DozeClock › Caffeine › Cutoff grid

Caffeine cutoff time by bedtime

Updated October 15, 2026 · Sources: J Clin Sleep Med · National Academies Press · U.S. Food and Drug Administration

The rule

Finish each drink early enough that 25 mg or less is left at lights out. For an 8 fl oz cup of brewed coffee (96 mg) that is about 9 h 42 min before bed. A 2013 trial that gave 400 mg found sleep was still disturbed when the dose came 6 hours before bed, so 6 hours is the floor, not the target.

Row people look up most

1:15 pm

last brewed coffee (96 mg) for a 11 pm bedtime

Latest time to finish the drink, by bedtime

Scroll sideways on a phone. Times are the latest finish so that 25 mg or less is left at lights out.
Lights out6 h beforestudy lineEnergy shot200 mg · 2 fl ozCold brew coffee155 mg · 12 fl ozBrewed coffee96 mg · 8 fl ozEnergy drink72 mg · 8 fl ozMatcha70 mg · 8 fl ozEspresso64 mg · 1 fl ozInstant coffee62 mg · 8 fl ozBlack tea47 mg · 8 fl ozCola33 mg · 12 fl ozGreen tea28 mg · 8 fl ozDark chocolate24 mg · 1 ozDecaf coffee2 mg · 8 fl oz
9:00 pm3:00 pm6:00 am7:50 am11:15 am1:20 pm1:30 pm2:10 pm2:25 pm4:25 pm6:55 pm8:10 pmany timeany time
9:30 pm3:30 pm6:30 am8:20 am11:45 am1:50 pm2:00 pm2:40 pm2:55 pm4:55 pm7:25 pm8:40 pmany timeany time
10:00 pm4:00 pm7:00 am8:50 am12:15 pm2:20 pm2:30 pm3:10 pm3:25 pm5:25 pm7:55 pm9:10 pmany timeany time
10:30 pm4:30 pm7:30 am9:20 am12:45 pm2:50 pm3:00 pm3:40 pm3:55 pm5:55 pm8:25 pm9:40 pmany timeany time
11:00 pm5:00 pm8:00 am9:50 am1:15 pm3:20 pm3:30 pm4:10 pm4:25 pm6:25 pm8:55 pm10:10 pmany timeany time
11:30 pm5:30 pm8:30 am10:20 am1:45 pm3:50 pm4:00 pm4:40 pm4:55 pm6:55 pm9:25 pm10:40 pmany timeany time
12:00 am6:00 pm9:00 am10:50 am2:15 pm4:20 pm4:30 pm5:10 pm5:25 pm7:25 pm9:55 pm11:10 pmany timeany time
12:30 am6:30 pm9:30 am11:20 am2:45 pm4:50 pm5:00 pm5:40 pm5:55 pm7:55 pm10:25 pm11:40 pmany timeany time
1:00 am7:00 pm10:00 am11:50 am3:15 pm5:20 pm5:30 pm6:10 pm6:25 pm8:25 pm10:55 pm12:10 amany timeany time

Finish-by = lights-out time minus 5 × log₂(mg ÷ 25) hours, rounded down to 5 minutes · half-life 5 h (Institute of Medicine 2001) · 25 mg is this site's planning line, not a clinical threshold · "any time" = the serving starts at or under 25 mg · serving sizes in the column heads

01How the grid is computed

Every cell comes from one formula. Caffeine leaves the body by halves, so what remains is the dose times 0.5 to the power of hours divided by the half-life. The grid uses a 5-hour half-life, the average reported by the Institute of Medicine, and asks how long one serving takes to fall to 25 mg. That many hours come off your lights-out time, rounded down to 5 minutes. Doses come from the site's drink table, one serving each, and the serving size is printed in every column head.

Worked example: brewed coffee, 96 mg

Half-life 5 h, so 96 mg × 0.5^(hours ÷ 5) falls to 25 mg after 9 h 42 min. For a 11 pm bedtime that is 1:15 pm. The 6-hour column is the same cup finished 5:00 pm: 42 mg would be left.

