Blood Alcohol Concentration (BAC) Calculator

Background.

This calculator will not tell you whether you can drive, and no honest BAC calculator can. It estimates a range of plausible blood alcohol concentrations from the number of standard drinks you report, your body weight, your sex and the time since your first drink. NHTSA puts the problem plainly in its guide The ABCs of BAC: "Because of the number of factors that affect BAC, it is very difficult to assess your own BAC or impairment... A person will likely be too impaired to drive before looking — or maybe even feeling — 'drunk.'" The safest plan is the one NHTSA recommends: never drive after drinking.

Most blood alcohol calculators print a single number to three decimal places, which quietly implies a precision that does not exist. This one does not, because the forensic standard that governs these calculations forbids it. ANSI/ASB Best Practice Recommendation 122, published by the American Academy of Forensic Sciences' Academy Standards Board in 2024, states in clause 4.1.3.2 that "due to the high variability within the population, the use of a single fixed Vd is inappropriate", and in clause 5.4.4 that the calculation "shall be performed using a range of elimination rates". A toxicologist who testified to a single BAC figure in court would be departing from their own profession's best practice. This page follows the standard: it reports the low and high corners of the published parameter envelope and shows you how far apart they are.

The arithmetic is Widmark's, first published in 1932 and still the basis of forensic alcohol calculations. A dose of ethanol distributes into the water compartment of the body, so the resulting concentration is the dose divided by that compartment's effective volume. The Academy Standards Board gives that volume as a range rather than a constant: 0.58 to 0.83 litres per kilogram for males, 0.43 to 0.73 for females, and 0.45 to 0.81 when sex is unspecified. It then gives the elimination rate as a range too — 0.010 to 0.025 grams per decilitre per hour, a band chosen because it "encompasses the majority of the population regardless of age, sex, ethnicity, and drinking experience". Taking the extreme corners of those two ranges is what produces the width you see in the result.

That width is real, and it is the most useful thing on this page. At the page defaults — a 180 lb man, four standard drinks, two hours since the first — the estimate spans 0.033% to 0.098%. Both ends are defensible from the same published parameters and the same drinking history. If you had asked whether four drinks put you above or below 0.08%, the honest answer is that the arithmetic cannot tell you, and neither can you. A published 2025 case report in Laboratory Medicine that applied the same standard to a real forensic case found the same thing: the longer the interval between drinking and the moment in question, the wider the defensible range becomes.

One number on this page is not a range. The theoretical peak is the concentration the dose would reach if every gram were absorbed and none had yet been metabolised — clause 5.3.3, equation 8 of the standard. At the defaults that is 0.118%. It is a ceiling, not a prediction, and it is the figure a toxicologist would use to test whether a stated drinking history could possibly explain a measured result.

Three scope limits matter before you read the number. First, the estimate assumes you are past the absorption phase; if you drank within the last hour or so, alcohol is still moving from your gut into your blood and your true level may be above the range shown and still climbing. Second, the model is for adults, and in the United States drinking under 21 is prohibited. Third, 0.08% is a legal per-se threshold, not a safety threshold. NHTSA's own effects table starts at 0.02%, where it reports loss of judgment and a decline in visual function and divided attention. Impairment does not begin at the limit; the limit is simply where the law begins.

And if the question behind the visit is how long until the alcohol is gone: only time does that. NIAAA states that the body metabolises alcohol "at a steady rate, regardless of how much alcohol a person drinks or of attempts to sober up with caffeine or by other means". Coffee, food, a cold shower, exercise and fresh air change how alert you feel. They do not change your blood alcohol concentration, and feeling more alert while still impaired is the more dangerous of the two states.

What is bac calculator?

Blood alcohol concentration is the mass of ethanol carried in a given volume of blood. In the United States it is reported in grams per decilitre and spoken as a percentage, so 0.08 g/dL is "point oh eight" or 0.08%. Much of Europe reports the same quantity as grams per litre — 0.8 g/L is the identical concentration — and the UK and Ireland commonly use milligrams per 100 millilitres, where the same reading is 80 mg/100 mL.

A BAC calculator does not measure anything. It infers a concentration from a dose using a pharmacokinetic model. The model here has two parts. The first is distribution: ethanol mixes into total body water, so a fixed dose reaches a higher concentration in a person with less body water. That is why sex and weight appear as inputs, and why the answer is a range — body composition varies widely among people of the same sex and weight, which is exactly what the Academy Standards Board's 95 percent volume-of-distribution ranges are quantifying. The second part is elimination: once absorption is complete, ethanol leaves the blood at a near-constant rate rather than a proportional one, because the alcohol dehydrogenase enzyme is saturated at any concentration a drinker would notice. That is why the estimate falls by a fixed number of points per hour rather than halving.

