Crude Oil API Gravity and Sulfur, Explained

API gravity is a density scale in which higher means lighter, and sulfur content is the share of the crude that is sulfur by weight. Together they are the two numbers that sort crude into light or heavy and sweet or sour. In 2025 the average barrel entering a US refinery measured 33.76 degrees and 1.30 percent sulfur, the lightest in the EIA series that starts in 1985, while the average imported barrel was 25.6 degrees and 2.48 percent.

The gap between those two averages is the trade in one comparison. Domestic wells fill refinery tanks with light, low-sulfur oil, imports bring back the heavier and sourer barrels that many plants are built to run, and EIA landed-cost data put the 2025 difference between 40-to-45-degree crude and crude at 20 degrees or lower at $9.99 a barrel, down from $18.94 in 2022.

Two numbers on a crude label

Degrees API come from a formula the American Petroleum Institute adopted so that the scale would run in the opposite direction to specific gravity. The EIA states it as degrees API = (141.5 / specific gravity at 60 degrees F) minus 131.5. Water has a specific gravity of 1.0 and so sits at 10 degrees. Anything above 10 floats on water, which is why a crude at 35 degrees spreads across a spill while an extra-heavy bitumen at 8 degrees would sink in fresh water.

A worked check shows how the scale behaves. A crude with a specific gravity of 0.8348 gives 141.5 / 0.8348 = 169.50, minus 131.5 = 38.0 degrees, which is the EIA boundary between intermediate and light. A one-degree change near 38 moves specific gravity by 0.0050, so a single degree is small in density terms but is the unit in which the market quotes quality differentials.

The last column explains why traders convert between volume and weight at all. Cargoes are often bought by the barrel and shipped by the tonne, and a tonne of 20-degree crude is only 6.74 barrels while a tonne of 45-degree crude is 7.85. The same 100,000-tonne tanker therefore carries about 111,000 more barrels of light crude than of heavy. Generic conversions between these units belong in a calculator; the CalculatorVillage collection is the place for those, and this page keeps to what the quality numbers say about crude markets.

Sulfur content is simpler. It is reported as a percentage of weight, and it matters for two reasons. Sulfur compounds must be removed before gasoline, diesel and jet fuel can meet the roadway and air-quality limits in the diesel guide, and the hydrogen-treating equipment that removes it is expensive to build and to run. A refinery running a 3 percent sulfur crude needs more hydrotreating capacity and a place to put the recovered sulfur than one running a 0.3 percent crude.

Where the cut-offs come from

No law defines light, heavy, sweet or sour for trading, but several official documents draw lines that the market uses. They disagree at the edges, which is worth knowing before comparing any two quality tables.

Two things stand out. The WTI futures window reaches down to 37 degrees, one degree below the EIA's light line, and caps sulfur at 0.42 percent, below the 0.5 percent sour line, so the contract specifies a narrow slice of the light sweet range and says nothing about heavier barrels. And the SPR buys medium-gravity oil in both sulfur grades, which is what the EIA describes in its April 2026 note on medium sour supply. The SPR levels guide follows the reserve's 2026 releases week by week.

Forty years of the average US barrel

The EIA publishes the volume-weighted API gravity and sulfur content of all crude oil input to US refineries. That series is the best single measure of what the refining system actually processes, as opposed to what wells produce or ports receive. It tells a two-act story: the barrel got heavier and sourer until the mid-2000s, then reversed.

The 1985 barrel measured 32.46 degrees and only 0.91 percent sulfur. The average then fell to its heaviest in 2004, at 30.18 degrees, and its sourest in 2008, at 1.46 percent. The 33.76 degrees of 2025 is the highest in the series; the previous high was 33.21 in 2021. Since then the average has risen 3.58 degrees, with 2.30 of that coming after 2015. Sulfur has fallen less than gravity has risen: 1.30 percent in 2025 against 1.46 at the 2008 peak, a drop of 11%.

The first time series most readers meet is the monthly one, which is noisier. In the first seven months of 2026 the national average ran between 33.07 and 33.71 degrees and between 1.24 and 1.37 percent sulfur (EIA monthly series MCRAPUS2 and MCRS1US2, through July 2026). That is a swing of under a degree month to month, which is the scale at which an individual refinery's purchasing decisions can show up in a national average.