The trial behind the 6-hour column

400 mgfixed caffeine dose given in the trial, against a placebo
0 · 3 · 6 hhours before habitual bedtime that the dose was taken
0.05 peach timing, including 6 h, differed from placebo at p < 0.05

Drake and colleagues (J Clin Sleep Med, 2013) tested whether the common advice to stop caffeine about 6 hours before bed has any empirical basis. They gave a fixed 400 mg dose at bedtime, 3 hours before bedtime or 6 hours before, and compared each with placebo. Sleep was measured two ways: self-report at home and a validated portable sleep monitor.

Their result: caffeine at bedtime, 3 hours before and 6 hours before each had a significant effect on sleep disturbance compared with placebo. The authors concluded that the size of the drop in total sleep time suggests caffeine taken 6 hours before bed has important disruptive effects, which supports the advice to avoid substantial caffeine for at least 6 hours. The abstract does not give the sample size or the minutes lost, so this page does not quote them.

At a 5-hour half-life the trial's 400 mg dose still leaves about 174 mg six hours later, and your 96 mg cup leaves 42 mg.

Two lines in the grid therefore answer different questions. The 6-hour column is the earliest point the trial tested, with a dose about four cups strong. The drink columns use 25 mg, a stricter line DozeClock chose so that one ordinary serving leaves little behind. No study fixes 25 mg as the level that protects sleep, and the trial does not show that a smaller dose 6 hours out does harm. Use the grid as a cautious planning tool and the 6-hour column as the evidence-backed floor.

02Where the grid is wrong for you

The grid assumes a 5-hour half-life, and people differ. The Institute of Medicine reports a range of roughly 1.5 to 9.5 hours in healthy adults. The measurements below come from small studies in saliva or plasma, and each is a group average, not a prediction for you. The third column shows what the 96 mg cup of brewed coffee would leave at lights out if you finished it 6 hours before bed, at each half-life.

WhoMeasuredLeft at lights out if a 96 mg cup ended 6 h beforeSource and limits
Typical adult (grid default)half-life about 5 h42 mgAverage for healthy adults. [IOM 2001]
Fast metabolisers, low end of the rangehalf-life 1.5 h6 mgEdge of the reported range. [IOM 2001]
Slow metabolisers, high end of the rangehalf-life 9.5 h62 mgEdge of the reported range. [IOM 2001]
Late pregnancyaverage half-life 3 h in non-pregnant women, 10.5 h in the last 4 weeks of pregnancy65 mgSaliva samples; small groups; 1982. [Knutti 1982]
Pregnancy (mixed stages)half-life 9.4 ± 0.6 h in 20 pregnant women, rising with the length of pregnancy62 mgPlasma; compared with 10 non-pregnant volunteers; 1985. [Ortweiler 1985]
Starting an oral contraceptivehalf-life 4.0 ± 1.7 h before, 9.0 ± 4.5 h after the first cycle60 mg12 women began, 5 remained at 7 months; one pill brand. [Meyer 1988]
Smokerscaffeine clearance 114 ± 40 ml/min in 8 smokers, 64 ± 20 ml/min in 15 non-smokersnot convertedHealthy volunteers; reports clearance, not half-life. [Joeres 1988]

Conversion = 96 × 0.5^(6 ÷ half-life in hours), applied to each measured half-life; a clearance figure cannot be converted without a volume of distribution

Direction matters more than the exact figure. Pregnancy and the pill in these studies lengthened the half-life, so a cup finished at the grid time leaves more behind than the table says. Smoking in the one study listed sped clearance, so it leaves less. The FDA also notes that some medications can make people more sensitive to caffeine, without a figure, and advises talking to a health care provider if you are pregnant, trying to become pregnant or breastfeeding.

If any of these fits you, move your finish-by time earlier than the grid by a few hours and watch how you sleep. A clinician who knows your medicines can give you a number that this table cannot.