What the model cannot represent is as important as what it can. It has no term for how quickly you drank, whether you ate, what medications you take, your liver's condition, your drinking history, or genetic variation in alcohol dehydrogenase and aldehyde dehydrogenase. Those factors are not ignored; they are the reason the published ranges are as wide as they are. Adding a multiplier for any of them on top of the range would be inventing precision, not adding it.

Most importantly, BAC is not impairment. It is a proxy for impairment that legislatures adopted because it can be measured in a roadside breath test. Two people at the same concentration can perform very differently, and a driver can be arrested below 0.08% when an officer has probable cause based on driving behaviour. Treat any number on this page as a rough description of how much alcohol is in your body, never as a licence.

How to use this calculator.

1. 01Choose the sex assigned at birth. This picks the published volume-of-distribution range; select "Not specified" for the widest and most cautious band.
2. 02Enter your body weight and choose pounds or kilograms.
3. 03Enter the number of US standard drinks. One standard drink is 14 g of pure alcohol — a 12 oz 5% beer, a 5 oz 12% wine, or a 1.5 oz shot of 40% spirits. Larger pours and stronger drinks count as more than one.
4. 04Enter the hours since your FIRST drink, not your last. Elimination has been running since the first drink was absorbed.
5. 05Read the range, not either end on its own. The two numbers are the corners of the published parameter envelope, and the truth is somewhere inside.
6. 06Check the theoretical peak. It is the highest level that dose could reach with no elimination at all — a ceiling, not a forecast.
7. 07Read the NHTSA effects line and the safety notice. They apply whatever the number turned out to be.
8. 08Do not use any of it to decide whether to drive. Arrange a ride before you start drinking, not after.

The formula.

Widmark's relation says that a dose of ethanol spread through the body's water compartment produces a concentration equal to the dose divided by the effective volume of that compartment. ANSI/ASB Best Practice Recommendation 122 writes it in clause 5.3.1 as AC = D ⁄ (Vd × w), with AC in grams per litre, D the dose in grams, Vd the volume of distribution in litres per kilogram and w the body weight in kilograms. Clause 5.3.3, equation 8 gives the form used on this page, which divides by a further factor of ten to express the answer in grams per decilitre — the unit US blood alcohol readings use.

The dose comes from the drink count. NIAAA defines one US standard drink as "about 14 grams, or about 0.6 fluid ounces, of pure alcohol", found in 12 fl oz of 5% beer, 5 fl oz of 12% wine or a 1.5 fl oz shot of 40% spirits. Four standard drinks is therefore 56 g of ethanol, and that single number carries all the information the model uses about what you drank — the model genuinely does not care whether it arrived as beer or bourbon.

The volume of distribution is where the range comes from. Clause 4.1.3.2 of the standard states that a single fixed value is inappropriate and gives 95 percent ranges instead: 0.58 to 0.83 L/kg for males, 0.43 to 0.73 for females, 0.45 to 0.81 when sex is not specified. Because Vd sits in the denominator, the SMALLEST value produces the HIGHEST concentration. That is why the female bands, which run lower, produce higher estimates than the male bands at identical weight and dose: at 180 lb with four drinks after two hours, a man's estimate is 0.033% to 0.098% and a woman's is 0.044% to 0.140%.

Elimination is subtracted last. Once absorption is complete, ethanol leaves the blood at a near-constant rate — zero-order kinetics — because alcohol dehydrogenase is saturated at any concentration a drinker would notice. Clause 4.1.4.2 gives the rate as a range of 0.010 to 0.025 g/dL per hour, chosen because it "encompasses the majority of the population regardless of age, sex, ethnicity, and drinking experience". Multiplying that rate by the hours since the first drink and subtracting gives the current estimate; both bounds are floored at zero, because a negative blood alcohol concentration is not a physical quantity.

The two reported bounds are the extreme corners of that envelope. The HIGH bound pairs the narrowest volume of distribution with the slowest elimination; the LOW bound pairs the widest with the fastest. This is the same construction Wu and Benowitz used in a 2025 Laboratory Medicine case report, where computing all four corners produced an extreme range for a real forensic case. The corners diverge at 0.015 g/dL per hour — the difference between the two elimination rates — so the longer ago you started drinking, the wider the honest range becomes.

