The 25 Highest-Probiotic Foods (Strains Named, Evidence Graded)
Updated April 2026 · Sources: NIH ODS, Cochrane, PubMed, ISAPP
What makes a food probiotic? Three conditions must hold: the microorganisms must be alive at the time you consume the food, they must be present in adequate amounts (a threshold that varies by strain and condition), and there must be evidence - or at least strong biological plausibility - for a health benefit. Many commercially available "probiotic" foods fail one or more of these criteria. Heat-processed yogurts, pasteurised sauerkraut, and vinegar pickles contain no live cultures. We note these distinctions honestly throughout this guide.
Strains are listed in monospace format using standard nomenclature (genus, species, strain designation). Where a specific strain designation is not established for the food, only genus/species is listed.
Kefir
Emerging EvidenceTypical serving
1 cup (240 ml)
Approx. CFU
10-20 billion CFU
Key strains
L. kefiri, L. acidophilus, L. casei, Bifidobacterium species, Saccharomyces cerevisiae (yeast), 30+ total
The most diverse probiotic food. Fermented dairy; coconut or water kefir available for dairy-free. Strain composition varies by starter culture. Strong evidence for lactose digestion improvement.
Source: Bourrie et al., Front Microbiol 2016; PMC4854945
Yogurt (live cultures)
Strong EvidenceTypical serving
6 oz / 170 g
Approx. CFU
1-10 billion CFU
Key strains
L. delbrueckii subsp. bulgaricus + S. thermophilus (mandatory); added L. acidophilus, B. lactis common
Look for the IDFA 'Live and Active Cultures' seal. Plain, unsweetened yogurt is best for gut health. Greek yogurt has the same cultures; straining just concentrates protein. Flavoured varieties often have heat-processed cultures.
Source: EFSA Panel on Dietetic Products, 2010; Harvard Health 2023
Kimchi
Emerging EvidenceTypical serving
1/4 cup / 40 g
Approx. CFU
100 million - 1 billion CFU
Key strains
Leuconostoc mesenteroides, L. plantarum, L. kimchii, Lactococcus lactis - species shift as fermentation progresses
The dominant strain shifts over fermentation time. Young kimchi (1-3 weeks) has more Leuconostoc; older kimchi (>4 weeks) has more L. plantarum. Both have prebiotic fibre from cabbage and radish.
Source: Dimidi et al., Adv Nutr 2019; Park et al., J Microbiol 2014
Sauerkraut (unpasteurised, refrigerated)
Emerging EvidenceTypical serving
1/4 cup / 40 g
Approx. CFU
1-10 billion CFU
Key strains
L. plantarum, L. brevis, Leuconostoc mesenteroides, Pediococcus acidilactici
Must be raw, refrigerated, and unpasteurised. Shelf-stable canned sauerkraut is heated and has no live cultures. Also contains vitamin K2 and vitamin C. The brine is itself probiotic-rich.
Source: Zabat et al., Front Microbiol 2018
Miso
Limited EvidenceTypical serving
1 tbsp / 17 g
Approx. CFU
Variable (10 million - 100 million CFU)
Key strains
Aspergillus oryzae (koji mold, not a probiotic bacterium), plus various Tetragenococcus halophilus and Lactobacillus species in some types
High sodium - 1 tbsp has ~600-700 mg sodium. Health benefits are likely from the isoflavones and amino acid profile as much as the microbial content. Add to soup after removing from heat to preserve cultures.
Source: Uehara et al., J Sci Food Agric 2020
Natto
Emerging EvidenceTypical serving
1/2 cup / 85 g
Approx. CFU
Not typically measured by CFU; Bacillus subtilis natto is the primary organism
Key strains
Bacillus subtilis var. natto - produces nattokinase enzyme and vitamin K2 (MK-7)
Fermented soybean with a strong, sticky, ammonia-like flavour. High in nattokinase (thrombolytic enzyme) and vitamin K2 MK-7. Strong evidence for K2 contribution to bone health. Probiotic classification of B. subtilis natto is debated.
Source: Sumi et al., Experientia 1987; EFSA Journal 2009
Tempeh
Limited EvidenceTypical serving
3 oz / 85 g
Approx. CFU
Variable; Rhizopus oligosporus is the fermentation mold
Key strains
Rhizopus oligosporus (mold, not typically classified as probiotic bacterium); some preparations contain Bacillus subtilis
Fermented soybean cake, high in protein (16g per 3 oz). Better evidence base as a protein source and isoflavone source than as a probiotic per se. Not raw - typically cooked before eating, which may reduce live organisms.
Source: Nout & Kiers, J Appl Microbiol 2005
Traditional brine pickles (non-vinegar)
Limited EvidenceTypical serving
1 medium / 65 g
Approx. CFU
1-100 million CFU
Key strains
L. plantarum, L. brevis, Pediococcus species
Only naturally fermented pickles in salt brine qualify. Vinegar pickles (most supermarket brands) are acidified, not fermented, and contain no live cultures. Kosher-style dill pickles are a common brine-fermented option.