03The bedtimes people look up most

If you go to bed at 9 pm

  • Brewed coffee11:15 am
  • Cold brew coffee7:50 am
  • Black tea4:25 pm
  • Cola6:55 pm

A 9 pm bedtime usually belongs to someone up before 5: early shifts, bakers, parents of small children. The coffee line lands at 11:15 am, so a single morning cup is the whole day's allowance and a lunchtime cup already misses it.

Cold brew is the trap here. A 12 fl oz serving carries 155 mg, which pushes its finish time to 7:50 am, before many people are at work. Cola, at 33 mg a can, can wait until 6:55 pm.

Back to the 9 pm row

If you go to bed at 9:30 pm

  • Brewed coffee11:45 am
  • Espresso2:40 pm
  • Green tea8:40 pm
  • Dark chocolateany time

Half past nine is the bedtime for a 5 am alarm with five full cycles. The 6-hour study line is 3:30 pm, but one brewed cup still needs to be finished by 11:45 am to be under 25 mg at lights out.

Green tea (28 mg a cup) and dark chocolate (24 mg an ounce) start close to the line, which is why an afternoon tea or a square of chocolate after dinner barely moves the numbers. A single espresso shot needs until 2:40 pm.

Back to the 9:30 pm row

If you go to bed at 10 pm

  • Brewed coffee12:15 pm
  • Espresso3:10 pm
  • Energy drink2:20 pm
  • Black tea5:25 pm
  • Cola7:55 pm

At a 10 pm bedtime a normal cup of coffee has to be finished around 12:15 pm. Cola and black tea are far looser because their doses are small: 7:55 pm and 5:25 pm.

If your cup is a 12 fl oz size, the dose is closer to 144 mg, and the finish time moves to about 9:20 am. The drink pages list the size ranges.

Back to the 10 pm row

If you go to bed at 10:30 pm

  • Brewed coffee12:45 pm
  • Matcha (1 tsp powder)3:00 pm
  • Instant coffee3:55 pm
  • Cola8:25 pm

Ten thirty suits a 6 am alarm with five cycles, one of the most common weekday plans. The brewed-coffee line falls at 12:45 pm; instant coffee, at 62 mg a cup, gets until 3:55 pm.

Matcha sits between them at 70 mg, finish by 3:00 pm. People who switch to matcha in the afternoon to "go easier" are often drinking nearly as much caffeine as an instant coffee.

Back to the 10:30 pm row

If you go to bed at 11 pm

  • Brewed coffee1:15 pm
  • Espresso4:10 pm
  • Energy drink3:20 pm
  • Black tea6:25 pm
  • Cola8:55 pm

Eleven o'clock gives brewed coffee until 1:15 pm. An energy drink at 72 mg per 8 fl oz looks lighter than a cup of coffee, but brands run from about 40 to well over 200 mg a can, so a large can can pass a cup.

A 5:00 pm coffee is the case where the 6-hour column and the drink column disagree most: it passes the trial line, yet about 42 mg is still in you at lights out. That is the answer to the common search "if I go to bed at 11 pm, when should I stop drinking coffee": 5 pm at the latest, early afternoon to be safe.

Back to the 11 pm row

If you go to bed at 11:30 pm

  • Brewed coffee1:45 pm
  • Energy shot8:30 am
  • Black tea6:55 pm
  • Decaf coffeeany time

Half past eleven is a typical bedtime for a 7 am alarm. The brewed-coffee line is 1:45 pm; a black tea can wait until 6:55 pm.

Energy shots are the outlier at 200 mg in 2 fl oz: finish by 8:30 am. Decaf coffee (2 mg a cup) starts under the line, so it shows "any time".

Back to the 11:30 pm row

If you go to bed at 12 am

  • Brewed coffee2:15 pm
  • Espresso5:10 pm
  • Energy drink4:20 pm
  • Black tea7:25 pm
  • Cola9:55 pm

Midnight bedtimes gain two hours on every drink compared with 10 pm, but the ordering stays the same. Cold brew and energy shots still need to be finished by 10:50 am and 9:00 am.

