a straw you drink only straight from the source and cannot fill a bottle for later. A bottle filter or a filter bottle lets you carry filtered water with you. A gravity bag filters on its own while it hangs, so it helps when several people drink. A pump draws water through a hose, also from a shallow source.
• The test behind the claim. A number like 99.9999% is only as good as the test behind it. A separate column in the table shows this, and the evidence classes are explained below.
• Carbon cartridge. It can hold back some chemicals, mainly organic ones, and improves the taste of the water⁴. It does not replace a filter against germs, though, and it wears out much sooner. More in the section on carbon.
• Frost. If you go out in winter or early spring, check what the maker says about freezing. A wet filter that has frozen may be damaged, and you cannot see it from the outside⁵. More in the section on frost.
• Flow. The litres per minute on the listing apply to a clean filter. The more it clogs, the slower the water runs^{5, 2}. More in the section on backflushing.

Pay less attention to slogans like “water from any source”. No maker in this comparison claims desalination⁶, and neither a mechanical filter, nor boiling, nor disinfection removes chemical contamination⁷.

How did I choose the filters for this comparison?

I looked for popular, well-rated models in every price band. I considered filters that had an average rating of at least 4.3 from at least 200 ratings on Amazon.pl, or a badge showing purchases in the past month (as of 15.10.2026). Two models are included despite lower numbers. LifeStraw Go averages 4.2 from 741 ratings, but it is the only branded filter bottle in its price band. Platypus QuickDraw has 166 ratings, but a high average and a gravity mode.

I left out models whose only distinctive feature was a virus removal claim with no test or laboratory named. I treat the numbers about bacteria and protozoa as makers’ and sellers’ claims, not as measurements. That is why the comparison has a separate column, “Test named by the maker”, showing whether a public test report, just a protocol name or nothing stands behind the number.

Affordable filters

This price band has three straws, a straw with a hose and a bottle filter. LifeStraw Personal is the classic straw for drinking straight from a stream. The maker cites the US EPA and NSF P231 protocols for it, for bacteria and parasites⁸. LifeStraw Peak Personal has a screw-off base with a 28 mm thread, and Purewell screws onto a bottle with its hose. ENONEO is a straw without a known brand, whose seller backs its claims with no test. LifeStraw Peak Solo is a small bottle filter weighing 48.2 g. According to the maker, the filter alone does not remove viruses or chemicals⁸.

Mid-range filters

This is where filters with their own container start. Katadyn BeFree AC is a soft 1 l bottle that weighs 76 g with the filter. LifeStraw Go is a rigid bottle with a carbon insert, BAYTIZ is a straw with a soft bottle, and Platypus QuickDraw works squeezed or by gravity. Sawyer Mini is the only model whose maker publishes a microbiological test report⁹. The maker rates it for 100,000 gallons, but that is the maker’s rating, not a test result.

Upper-range filters

The top price band has three Katadyn pumps, a gravity bag, Sawyer Squeeze and GRAYL UltraPress. Katadyn Pocket has a ceramic cartridge which, according to the maker, can be cleaned in the field, and a 20-year warranty. Hiker Pro and Vario are lighter pumps, and Vario has two flow modes. LifeStraw Peak Gravity is a 3 l bag that filters water with no pumping. The maker warns that freezing can damage the filter membrane⁸. Sawyer Squeeze is the Mini’s bigger brother, with 1 l pouches. GRAYL UltraPress is the only model whose maker claims virus removal under the EPA protocol for purifiers¹⁰. The report of that test could not be found, so this is the maker’s claim.

How is a filter different from a purifier?

A filter works like a sieve with very fine holes and stops whatever is larger than them¹. Microfilters, which covers most filters with 0.1-0.2 µm pores, stop bacteria and protozoan cysts, but do not remove all viruses². Some viruses stay in the filter because they clump together or stick to its surface, but that is not enough to remove them all². Viruses are stopped only by filters with pores of about 0.02 µm and by reverse osmosis filters².

A purifier must remove, kill or inactivate all types of disease-causing organisms, viruses included¹. Few devices meet that, and most of those that do have a disinfectant inside¹. Put simply, a filter has pores smaller than 0.3 µm and a purifier smaller than 0.03 µm³.