Source: Perez-Diaz et al., J Ferment Bioeng 2020
Aged hard cheese (Gouda, Swiss, Gruyere)
Limited EvidenceTypical serving
1 oz / 28 g
Approx. CFU
100 million - 1 billion CFU
Key strains
L. casei, L. rhamnosus, Bifidobacterium species survive in aged cheeses; S. thermophilus in Swiss-type
Soft aged cheeses (brie, camembert) and aged hard cheeses preserve more cultures than processed cheese. High fat content may protect bacteria through the GI tract. Evidence for health benefit specifically from cheese-derived probiotics is limited.
Source: Milesi et al., Front Microbiol 2020
Cottage cheese (with live cultures)
Limited EvidenceTypical serving
1/2 cup / 113 g
Approx. CFU
Varies by brand; some add 1-5 billion CFU
Key strains
L. acidophilus, Bifidobacterium species, Lactococcus lactis
Not all cottage cheese contains live cultures - check label. Good Cultures and similar brands specifically add probiotics post-fermentation. Mild flavour makes it easy to consume daily.
Source: NIH ODS Probiotics Fact Sheet 2023
Cultured buttermilk (traditional)
Limited EvidenceTypical serving
1 cup / 240 ml
Approx. CFU
1-10 million CFU
Key strains
L. acidophilus, Lactococcus lactis subsp. lactis, Leuconostoc cremoris
Traditional fermented buttermilk only - the liquid left after churning cultured cream. Commercial 'buttermilk' is usually acidified milk with no live cultures. Relatively low CFU count compared to kefir.
Source: FAO/WHO Probiotic Guidelines 2006
Skyr
Limited EvidenceTypical serving
5.3 oz / 150 g
Approx. CFU
1-10 billion CFU
Key strains
L. delbrueckii subsp. bulgaricus, S. thermophilus; traditional Icelandic skyr also contains Lactococcus strains
Icelandic cultured dairy product, similar to Greek yogurt but technically closer to fresh cheese. High protein (~17g per serving), low fat, live cultures. A nutritious daily food regardless of probiotic classification.
Source: Sigurdardottir et al., Microorganisms 2022
Lassi / Ayran / Doogh
Limited EvidenceTypical serving
1 cup / 240 ml
Approx. CFU
100 million - 1 billion CFU
Key strains
L. acidophilus, S. thermophilus, and the cultures of the yogurt used to make them
Yogurt-based drinks popular in South Asian (lassi), Turkish (ayran), and Persian (doogh) cuisines. Live culture content depends on the yogurt base. Easy probiotic delivery format; watch salt content in savoury versions.
Source: Yildiz, J Dairy Sci 2010
Labneh
Limited EvidenceTypical serving
2 tbsp / 30 g
Approx. CFU
100 million - 1 billion CFU
Key strains
Inherited from yogurt base: L. acidophilus, S. thermophilus
Strained yogurt or yogurt cheese popular in Middle Eastern cuisine. Retains live cultures if made from quality yogurt. Tangy flavour, spreadable. Probiotic content varies by how long it is strained.
Source: Haddadin et al., World J Microbiol Biotechnol 2008
Kombucha
Limited EvidenceTypical serving
8 fl oz / 240 ml
Approx. CFU
1-100 million CFU; highly variable by brand and batch
Key strains
SCOBY (symbiotic culture of bacteria and yeast): Acetobacter, Gluconobacter, Komagataeibacter species; Saccharomyces, Zygosaccharomyces yeasts
The SCOBY is primarily acetic acid bacteria and yeast - different from Lactobacillus/Bifidobacterium. Evidence for specific health benefits in humans is limited. Kombucha is acidic (helps palatability) and contains small amounts of alcohol (0.5-3%). Not suitable for pregnant women or those avoiding alcohol.
Source: Jayabalan et al., Compr Rev Food Sci Food Saf 2014
Kvass (traditional grain fermented)
Limited EvidenceTypical serving
8 fl oz / 240 ml
Approx. CFU
Variable
Key strains
L. plantarum, L. brevis, Saccharomyces cerevisiae
Eastern European fermented beverage traditionally made from rye bread. Lactic acid bacteria are present but at lower counts than dairy-based ferments. Low alcohol (0.5-1%). Beetroot kvass is a modern variant.
Source: Gawel et al., Fermentation 2021
Amazake
Limited EvidenceTypical serving
1/2 cup / 120 ml
Approx. CFU
Not typically CFU-classified
Key strains
Aspergillus oryzae (koji fermentation); some preparations are not bacterial-fermented
Japanese sweet rice drink made with koji mold. Unlike most probiotic foods, the key microorganism is a mold, not a bacterium. The probiotic classification is debated. Nutritional profile includes B vitamins and amino acids.
Source: Sawada et al., J Funct Foods 2019
Vinegar with 'mother' (apple cider)
Limited EvidenceTypical serving
1-2 tbsp / 15-30 ml
Approx. CFU
Not CFU-classified
Key strains
Acetobacter species (acetic acid bacteria in the 'mother')
The 'mother' is a cellulose matrix containing acetic acid bacteria - not classified as a traditional probiotic. Health claims for ACV are largely anecdotal or very limited evidence. Blood glucose modulation has some small-study evidence but is not replicated at scale.
