Caffeine cutoff time for sleep: the half-life math, and how to find your own
Caffeine's half-life averages around 5 hours but ranges roughly 2 to 10 hours across healthy adults, so there's no single cutoff time that works for everyone. The commonly cited rule of thumb — stop 8 to 10 hours before bed — works reasonably well for an average metabolizer, but a genuinely fast metabolizer can often have a mid-afternoon coffee with little effect, while a slow metabolizer may still be carrying a meaningful dose at bedtime from lunch. The only way to know which end of that range you're actually on is to run the arithmetic once, then test it against your own deep sleep data for a couple of weeks.
The arithmetic: how much caffeine is still in your system
Half-life math is simple once you see it laid out: every half-life period, whatever caffeine is left in your body gets cut in half. The catch is that your personal half-life could be anywhere from about 2 to 10 hours, so the same clock time means a very different remaining dose depending on which kind of metabolizer you are.
Here's the percentage of caffeine still in your system by hours since your last dose, for three rough metabolizer profiles:
| Hours since intake | Fast metabolizer (~3h half-life) | Average (~5h half-life) | Slow metabolizer (~8h half-life) |
|---|---|---|---|
| 4 hours | ~40% remaining | ~57% remaining | ~71% remaining |
| 6 hours | ~25% remaining | ~44% remaining | ~59% remaining |
| 8 hours | ~16% remaining | ~33% remaining | ~50% remaining |
| 10 hours | ~10% remaining | ~25% remaining | ~42% remaining |
| 12 hours | ~6% remaining | ~19% remaining | ~35% remaining |
Turn that into an actual dose and it gets more concrete. Say you have a 200mg cup of coffee (a strong 16oz brew or a double espresso) at 2pm and go to bed at 10pm — that's 8 hours later:
| Metabolizer type | Caffeine remaining at bedtime (from 200mg at 2pm, 10pm bedtime) |
|---|---|
| Fast (~3h half-life) | Roughly 30mg |
| Average (~5h half-life) | Roughly 65mg |
| Slow (~8h half-life) | Roughly 100mg |
That's close to a 3x spread in how much is actually on board at lights-out, from the exact same cup of coffee at the exact same time. This is why "everyone should stop caffeine by 2pm" advice works fine for some readers and does nothing for others — the clock time is identical, the physiology underneath it isn't.
Why the same cup of coffee affects people so differently
Almost all of the variation comes down to one enzyme, plus a couple of modifiers:
- CYP1A2 genetics. This liver enzyme handles more than 95% of caffeine clearance, and common variants in the gene that codes for it split people into roughly "fast" and "slow" metabolizer groups — with slow metabolizers taking something like two to three times longer to clear the same dose. There's no home test that's necessary here; if coffee after 3pm reliably wrecks your sleep, or you can drink it at dinner and sleep fine, that's your genetics showing up in real life.
- Age. CYP1A2 activity tends to slow somewhat with age, which is part of why a cutoff that worked in your 20s can stop working in your 40s or 50s even with no other changes to your habits.
- Hormonal contraceptives and pregnancy measurably slow caffeine clearance — pregnancy in particular can roughly double or triple the half-life by the third trimester, which is one reason pregnancy caffeine guidance is more conservative than general adult guidance.
- Smoking speeds CYP1A2 up, which is part of why smokers often report less caffeine sensitivity — though this isn't a reason to start, given the far larger health costs.
None of this is fixed forever, either — the enzyme's activity can shift somewhat over months, which is part of why a self-test (below) needs occasional rechecking rather than a one-time answer.
What controlled sleep research actually shows
A 2025 randomized, double-blind crossover trial in the journal Sleep tested 100mg and 400mg doses of caffeine consumed 12, 8, or 4 hours before bedtime against a placebo. The 100mg dose — roughly one cup of coffee — showed no significant effect on objective or subjective sleep even taken as late as 4 hours before bed. The 400mg dose — closer to three or four cups, or a couple of strong energy drinks — was a different story: it significantly delayed sleep onset and altered sleep architecture when taken within 12 hours of bedtime, fragmented sleep further when taken within 8 hours, and measurably reduced perceived sleep quality at the 4-hour mark.
The practical read: dose matters as much as timing. A single modest cup well before bed is a low-risk combination for most people; a large dose, or caffeine stacked from multiple drinks across the afternoon, is the combination most likely to cost you sleep even hours before lights-out — and that cost tends to show up most clearly in sleep stages like deep sleep rather than in how long you're technically asleep.