If a maker writes that the filter meets the requirements for bacteria and protozoa and says nothing about viruses, it is describing a filter, not a purifier¹.

What do 0.1-0.2 µm pores stop, and what do they not?

A micrometre (µm) is a thousandth of a millimetre. Giardia cysts measure 8-12 µm, Cryptosporidium oocysts 4-6 µm, bacteria such as E. coli or Salmonella 0.2-4 µm, and viruses 0.004-0.1 µm¹. A filter with 0.1-0.2 µm pores therefore stops protozoa and bacteria, while viruses are smaller than those pores, so the filter does not remove them all^{1, 2}.

It also matters how the maker states pore size. An absolute rating is the largest hole in the filter, a nominal rating is the average hole, so some holes are larger¹. For a filter against protozoa, look for an absolute rating¹. Sawyer states 0.1 µm absolute for its filters⁵. If a maker does not say which rating it gives, do not assume it is absolute.

Pores are not everything. A cracked filter, or one with a channel worn through it, lets water pass unfiltered, whatever the packaging says².

What do log 3, log 4 and log 6 mean?

Filter performance is given as a logarithmic reduction, “log” for short. It is the base-10 logarithm of the number of germs before the filter divided by their number after it¹¹. Each extra “log” means ten times fewer germs in the filtered water.

• log 3 is 99.9%, so one in every thousand gets through,
• log 4 is 99.99%, so one in ten thousand gets through,
• log 6 is 99.9999%, so one in a million gets through.

The gap between 99.9% and 99.9999% looks cosmetic, yet it means a thousand times fewer germs. That is why the number of nines matters, but only when you know which test produced it. In the lab, very large numbers of germs are added to the water so that such large reductions can be measured⁹.

What do the EPA, NSF P231 and NSF/ANSI 53 protocols say?

The protocol cited most often is the “Guide Standard and Protocol for Testing Microbiological Water Purifiers” from 1987, written by the US Environmental Protection Agency (EPA). According to the Sawyer filter test report, which cites this protocol, it requires a 6 log reduction of bacteria and a 3 log reduction of protozoa⁹. NSF P231 is a protocol for purifiers from NSF International. The data sheet of a home unit tested under both protocols lists 6 log for bacteria and 4 log for viruses as the requirements¹². The World Health Organization lists both protocols as internationally recognised testing standards and calls them rigorous and detailed¹¹.

Both protocols describe testing a purifier, a device that also deals with viruses¹. A filter whose maker cites them only for bacteria and protozoa is still a filter¹. The name of a protocol on the box does not tell you who ran the test or what the result was, either. That is why the table separates a public report from a protocol name alone.

NSF/ANSI standards for water treatment devices come in two groups: one for health and one for the look and taste of water¹. NSF/ANSI 53 covers contaminants that affect health, such as lead, some pesticides and protozoan cysts^{12, 13}. NSF/ANSI 42 covers the look and taste of water, such as the taste and smell of chlorine or fine particles, so it says nothing about germs^{12, 13}.

What does a carbon cartridge do, and what does it not?

Activation increases the surface of the carbon, and some chemicals dissolved in water get trapped on that surface⁴. It works well on organic compounds, chlorine and hydrogen sulphide, the rotten-egg smell⁴. Some carbon filters also reduce lead or pesticides¹.

Carbon does not kill germs and is not designed to remove them². Nor does it remove nitrates, iron or dissolved minerals^{13, 4}.

The carbon’s ability to bind chemicals runs out over time, and the cartridge has to be replaced⁴. Regular replacement also prevents bacteria from building up in it⁴. That is why makers give carbon its own lifespan, much shorter than for the rest of the filter. In the LifeStraw Go bottle the maker rates the membrane at 4000 l and the carbon cartridge at 100 l. In the Katadyn Vario pump the ceramic is rated at about 2000 l and the carbon at about 400 l⁶.

When do you still need to boil or disinfect water after filtering?

An ordinary mechanical filter does not protect against viruses unless the maker plainly claims to remove them². Nor does it remove chemical contamination⁷. If the water may contain viruses, boil or disinfect it after filtering³.

The surest way to deal with germs is boiling³. It works on bacteria, viruses and protozoa, even in cloudy water and at high altitude¹⁴. If the water may contain viruses and you cannot boil it, the next best option is a filter plus a disinfectant, or a purifier³.