The grid only cares about lights out, so it holds whether you wake at 7 or at 9.

Back to the 12 am row

If you go to bed at 12:30 am

  • Brewed coffee2:45 pm
  • Cold brew coffee11:20 am
  • Energy drink4:50 pm
  • Cola10:25 pm

A 12:30 am bedtime usually means an evening shift or a late finish in a restaurant or bar. The coffee line moves to 2:45 pm, so a cup at the start of a 3 pm shift is fine and the one at the 7 pm break is not.

On a late shift the drink that slips through is the energy drink at 4:50 pm, or a cola during service, which can wait until 10:25 pm.

Back to the 12:30 am row

If you go to bed at 1 am

  • Brewed coffee3:15 pm
  • Espresso6:10 pm
  • Black tea8:25 pm
  • Dark chocolateany time

At 1 am the 6-hour study line is 7:00 pm and the brewed-coffee line 3:15 pm. If you are up this late because of work, the last cup belongs before the evening rush, not during it.

If you are up this late by drift rather than by shift, the cutoff grid will not fix it on its own: holding the alarm steady is what pulls bedtime back. The bedtime table shows which lights-out times fit your alarm.

Back to the 1 am row

Questions people ask about coffee before bed

How many hours before bed should you stop drinking coffee?
At least 6 hours, which is the earliest point the 2013 trial still found sleep disturbed with a 400 mg dose. For one 96 mg cup of brewed coffee the grid's stricter line is about 9 h 42 min, so for an 11 pm bedtime that means finishing by 1:15 pm.
Coffee before bed makes me sleepy. Is it still a problem?
Feeling sleepy is not the same as sleeping deeply. The 2013 timing trial recorded sleep with a portable monitor as well as with diaries, because what people notice and what a monitor records can differ. Falling asleep easily after a late coffee does not show the caffeine had no effect on the rest of the night.
Is decaf coffee OK before bed?
For caffeine, yes. An 8 fl oz cup of decaf has about 2 mg, under the 25 mg line from the start, so the grid shows "any time". A large decaf or several cups add up slowly, and decaf is not caffeine-free.
How much caffeine is left at bedtime from a 3 pm coffee?
With a 5-hour half-life, a 96 mg cup at 3 pm leaves about 36 mg at a 10 pm bedtime and about 32 mg at 11 pm. Slow metabolisers keep more, as the table above shows.

General information from population studies (NIH, CDC, NSF, AASM). It is not medical advice and does not replace a clinician who knows you.

Sources

  1. Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed — J Clin Sleep Med 9(11):1195-1200, 2013 (Drake C. et al.) — PubMed 24235903
  2. Caffeine for the Sustainment of Mental Task Performance: Formulations for Military Operations — National Academies Press, 2001 (half-life about 5 h, range 1.5-9.5 h, as stored in site data)
  3. Spilling the Beans: How Much Caffeine is Too Much? — U.S. Food and Drug Administration
  4. The effect of pregnancy on the pharmacokinetics of caffeine — Arch Toxicol Suppl 5:187-92, 1982 (Knutti R., Rothweiler H., Schlatter C.) — PubMed 6954898
  5. Determination of caffeine and metamizole elimination in pregnancy and after delivery — Biomed Biochim Acta 44:1189-99, 1985 (Ortweiler W. et al.) — PubMed 4084271
  6. Intra-individual comparison of the long term effect of Non-Ovlon on caffeine pharmacokinetics — Zentralbl Gynakol 110:1449-54, 1988 (Meyer F.P. et al.) — PubMed 2464882
  7. Influence of smoking on caffeine elimination in healthy volunteers and in patients with alcoholic liver cirrhosis — Hepatology 8:575-9, 1988 (Joeres R. et al.) — PubMed 3371873

General information from population studies (NIH, CDC, NSF, AASM). It is not medical advice and does not replace a clinician who knows you.

Photo by Haley Truong on Unsplash

Last reviewed by the DozeClock team: formulas re-run against the data files, sources re-checked. How we calculate · Corrections policy and log