Building your own cutoff: a 2-week self-test protocol
Population averages and the table above are a starting point, not your answer — the only way to find your actual cutoff is to test it against your own data instead of a stranger's average.
- Days 1-3 — baseline. Keep your normal caffeine habits, but log the time and rough dose of your last caffeine each day (a coffee ≈ 80-100mg, espresso shot ≈ 65mg, black tea ≈ 30-50mg, most energy drinks ≈ 80-160mg). Each morning, check your deep sleep minutes and sleep efficiency in sleep analysis — this is your reference point.
- Days 4-7 — pick a cutoff. Stop caffeine 8-10 hours before your usual bedtime and hold it steady for four days. Compare deep sleep and efficiency against your baseline days.
- Days 8-11 — push the boundary. If days 4-7 looked no different from baseline, nudge your cutoff two hours later and repeat. If they looked clearly better, nudge two hours earlier instead to see if it improves further or plateaus.
- Days 12-14 — confirm. Hold whichever cutoff performed best for a few more nights to make sure it wasn't noise from an unrelated bad-sleep night (alcohol, late meal, travel, stress).
Look for a pattern across the two weeks rather than reacting to any single night — one poor night after a normal cutoff is common and usually unrelated to caffeine, but a consistent dip in deep sleep on the nights you had caffeine later is a real signal worth acting on. If you're also carrying ongoing sleep debt, settle that separately first — a caffeine self-test run on top of an existing deficit is harder to read cleanly, and our guide to recovering sleep debt covers how to clear that baseline before you start.
If two weeks of honest testing still shows no clear pattern either way, that itself is useful information — it likely means caffeine timing isn't your limiting factor, and something else (light exposure, alcohol, stress, a late workout) is worth testing next using the same before/after approach against your own data.
Tracking the pattern instead of guessing from a chart
The tables above are the same starting math anyone would use, but your actual half-life, your actual bedtime, and your actual dose are specific to you — and the only way to close that gap is to look at your own trend instead of a population number. Vita's sleep analysis breaks out deep sleep, REM, and efficiency night by night from either a WHOOP band or Apple Health alone, so the self-test above works whether or not you own a dedicated recovery wearable. If you'd rather talk through what your own data is showing, Vita's AI coach can walk through your recent nights with you and help spot whether a pattern is really there.
FAQ
What time should I stop drinking coffee if I go to bed at 10 or 11pm?
For an average metabolizer, roughly noon to 2pm is the conservative cutoff for a 10pm bedtime — that's the commonly cited 8-10 hour buffer. Fast metabolizers can often push this to mid-afternoon with little effect; slow metabolizers may need to stop by late morning. There's no single correct answer without knowing which end of that range your own body falls on.
Is an 8-hour caffeine cutoff actually enough?
It's a reasonable population-level default, but caffeine's half-life ranges roughly 2 to 10 hours across healthy adults, so 8 hours leaves a meaningfully larger dose in a slow metabolizer's system than a fast one's. For most people it clears the largest, most disruptive share of a dose; for slow metabolizers it may still leave enough caffeine at bedtime to measurably affect deep sleep.
Why doesn't coffee at night keep some people awake at all?
Mostly genetics. Variants in the CYP1A2 liver enzyme, which handles more than 95% of caffeine clearance, mean some people break caffeine down two to three times faster than others. Someone with a naturally short half-life can drink coffee at 6pm and have cleared most of it by bedtime, while a slow metabolizer with the same cup still has a substantial dose on board.
Does caffeine tolerance mean it stops affecting sleep?
Tolerance dulls the subjective jittery or alert feeling, but it doesn't change how fast your liver clears caffeine or fully eliminate its effect on sleep architecture. Regular coffee drinkers can feel like caffeine "doesn't affect them" while a controlled sleep study would still show delayed sleep onset or reduced deep sleep from a late dose.
Does the type of drink — coffee, tea, or an energy drink — change the cutoff?
What matters is total milligrams of caffeine, not the beverage. A cup of drip coffee (roughly 80-100mg), an energy drink (similar or higher), and strong tea (30-50mg) just put different amounts into the same clearance math — a large energy drink close to bedtime is a bigger dose than a small tea at the same time, not a different kind of effect.
How do I find my own personal caffeine cutoff instead of guessing?
Track your last caffeine time and rough dose alongside your deep sleep and sleep efficiency for a couple of weeks, nudging the cutoff earlier or later every few days. Your own trend line — not a population average — is what tells you where your personal threshold actually sits.
This article is general health and training reference, not medical advice — see our sources & methodology. Consult a doctor for health concerns.