Source: Johnston et al., J Acad Nutr Diet 2004
Fermented vegetables (kimchi-style, wildbrine)
Emerging EvidenceTypical serving
1/4 cup / 40 g
Approx. CFU
100 million - 1 billion CFU
Key strains
L. plantarum, Leuconostoc species
Commercial fermented vegetables (Wildbrine, Farmhouse Culture) use natural lacto-fermentation without vinegar. Similar microbial profile to kimchi. Check that they're stored refrigerated and label says 'raw, unpasteurised'.
Source: Dimidi et al., Adv Nutr 2019
Gari (fermented cassava)
Limited EvidenceTypical serving
1/4 cup / 40 g
Approx. CFU
Variable
Key strains
L. plantarum, L. brevis, Leuconostoc mesenteroides
West African staple made from fermented and roasted cassava. Lactic acid fermentation naturally occurs during processing. Probiotic viability in cooked preparations may be reduced.
Source: Kostinek et al., J Appl Microbiol 2005
Injera
Limited EvidenceTypical serving
One large round (~200 g)
Approx. CFU
Variable
Key strains
Leuconostoc mesenteroides, L. plantarum
Ethiopian/Eritrean sourdough flatbread made from teff. Unlike wheat sourdough, injera is not fully baked at high heat and retains more live cultures. Also contains prebiotic fibre from teff - making it unique as both a probiotic and prebiotic food.
Source: Girma & Timotewos, Afr J Microbiol Res 2019
Dosa (fermented rice-lentil)
Limited EvidenceTypical serving
Two medium dosas
Approx. CFU
Variable
Key strains
Leuconostoc mesenteroides, Lactococcus lactis - primary fermenters of the batter
South Indian fermented crepe. The batter is fermented overnight; cooking on a hot griddle likely reduces live cultures but postbiotic organic acids remain. The lentils contribute prebiotic fibre.
Source: Soni & Sandhu, J Appl Bacteriol 1990
Poi (Hawaiian fermented taro)
Limited EvidenceTypical serving
1/2 cup / 120 g
Approx. CFU
Variable
Key strains
L. mesenteroides, various Lactobacillus species depending on fermentation stage
Traditional Hawaiian fermented staple with live Lactobacillus bacteria. Sour poi (fermented for several days) has higher probiotic counts. Also provides prebiotic starch from taro.
Source: Huang et al., J Ethnopharmacol 2014
Water kefir
Emerging EvidenceTypical serving
1 cup / 240 ml
Approx. CFU
5-10 billion CFU
Key strains
L. casei, L. mesenteroides, Bifidobacterium species, Saccharomyces cerevisiae
Dairy-free probiotic beverage fermented with water kefir grains (SCOBY in sugar water). Good option for those who cannot tolerate dairy. Lower microbial diversity than milk kefir but still a solid probiotic delivery vehicle.
Source: Jeong et al., J Dairy Sci 2021
Coconut yogurt (live cultures)
Limited EvidenceTypical serving
5.3 oz / 150 g
Approx. CFU
1-5 billion CFU
Key strains
L. acidophilus, S. thermophilus, B. lactis (varies by brand)
Dairy-free alternative fermented from coconut cream. Brands vary widely in strain choice and CFU count. Coyo, Cocojune, and Anita's have added live cultures. Check label for specific strains - some brands use only acidifiers, not live fermentation.
Source: NIH ODS Probiotics Fact Sheet 2023
How to Maximise Probiotic Value from Food
Storage is the first variable. Refrigerated live-culture products degrade faster at room temperature. After opening, consume within 3-5 days for optimal CFU counts. Avoid heating probiotic foods above 40-45C (104-113F) - this kills most Lactobacillus and Bifidobacterium species. Add miso to soup after taking the pot off the heat. Eat kimchi cold as a condiment, not stir-fried.
Label reading matters. Yogurt labels should state the bacterial species used. The IDFA Live and Active Cultures seal means at least 100 million CFU per gram at the time of manufacture - roughly 1-10 billion CFU per standard serving. "Contains live and active cultures" without the seal is less reliable. "Made with active cultures" can mean cultures were added but then heat-killed.
If you cannot eat fermented foods reliably (travel, taste, allergy, intolerance), supplements are a reasonable alternative for specific conditions. See our L. rhamnosus GG guide (Culturelle) or our S. boulardii guide for the strains with the most consistent evidence.
A Note on Food vs Supplement CFU Counts
Clinical probiotic trials typically use doses of 1-50 billion CFU per day. Food sources often deliver 100 million to 10 billion CFU per serving - within range, but on the lower end. The more important variable is consistency: daily consumption of moderate-CFU food sources may be as effective as intermittent high-dose supplementation. The gut microbiome responds to regular inputs, not one-off high doses.
The key advantage of foods over supplements is the matrix effect: the fat, protein, and carbohydrate in dairy products protect probiotic bacteria through stomach acid better than a dry capsule alone. A meta-analysis by Govinden et al. (2019) showed significantly higher survival rates for probiotics delivered in a food matrix versus powder in stomach acid models.