Korea built gochujang a heat scale, then made it optional

Most Korean gochujang tubs carry a number from 1 to 5 in a small box on the front. English-language writing about gochujang almost never mentions it, and when it does, it treats the number as a manufacturer’s whim.

It is not. There is an official formula behind it, published by a government agency, with a conversion factor given to three decimal places. I went looking for that formula and found it. Then I found that nothing requires anyone to use it, that there is no way to translate it into the unit the rest of the world uses, and that the standard cannot decide what goes into the top of its own equation.

What the English pages say, and the one thing they get right

Four of the most-read English pages on gochujang, opened and quoted directly.

MasterClass says the Scoville rating is “less than 1,000 units, while gochujang is on the even milder side,” and that “It takes about three months to ferment gochujang properly.” Neither figure carries a footnote.

The Daily Meal gives the same Scoville claim — “the average gochujang rates at less than 1,000 units on the Scoville scale” — and adds that it “is typically made during the fall and allowed to ferment from October to March.” Also unsourced.

Bon Appétit says it is “fermented over years in an earthenware pot kept outdoors.” Also unsourced.

So three pages give three fermentation times: three months, roughly six months, and years. None cites anything. That is the same failure that produces “miyeok expands ten times” — a number with no condition attached, described in detail in the miyeok article.

But the Scoville figure is different, because it happens to be right.

In 2012 a group at Chungbuk National University ran HPLC on commercial pepper products bought from a large supermarket in Cheongju, including eight jars of gochujang (J Korean Soc Food Sci Nutr 41(6):870-874). Their Table 1 gives capsaicin of 2.38 to 3.65 mg per 100 g, dihydrocapsaicin of 1.53 to 2.30, and Scoville values from 630.32 to 949.10. Every one of the eight is under 1,000.

The paper’s own footnote defines how it got there: “Scoville Heat Unit, (capsaicin×16.1 + dihydrocapsaicin×16.1)×10.” The English pages arrived at a correct number without saying where it came from, which is luck rather than sourcing — and the conditions matter: eight jars, one city, one supermarket, 2012.

The formula exists, and it is in an attachment

Korea’s heat scale for gochujang is called GHU, for Gochujang Hot taste Unit. Wikipedia reproduces the five steps correctly but prints the table without units, citing a 2016 agency document.

The primary source is Traditional Food Standard T014, issued by the National Agricultural Products Quality Management Service. It is not in the national law database. It is an HWPX file attached to a page on the agency’s own website, and the clause reads:

8.3.1 고추장의 매운맛 표시 제품의 표시면에 소비자가 알아보기 쉽도록 고추장의 매운맛을 6.4(캡사이신)에 따라 시험하여 구한 캡사이신(mg/kg)에 환산계수 0.769를 곱하여 나타낸 값(GHU, Gochujang Hot taste Unit)이 30 미만일 경우 ‘1 단계’, 30 이상 45 미만일 경우 ‘2 단계’, 45 이상 75 미만일 경우 ‘3 단계’, 75 이상 100 미만일 경우 ‘4 단계’, 100 이상일 경우 ’5 단계’로 구분하여 표시한다.

So GHU = capsaicin in mg/kg × 0.769, and the five steps are under 30, 30–45, 45–75, 75–100, and 100 or more. The standard also fixes the English wording for each step: ‘Mild Hot’, ‘Slight Hot’, ‘Medium Hot’, ‘Very Hot’, ‘Extreme Hot’.

Working backwards, the step boundaries correspond to capsaicin concentrations of 39.0, 58.5, 97.5 and 130.0 mg/kg.

The parent regulation that creates these standards (NAQS Notice 2023-13) has four articles and no numbers at all. Article 2 simply says the standards will be published on the agency’s website. Every figure that matters is in the attachment.

It is a recommendation

The clause above sits under heading 8.3, and 8.3 is titled 표시권고사항 — recommended labelling — introduced by the sentence “다음의 표시시항을 표시할 수 있다”, the following may be shown. (The typo is in the original: 표시시항 for 표시사항.)

The headings around it are written differently. 8.1 and 8.2 both end in 표시하여야 한다, must be shown. Under 8.2, a certified maker must print the percentage of chilli powder, glutinous rice and rice on the label.

So inside the one document that defines the heat scale, printing the chilli percentage is mandatory and printing the heat step is not — and this applies only to products carrying traditional-food certification in the first place. For ordinary supermarket gochujang, neither is required.

The reporting around this is confused. A 2025 trade piece states that GHU labelling became mandatory for KS-certified gochujang after 2010. I could not find a primary source for that. I read the KS H 2120 record and its revision history; the 2010 revision’s stated reason mentions harmonisation with compulsory law and gaps in usability, and says nothing about GHU. The body text of KS H 2120 is behind the standards service and I have not read it.

There is no bridge to Scoville

T014 allows a maker to print other units beside GHU:

단, GHU(GochujangHottasteUnit)외 SHU(Scoville Heat Unit), ppm 등의 단위를 병행하여 표시할 수 있으며

It permits the comparison and then declines to enable it. No conversion between GHU and SHU appears anywhere in the standard. It does not appear in the Codex standard either — I searched the full text of CXS 294-2023 and the strings “GHU” and “Scoville” occur zero times.

The closest thing to a conversion I found is an example on a Ministry of Agriculture portal: a mild green chilli at 1,500 SHU, it says, makes a gochujang of “약 150 GHU”, about 150 GHU. It does not show the arithmetic. It also contradicts itself in the same paragraph, describing the scale as running “0부터 100까지”, from 0 to 100, while its own top step is 100 or more and its own example is 150.

The standard cannot decide what it is measuring

This is the part that undermines the whole scale.