Five regions, five different barrels

A national average hides the fact that each region of the country runs a different crude. The West Coast processes the heaviest slate in the country and the Gulf Coast the lightest; in 2005 the Gulf Coast was the second-heaviest region, behind only the West Coast.

The Gulf Coast moved from 29.66 to 34.63 degrees between 2005 and 2025, a gain of 4.97 degrees, and its sulfur fell from 1.64 to 1.29 percent. That one region holds the largest share of US refining capacity, which is why the national average follows it. The West Coast gained only 1.82 degrees over the same twenty years and still ran 5.14 degrees heavier than the Gulf Coast in 2025. The region is the most isolated in the country, a point the California gasoline guide documents for fuel, and its slate is the heaviest because of it. Sulfur is highest in the Midwest, at 1.46 percent in 2025 (the East Coast is lowest at 0.95), which fits the heavy Canadian crude that enters through Midwest ports in Table 5. The PADD districts guide maps the regions.

What the country buys from abroad

The EIA's Company Level Imports files list every crude oil import by importer, port and country of origin, with the sulfur and API gravity the importer reported. Weighting those reports by volume gives the quality of each country's shipments to the United States. In 2025 the file totals 6,168 thousand barrels a day, at 25.60 degrees and 2.477 percent sulfur.

Three patterns are visible. The first is the shift in the source of the heaviest barrels. Venezuela supplied 741 thousand barrels a day in 2016 at 17.6 degrees, and 140 in 2025 at 14.1. Canada filled the space: it rose from 3,227 to 3,910 thousand barrels a day, and the weighted sulfur of its shipments rose from 2.38 to 2.91 percent, because the growth was in heavy bitumen blends. The second is the fall in medium sour Middle Eastern crude. Saudi Arabia went from 1,099 to 269 thousand barrels a day, and Kuwait from 209 to 14. The third is a set of new sources that are light or sweet: Guyana (208 thousand barrels a day at 31.3 degrees), Argentina (from 12 to 130, 40.2 degrees) and Kazakhstan at 47.1 degrees.

Canada is two crudes, not one

The Canadian average of 24.1 degrees and 2.91 percent describes almost no actual barrel. Splitting the 2025 shipments into the EIA classes shows why:

Heavy sour crude was 45.4% of Canadian volume and medium sour another 41.0%, while sweet crude of any gravity was 12.5%. This is why a refinery that buys Canadian crude is not necessarily buying heavy crude, and why the WCS-WTI differential guide treats Western Canadian Select as a specific blend and not as a country.

How the import mix changed in nine years

Grouping every import row into the six EIA-style classes puts the origin changes in one table. Medium sour and heavy sour crude together made up 78.7% of imports in 2016, 77.9% in 2020 and 80.5% in 2025.

The sourness of the import stream has crept up while its volume fell. Total imports dropped from 7,850 thousand barrels a day in 2016 to 6,168 in 2025, a 21% decline, and the weighted sulfur rose from 2.18 to 2.48 percent. Nigeria, the clearest light sweet supplier in Table 5, fell from 207 to 137 thousand barrels a day over the same period. What remained, in volume terms, was mostly medium and heavy sour crude, and the next two sections look at why domestic supply does not replace it.

Where imports land

Port of entry shows the other half of the pattern. In 2025 the Midwest ports (PADD 2) received 2,903 thousand barrels a day at 23.3 degrees and 3.14 percent sulfur, the heaviest and sourest stream in the country, while East Coast ports took 596 at 33.0 degrees and 1.18 percent. The two regions take different crude through different ports with different equipment, and neither is well described by the national average.

Domestic production is getting lighter

Why, then, does the refinery slate keep lightening while imports stay heavy? The answer is the oilfield. EIA's production statistics break Lower 48 crude and lease condensate into gravity bands, and the bands above 40 degrees account for a growing share.

Between 2015 and 2023 Lower 48 output of crude and condensate rose from 8,950 to 12,505 thousand barrels a day, and the share above 40 degrees went from 51.2% to 55.7%. The share at 30 degrees or below fell from 14.9% to 9.5%. Condensate, which is very light, accounts for part of the rise in the top bands, a point worth remembering whenever an average API figure is quoted for US production: the average is a blend of sub-30 California and Gulf of America heavy oil and shale condensate above 50 degrees, and neither end looks like the mean.

