Every conversation about MSG eventually becomes a conversation about glutamate, and every conversation about glutamate eventually needs numbers. Here they are, in two tables that measure two different things. Keeping them separate is the whole point of this page.

By Travis Leleu. Published October 24, 2026.

First, the one distinction that matters

Total glutamic acid is all the glutamate in a food, including the large share locked inside protein. Protein-bound glutamate has almost no taste; your body frees it during digestion and uses it as building material and fuel. Free glutamate is glutamate already on its own, and it is the only kind your tongue can taste. It is what makes aged cheese, cured ham, and a long-fermented sauce taste deep and savory.

Table A ranks foods by total glutamic acid, from USDA laboratory data. Table B ranks cheeses and fermented foods by free glutamate, from the Umami Information Center's published measurements. Comparing a number from Table A to a number from Table B is a category error, and we will not do it even when it would make a better headline.

Table A: Total glutamic acid, top 40 of 5,090 foods

Grams per 100 grams of food, USDA FoodData Central (Foundation and SR Legacy foods). Near-identical variants of the same food are grouped under the highest measured value.

The top of this list is protein concentrates and seed flours, which is exactly what you would expect: total glutamic acid mostly tracks protein content. For everyday calibration, a whole egg carries about 1.4 g per 100 g, canned tomato in juice about 0.45 g, and white mushrooms about 0.34 g. High total numbers tell you a food is protein-rich. They do not, by themselves, tell you how savory it tastes. For that, you need the free numbers.

Table B: Free glutamate, the kind you can taste

Milligrams per 100 grams. Free glutamate only. No Table A figure appears in this table, and no sentence here borrows a Table A number to imply taste. Values vary with sample, season, age, and brand, which is why several entries are ranges, and why the source for each value is named in the table.

Age is the headline. Ripening slowly breaks milk proteins down and sets glutamate free, so the same cheese at two ages is almost two different foods. Comté measures 539 mg per 100 g at 8 months and 1,570 at 18 months (Umami Information Center, analyses by Japan Food Research Laboratories). Parmigiano Reggiano measures 1,680 at 24 months and 2,220 at 48 months on the same analyses, while the 1,200 to 1,680 range from the older Giacometti compilation stays honest for typical retail ages. Cheddar measures 173 at 1 year and 718 at 10 years. Time is an ingredient, and these three gradients are the proof.

Cheeses

Roquefort note. The published Roquefort value above is 471, the modern direct analysis. An older compilation, Giacometti (1979), lists 1,280. That is a 2.7 times gap, the largest disagreement in the data on this page, between a modern analysis of a named product and a 1979 compilation entry of unknown sample provenance. The two figures are not averaged away. We publish 471 and footnote the older figure here so the disagreement stays visible.

Pecorinos: one study, a wide spread

Di Cagno et al. (2006, Journal of Dairy Science) measured free glutamate in nine regional Pecorino cheeses at the end of ripening. Eight are in the table below. The spread itself is the story: same family, same craft, wildly different numbers depending on the regional cheese. Values were published in mg per g and converted by multiplying by 100.

Across the 26 cheeses in these two tables, the pattern is time plus craft. Fresh Mozzarella starts at 0 and Sainte-Maure de Touraine sits at 60, while Parmigiano Reggiano across typical retail ages (1,200 to 1,680) and Comté at 18 months (1,570) sit far above the soft and young cheeses in the same tables, and Cheddar climbs from 173 to 718 with nothing but a decade of patience. That is the cheese case as the sources actually show it: not one champion towering over a thin field, but a whole shelf of makers who learned that waiting makes dinner taste deeper.

Fermented and cultured foods

Fermentation does the same work as ageing, with microbes instead of months in a cellar. Type and brand matter more than the category label: fish sauce and soy sauce each span a wide range depending on which type you mean, so the table names types instead of publishing one number for a whole category.

Miso: a sourced range, and an open gap behind it. The Umami Information Center's published rich-foods table lists miso at 200 to 700 mg of glutamate per 100 g, and that compilation figure is what this page reports. It is a table range, not a primary laboratory measurement: no primary absolute measurement for miso was located in the reachable literature in this research pass, and the Center's individual miso page could not be fetched (403). Kim et al. (1998) establish only that glutamate is 10.4 to 12.5 percent of total free amino acids in Japanese pastes, against 11.5 to 22.6 percent in Korean ones, which is a share, not an amount per 100 g. Unsourced content-farm ranges were excluded on sight because they carry no primary citations. The primary measurement for miso remains an open gap, and the same rule applies to every number above: no citation, no number.

Every entry in the fermented table is a food whose makers learned, long before anyone could name the chemistry, that letting microbes digest protein makes dinner taste better. That is worth celebrating plainly: these are not accidents or shortcuts, they are centuries of careful craft, and the savory depth people love in them is glutamate doing exactly what it has always done.