A published elimination rate that this page does NOT use is worth naming. Jones's 2010 evidence-based survey in Forensic Science International puts the full physiological range at 10 to 35 mg/100 mL per hour, i.e. 0.010 to 0.035 g/dL/h, with 10 to 15 on an empty stomach, 15 to 20 when not fasted, and 25 to 35 in alcoholics undergoing detoxification, whose microsomal CYP2E1 activity is elevated. This page uses the narrower ANSI/ASB range because that upper tail describes a clinical sub-population rather than a general audience. The wider Jones range is selectable on the alcohol elimination time calculator.

Rounding happens exactly once, at the end. Every intermediate value is carried at full precision, and the result is rounded to three decimal places — the precision at which blood alcohol is universally reported — only when it is displayed. The NHTSA effects band is looked up from that same rounded number, so the band and the printed figure can never contradict each other. The published standard's own worked examples note the opposite convention for hand calculation: "For accuracy in the text, rounding was performed at each step. Numbers may vary slightly when calculations are performed using a spreadsheet and rounding is not applied until the end."

The engine reproduces the standard's own example. Annex A.1.2 of ANSI/ASB 122 works through a 230 lb (104 kg) man who has consumed 32 g of ethanol and prints a maximum achievable range of approximately 0.037 to 0.053 g/dL. Entering 104 kg, 32 ÷ 14 = 2.286 standard drinks and zero elapsed hours returns 0.037% to 0.053% — a match to the standard's printed precision, and the test that guards this page.

A worked example.

Frequently asked questions.

References& sources.

1. [1]AAFS Academy Standards Board (2024). ANSI/ASB Best Practice Recommendation 122, 1st Ed., "Best Practice Recommendation for Performing Alcohol Calculations in Forensic Toxicology." §4.1.3.2 (Vd ranges), §4.1.4.2 (elimination rate range), §5.2.2, §5.3.3 Eq. 8, Annex A.1.2 (worked example reproduced by this calculator). Approved 21 June 2024; added to the OSAC Registry 1 October 2024.
2. [2]National Highway Traffic Safety Administration (2016). "The ABCs of BAC: A Guide to Understanding Blood Alcohol Concentration and Alcohol Impairment." DOT HS 809 844, July 2016 (Rev). Source of the BAC effects table at .02, .05, .08, .10 and .15 and of the crash-risk multiples quoted on this page.
3. [3]National Center for Statistics and Analysis (2025). "Alcohol-Impaired Driving: 2023 Data." Traffic Safety Facts, Report No. DOT HS 813 713. National Highway Traffic Safety Administration. Overview, p.2 (per-se limits; Utah 0.05; commercial 0.04; under-21 prohibition) and Important Safety Reminders, p.11.
4. [4]National Institute on Alcohol Abuse and Alcoholism (updated December 2025). "What Is A Standard Drink?" — "one standard drink contains about 14 grams, or about 0.6 fluid ounces, of pure alcohol."
5. [5]National Institute on Alcohol Abuse and Alcoholism. Core Resource on Alcohol — "The Basics: Defining How Much Alcohol Is Too Much." Source of the statement that alcohol metabolism "proceeds at a steady rate, regardless of how much alcohol a person drinks or of attempts to sober up with caffeine or by other means", and of the 0.08% binge-drinking definition.
6. [6]Jones AW (2010). "Evidence-based survey of the elimination rates of ethanol from blood with applications in forensic casework." Forensic Science International 200(1–3):1–20. doi:10.1016/j.forsciint.2010.02.021. PMID 20304569. Full text paywalled; figures cited here are from the published abstract.
7. [7]Wu AHB, Benowitz N (2025). "Theoretical and retrograde calculation of blood alcohol: wide estimates using the Academy Standards Board for calculations over an extended period of time." Laboratory Medicine 56(6):829–831. doi:10.1093/labmed/lmaf055. Open access. Independent application of ANSI/ASB 122 that corroborates the parameter ranges and the extreme-corner method used here.
8. [8]Widmark EMP (1932). Die theoretischen Grundlagen und die praktische Verwendbarkeit der gerichtlich-medizinischen Alkoholbestimmung. Berlin: Urban & Schwarzenberg. Original derivation of the relation used here. Print monograph — bibliographic reference only, no online edition.