A filter and chlorine complement each other, because chlorine kills most bacteria and viruses, but some protozoa, notably Cryptosporidium, resist it⁷, and those are what the filter stops². Chlorine or iodine tablets work more slowly in cold water, so you need a longer contact time or a higher dose². A UV lamp works only in clear water³.

Advice differs on how long to boil. The shortest says that bringing the water to a rolling boil is enough¹⁴. Other advice says 1 minute, or 3 minutes above roughly 1600 m^{3, 1}. A Polish notice about tap water contaminated with coliform bacteria says at least 2 minutes¹⁵. If you want to follow the most cautious of these, boil water for 2 minutes, and for 3 high in the mountains. Then let it cool on its own, without adding ice^{14, 15}.

Water from flooded private wells is not fit to drink, and after a flood the well has to be disinfected and the water tested¹⁶. Do not drink untreated water even if it looks clean, and carry at least two ways to treat water in case one fails³.

What happens to a filter in frost?

Sawyer writes that a filter that has never been wet is safe from frost. If a wet filter freezes, though, there is no way to tell whether it is damaged, so the maker advises replacing it and gives no warranty for it⁵. The damage matters because a crack in a filter lets water pass unfiltered².

The same maker advises keeping the filter in a pocket or close to your body in freezing weather so it does not freeze⁵. If your filter has a membrane or a ceramic cartridge, check what the maker’s instructions say about frost.

Why does a filter slow down, and how do you backflush it?

Every filter clogs eventually, because even a clear stream carries fine particles². Cloudy water clogs it much faster², so let it stand first for the sediment to settle, or strain it through a tightly woven cloth⁷.

Do not force the water through. The more clogged the filter, the more pressure it takes, and high pressure can push germs through the filter or damage it². Sawyer advises against attaching its filters directly to a tap, because the pressure in the plumbing can damage the fibres⁵.

Hollow-fibre membrane filters are cleaned by backflushing, which means pushing clean water through in the opposite direction. Sawyer recommends it when the flow starts to drop, before long storage and before using the filter again⁵. Other designs are cleaned as the maker’s instructions describe⁷. Flow also depends on altitude, and according to Sawyer the higher you go, the slower it gets⁵.

How many litres will it really filter?

Makers state lifespan each in their own way: in litres or in gallons, separately for the membrane and for the carbon. Sawyer gives up to 100,000 gallons for the Mini⁵. The public test report for that model describes three filters sampled after 100, 500 and 900 ml of water had passed through⁹. The report checked performance, not lifespan, so it neither confirms nor contradicts the gallon figure.

A filter whose flow does not come back after cleaning has to be replaced². When that happens depends largely on how cloudy the water you filter is². A carbon cartridge also has a set shelf life and is replaced regularly, as the maker’s instructions say¹.

For these reasons, litres from different listings are hard to compare one to one, and the table has no column for them. You will find the maker’s lifespan in each filter’s description.

How do you know whether a claim has been checked?

Every health claim from the product listings went into one of three classes.

• Class A. The maker names the test protocol, and the report is available and was read. In this comparison that is only the Sawyer Mini report⁹.
• Class B. The maker names a protocol, a lab or a report number, but the report itself could not be found. The GRAYL UltraPress virus claim is one example¹⁰.
• Class C. Just a number or a slogan, with no protocol or lab named. A claim like that cannot be checked.

In this comparison 2 claims are class A, 17 class B and 14 class C. The class does not tell you whether a filter works, only how much of what the maker says can be checked. A filter with a class C claim may work well, and a class A report covers particular units in a lab, not your filter out in the field.

How do you read the comparison table?

• Type tells you how you drink or collect water: straw, squeeze (pouch or bottle), filter bottle, press bottle, pump or gravity.
• Viruses shows what the maker says about viruses. Only GRAYL UltraPress has “maker’s claim under the EPA protocol”¹⁰. “Maker: not for viruses” applies to the Sawyer Mini and the LifeStraw Peak Solo. The Amazon.pl listing of Sawyer’s SP128 version says plainly that the filter is not suitable when you suspect viruses in the water⁶, and LifeStraw writes that the Peak Solo filter alone does not remove viruses or chemicals⁸. “No claim” means the maker does not mention viruses.
• Test named by the maker matches the evidence classes. “Public test report” is class A, “maker names a protocol” and “report cited, not public” are class B, and “test organism only” and “none named” are class C.
• Pore size is the value the maker gives. Two straws show 0.01 µm marked “seller”, because only the seller states it, with no test. Where it says “no data”, the maker does not state pore size.
• Weight comes from the Amazon.pl listings or the maker pages, as of 15.10.2026 and 19.10.2026. Price on Amazon.pl is a band as of 15.10.2026, not an amount, because it changes from day to day.