What we left out, and why. Dried bonito and shiitake are not in the glutamate tables because their umami is inosinate and guanylate, not glutamate; they belong in a synergy note, where guanylate complements glutamate, not in a glutamate count. Pecorino Sardo from Di Cagno et al. (2006) is excluded because the paper's table reports free glutamate of exactly 0 despite 31 mg per g of total free amino acids, which is not credible as a measurement of a ripened cheese and reads as an assay or reporting artifact. Dried doenjang powder figures of about 3,700 to 4,200 micrograms per g are quarantined because they were measured on dried powder, a concentrated basis that is not comparable to pastes as eaten; the as-eaten doenjang figure of 252 to 284 above is the comparable one. Jeju doenjang figures of 2.12 to 2.55 percent are excluded because the basis in the abstract, percent of paste versus percent of total free amino acids, is ambiguous.

The cultures that built cooking on glutamate

Kikunae Ikeda isolated glutamate in 1908. By then, kitchens on several continents had already spent centuries concentrating it on purpose.

Japan built the single most glutamate-dense pantry on this page: makombu kelp at 3,190 mg per 100 g (glutamate flavoring table, via Yamaguchi and Ninomiya 2000), with kombu generally at 1,200 to 3,400 (Kurihara 2015), Japanese-type fish sauce at 1,383 (Giacometti 1979), and nori at 1,378 to 1,380 (Kurihara 2015). Italy is the Western peak, by long ageing rather than by seaweed: Parmigiano Reggiano at 1,680 at 24 months and 2,220 at 48 months (Umami Information Center, Japan Food Research Laboratories analyses). Korea is the volume player: no single blockbuster, but Korean soy sauce at 1,264 (Giacometti 1979, cross-checked by Kim et al. 2020), doenjang at 252 to 284, and Korean commercial kimchi at about 178 (Foods 2021) stacking glutamate at every meal. Marmite at 1,960 in the United Kingdom and Vegemite at 1,431 in Australia are industrial yeast-extract staples, not farmhouse ferments, and they are labeled that way in the table above; they count here because they are genuine daily staples, and because they show the same discovery arriving by factory as well as by cellar.

The table below orders cultures by each culture's single most glutamate-dense staple seasoning. That is a density ranking, not a ranking of how much glutamate a whole diet carries. By that second measure, Korea's stacking at every meal and Japan's layered dashi cooking would both deserve a longer argument than one number can settle.

Different continents, different ingredients, one shared discovery: break protein down slowly, and the savory flavor concentrates itself. MSG is that discovery in a shaker. If your favorite food traditions are on this page, and most are, then the flavor you grew up loving was glutamate all along.

Sources and method

• Table A: USDA FoodData Central, full CSV release dated October 31, 2024. Glutamic acid is nutrient 1224 in the SR Legacy and Foundation Foods data; 5,090 foods carry a measured value. These are total glutamic acid figures, and any reuse of this table should keep that label attached.
• Table B, cheeses: Umami Information Center cheese page, analyses by Japan Food Research Laboratories; Giacometti (1979, in Glutamic Acid: Advances in Biochemistry and Physiology, Raven Press) as reproduced in the Academy of Nutrition and Dietetics Evidence Analysis Library Umami in Foods table; Di Cagno et al. (2006, Journal of Dairy Science 89, DOI 10.3168/jds.S0022-0302(06)72458-4), published in mg per g and converted by multiplying by 100. Drake et al. (2007, Journal of Food Science 72:S360) is cited as qualitative support only, because its abstract does not publish concentrations.
• Table B, fermented and cultured foods: Yamaguchi and Ninomiya (2000, Journal of Nutrition 130:921S-926S) as tabulated in Kurihara (2015, BioMed Research International, Table 1, DOI 10.1155/2015/189402); Giacometti (1979) as reproduced in the Academy of Nutrition and Dietetics table and in the glutamate flavoring table tracing the same sources; Jinap et al. (2010, Appetite 55:238-244, DOI 10.1016/j.appet.2010.06.007); Liu et al. (2017, Journal of Food Science and Technology, DOI 10.1007/s13197-017-2734-1); Foods (2021) commercial kimchi comparison; traditional Korean doenjang and samjang studies; Kim et al. (2020, Foods) as a cross-check on Korean soy sauce; Glutamate Info Australia seasoning table as an exact-match cross-check where noted. The Umami Information Center notes that values vary with the individual sample and the season, and the same holds for brand, age, and origin throughout Table B.
• Method: free glutamate only, throughout Table B. Conversions are stated where they were needed: Di Cagno mg per g to mg per 100 g, Liu mg per 100 mL to mg per 100 g by fish-sauce density, kimchi mg per kg to mg per 100 g, and doenjang mg percent read as mg per 100 g. Where sources disagree, the disagreement is published, not averaged away; see the Roquefort note and the Pecorino Piemontese outlier label. Figures that could not be classified as free glutamate, or whose basis was ambiguous, were excluded or quarantined as described in the left-out note above, never silently converted.

Curious which cheese matches your personality, and your glutamate level? Take the Which Cheese Are You quiz.