Clause 5.1, the quality table, defines the pungent component as “매운성분(mg/kg, w/w) 10.0 이상(캡사이신 및 디하이드로캡사이신의 합으로서)” — the sum of capsaicin and dihydrocapsaicin. Clause 6.4, the test method, is also titled 매운성분, pungent component, and its reagents section prepares standards of both compounds.

But clause 8.3.1 refers to that same clause as “6.4(캡사이신)” — 6.4 (capsaicin) — and multiplies 캡사이신(mg/kg), capsaicin alone, by 0.769.

Capsaicin alone and the sum of the two are different quantities, and in gochujang the difference is large. Take the 2012 measurements and run them through the formula both ways:

basis	capsaicin, mg/kg	× 0.769	step
capsaicin alone	23.8–36.5	18.3–28.1 GHU	step 1
capsaicin + dihydrocapsaicin	39.2–59.0	30.1–45.4 GHU	step 2

One reading of the standard puts those eight jars in step 1. The other puts them in step 2. The document supports both.

And neither reading reaches what the market actually prints. The same 2025 trade piece reports that among one major maker’s gochujang products, “3단계(보통 매운맛) 제품이 가장 많습니다” — step 3 is the most common. Step 3 begins at 45 GHU, which needs 58.5 mg/kg of capsaicin. The highest capsaicin figure in the 2012 paper is 36.5 mg/kg.

That comparison has to be read with its conditions: the measurements are from 2012 and eight jars, the labelling report is from 2025 and a different set of products, and the standard itself is ambiguous about the numerator. It is not proof that the labels are wrong. It is a gap that nobody appears to have closed, between the only published measurements of Korean gochujang and the steps printed on the tubs.

The one number the law does enforce, and where the market landed

The Korean Food Code says almost nothing numerical about fermented pastes. Its specifications for the whole 장류 category are four items: total nitrogen (soy sauce only), tar colourants, coliforms, and preservatives. There is no moisture limit, no salt limit, no protein limit.

There is one number, and it is in the definition rather than the specifications:

(10) 고추장 두류 또는 곡류 등을 주원료로 하여 누룩균 등을 배양한 후 고춧가루(6% 이상), 식염 등을 가하여 발효･숙성하거나 숙성 후 고춧가루(6% 이상), 식염 등을 가한 것을 말한다.

Chilli powder, 6% or more. That is the floor for anything sold as gochujang in Korea.

In 2022 a consumer group surveyed the category and a newspaper published the results. The five lowest products, by combined chilli powder and chilli seasoning, were 6.0%, 6.2%, 6.23%, 6.5% and 7.0%. The report should be read carefully — its own subheading and body text use “chilli powder content” and “chilli powder plus chilli seasoning” interchangeably, which are not the same thing. But the clustering is hard to miss. Set a legal minimum of 6% and the bottom of the market arrives at 6%.

One maker quoted in that article explained why buyers cannot check this themselves: “후면 라벨에는 고춧가루 함량은 필수 기재사항이 아니라서 별도 기재 안했다” — the chilli powder percentage is not a required label item, so they did not print it. Which is the mirror image of T014: the voluntary certification scheme demands the percentage, and the mandatory food code does not.

A current supermarket label for a major brand lists its ingredients in descending order by weight as 조청쌀엿, 정제수, 쌀, 고춧가루 — rice syrup, water, rice, chilli powder. The first ingredient is rice syrup. Chilli powder is fourth, with no percentage given.

Codex does not consider gochujang a chilli paste

Nearly every English page introduces gochujang as Korean chilli paste. The international standard adopted in 2023 explicitly excludes that category:

This standard does not apply to chili paste or chili sauce products having red pepper as the main ingredient.

Codex builds its definition around the saccharified grain instead: starch is saccharified with powdered malt or by naturally occurring microorganisms, salt is mixed in, the mixture is fermented and aged, and only then is red pepper powder added. The chilli enters a product that is already fermenting.

Its three quality factors are capsaicin “not less than 10.0 μg/mL (w/w)”, crude protein “not less than 3.0% (w/w)” and moisture “not more than 60.0% (w/w)”. Two things about that list are worth noting. The capsaicin figure is given in μg/mL, a concentration by volume, and then marked (w/w), by weight — the unit contradicts its own qualifier. And the moisture ceiling of 60.0% does not match T014’s, which is 50.0%. Two official standards for the same food, ten percentage points apart.

What I could not verify

• Whether KS-certified gochujang must print GHU. A trade report says so. I read the KS H 2120 record, its scope and its full revision history, and found nothing stating it; the 2010 revision’s reason does not mention GHU. The standard’s body text is not publicly readable and I have not seen it.
• Where 0.769 comes from. The factor is stated in T014 without derivation. A 2010 paper from the Korea Food Research Institute — the body credited with developing the scale, in the same year — describes a rapid UHPLC method for capsaicin in gochujang, but its abstract contains no GHU, no conversion factor and no survey of commercial products. The full text is paywalled, so I cannot say whether it is the source.
• Whether GHU uses capsaicin alone or the sum. The standard says both, in different clauses, as set out above. I found no agency clarification.
• How long commercial gochujang actually ferments. T014 clause 7.3.8 requires a maker to set standards for temperature, humidity and mixing frequency during fermentation and ageing, but prescribes no duration. The food code prescribes none either. I found no published measurement of fermentation time in commercial production.
• The original consumer-group report behind the 6–7% figures. I read the newspaper’s account of it; I could not open the group’s own release, so every figure in that section is secondhand.
• Whether the chilli-powder floor has always been 6%. I read the current food code. I did not establish when the figure was introduced or whether it has changed.
• Storage. I looked for a government or academic recommendation on storing opened gochujang and found none. The English pages give confident shelf lives — two years unopened, one year refrigerated after opening — with no source behind them.