Frequently asked questions

Does a hiking filter remove viruses?

A filter with 0.1-0.2 µm pores stops bacteria and protozoa, but does not remove all viruses^{2, 3}. In this comparison only the maker of GRAYL UltraPress claims virus removal¹⁰. Sawyer says plainly that the Mini is not suitable when viruses are suspected⁶, and LifeStraw that the Peak Solo filter alone does not remove viruses⁸.

Do I still need to boil water after filtering?

Yes, if the water may contain viruses and the filter’s maker does not claim to remove them. Then boil or disinfect the water³. Neither boiling nor disinfection helps against chemical contamination⁷.

What is the difference between a filter and a purifier?

A filter stops bacteria and protozoa, while a purifier also deals with viruses^{1, 3}. There is more in the section on filters and purifiers.

Where do the ratings in the table come from?

They are Amazon.pl customers’ ratings as of 15.10.2026, not mine. The numbers about bacteria and viruses are makers’ or sellers’ claims.

And how do you treat water on the trail? If you have one of these filters, let me know in the comments how it works for you.

Sources

1. U.S. Environmental Protection Agency, Filtration Facts (2005), epa.gov
2. Backer H. D., Derlet R. W. et al., Wilderness Medical Society Clinical Practice Guidelines on Water Treatment for Wilderness, International Travel, and Austere Situations, Wilderness & Environmental Medicine 35(1 Suppl), 2024, pmc.ncbi.nlm.nih.gov
3. Alberta Health Services, Backcountry Water Treatment, myhealth.alberta.ca
4. Minnesota Department of Health, Water Treatment Using Carbon Filters: GAC Filter Information, health.state.mn.us
5. Sawyer, Sawyer Mini product page (specifications and questions about frost, backwashing, pressure and flow), read 15.10.2026, sawyer.com
6. Amazon.pl product listings, read 15.10.2026: BAYTIZ⁠ Amazon store, affiliate link, opens in a new tab, GRAYL⁠ Amazon store, affiliate link, opens in a new tab, ENONEO⁠ Amazon store, affiliate link, opens in a new tab, Purewell⁠ Amazon store, affiliate link, opens in a new tab, Sawyer SP128⁠ Amazon store, affiliate link, opens in a new tab, LifeStraw Go⁠ Amazon store, affiliate link, opens in a new tab, Katadyn Vario⁠ Amazon store, affiliate link, opens in a new tab
7. World Health Organization, Technical notes on drinking-water, sanitation and hygiene in emergencies, no. 5: Emergency treatment of drinking-water at the point of use, cdn.who.int
8. LifeStraw, product pages, read 19.10.2026: LifeStraw Personal, LifeStraw Peak Solo, LifeStraw Peak Gravity
9. Sawyer, Sawyer Mini Water Filter Microbiological Testing Report (16.12.2013), cdn.prod.website-files.com
10. GRAYL, UltraPress description, grayl.com
11. World Health Organization, Evaluating household water treatment options: health-based targets and microbiological performance specifications (2011), iris.who.int
12. California State Water Resources Control Board, Performance Data Sheet (an example home unit sheet with the NSF P231 and NSF/ANSI 42 and 53 requirements), waterboards.ca.gov
13. U.S. EPA WaterSense, Home Water Treatment Guide (2024), epa.gov
14. World Health Organization, Boil water, Technical brief WHO/FWC/WSH/15.02 (2015), iris.who.int
15. Powiatowa Stacja Sanitarno-Epidemiologiczna w Toruniu (district sanitary inspectorate, Toruń), Notice no. 10 of 17.07.2026 on conditional fitness of water for drinking (in Polish), gov.pl
16. Główny Inspektorat Sanitarny (Chief Sanitary Inspectorate of Poland), notice on access to drinking water in flooded areas (in Polish), gov.pl