A refinery cannot always use the light crude on its own doorstep. Plants built around cokers and hydrocrackers are designed for heavy feedstocks, and running very light oil leaves part of that equipment idle while overloading the distillation columns. This is the engineering reason light US crude is exported, a flow the crude exports guide measures, while heavy barrels come in. It is also the reason the imports by country guide and this page should be read together: volumes there, quality here.

Does quality show up in the price?

The EIA reports the landed cost of imported crude by gravity band, which is the price paid including transport and fees at the US port. Reading across the bands for seven years shows that the market does pay for lightness, with a spread that moves a lot.

The gap between the 40.1-to-45-degree band and the 20-degree-and-lower band was $16.18 a barrel in 2005, $15.21 in 2012, $8.65 in 2019, $18.94 in 2022 and $9.99 in 2025. The spread does not follow the price level. In 2005 and 2019 WTI averaged $56.64 and $57.00, nearly the same, yet the gap was $16.18 in one and $8.65 in the other, and in 2022 it reached $18.94 with WTI at $94.90. What the EIA describes is a market in which medium sour grades, which are harder to refine, normally sell at a discount to light sweet ones: the Mars discount to Light Louisiana Sweet averaged $2 a barrel in 2025, and since March 2026 Mars has traded at a $1 premium because of supply disruptions in the Middle East. Quality premiums are therefore a function of what refiners can find, not a fixed schedule per degree.

The domestic series complicates the picture. First-purchase prices, the price paid at the wellhead or lease, show no consistent premium for light crude in recent years:

In 2025 the 20-degree-and-lower class averaged $65.01, above the $62.67 for the 40.1-and-higher class. The heavy class is small, mainly California and Gulf of America production sold into nearby heavy-crude refineries, and the price it receives reflects local competition more than density. A page that told readers heavy crude is always cheaper would be wrong on these data, which is one reason the import series, with its common delivery point at the US border, is the cleaner measure of the quality premium. Both series are nominal annual averages, the import values are company-reported and the production bands stop at February 2024, so the tables describe how quality has been paid for, not what the next barrel will cost; the pricing relationships between benchmarks are in the WTI-Brent spread guide.

What a refinery gets out of each barrel

The reason the market pays more for light, sweet crude is yield. A light barrel contains more of the molecules that boil off as naphtha, gasoline blendstock and diesel, and fewer of the heavy residues that fetch less than crude itself. DOE's SPR specification states the yield windows it expects, and they show the practical difference between the two reserve grades. Table I of the manual gives naphtha, distillate, gas oil and residuum yield ranges by volume.

Read as ranges, the table is more about spread than about difference: a sweet barrel can carry up to 45 percent distillate, while the sour range tops out at 31. The point for a reader pricing quality is that the SPR sweet window allows a much wider range of outcomes than the sour window, so even inside a single sulfur class two barrels can have different refined values. Assays, the lab reports that give the full distillation curve, are the document refiners actually buy on. The API and sulfur numbers are the summary on the front page.

Data and sources: EIA petroleum bulk file (PET.zip, series last updated September to October 2026) checked October 23, 2026, for refinery crude input quality (MCRAPUS2, MCRS1US2 and PADD series), landed costs, domestic first-purchase prices, Lower 48 production by gravity and WTI. EIA Company Level Imports, annual files for 2016, 2020 and 2025, checked October 23, 2026. EIA Table Definitions for Qualities of Crude Oil Input (API gravity formula and classes), checked October 23, 2026. EIA Today in Energy, April 6, 2026 (medium sour definition, Mars-LLS spread, SPR specifications), checked October 23, 2026. EIA Today in Energy, July 16, 2012 (why light, sweet crude is priced higher), checked October 23, 2026. CME Group NYMEX Rulebook Chapter 200 (light sweet crude oil futures delivery specification), checked October 23, 2026. DOE Office of Petroleum Reserves, SPR Crude Oil Assay Manual, 5th edition, August 1, 2024, checked October 23, 2026. Unit conversions, volume-weighted import averages, class assignments, gaps and shares are this page's arithmetic.