Caffeine & Sleep
The 5-6 Hour Rule: When to Stop Drinking Coffee
By Nora Vale · December 4, 2025 · 7 min read
My rule used to be “coffee doesn’t affect me,” delivered with real confidence, because I fell asleep the moment I touched the mattress no matter what I’d drunk. Then I spent a month actually looking at the shape of my nights instead of the speed of my landings, and the shape told a different story. Falling asleep fast isn’t the test. Staying down in deep, unbroken sleep is the test, and the 5-6 hour rule everyone quotes is the first clue to why. This post is the full arithmetic, and the bigger caffeine chapter it comes from lives in the caffeine and sleep guide.
Key takeaways
- Caffeine has a half-life, dinner has a deadline, and both of them are set by your bedtime.
- The 5-6 hour half-life means a 4 p.m. coffee leaves a quarter dose in your blood at 2 a.m.
- The working guideline is bedtime minus eight hours, ten if your sleep is already thin.
- The cutoff doesn’t count sources. It counts milligrams, including tea, chocolate, decaf, and some painkillers.
- If your cutoff turns out to be innocent, that’s a real result, not a failed experiment.
Where the 5-6 hours comes from
Caffeine doesn’t leave your body on a schedule you choose. It decays, roughly by half every five to six hours in a typical adult, though the honest range is wide, and the Sleep Foundation puts it anywhere from two to twelve depending on the person. The half-life is the whole mechanism behind the rule: drink a coffee at 4 p.m. and half of it is still working at 9 or 10, a quarter at 2 a.m. You don’t feel a quarter dose as “caffeinated.” You feel it as sleep that doesn’t quite lock.
That’s the trick that fooled me for years. The drug hides its own receipt, because what you notice isn’t chemistry, it’s whether you fell asleep. Adenosine builds all day as your sleepiness receipt, and caffeine blocks the receptors rather than paying anything down. “caffeine doesn’t give energy, it hides the receipt.” The bill moves downstream. “The debt keeps moving downstream, and downstream is always night.”
So the 5-6 hour rule is real math but bad engineering, because it’s calibrated to “mostly gone” rather than “gone enough to sleep deeply on.” A quarter dose at 2 a.m. is exactly the size that doesn’t stop you falling asleep and does stop you staying under. Which is why the working rule needs more runway.
The better rule: bedtime minus eight, ten if you’re fragile
The guideline I settled on after one dismal experimental month: no caffeine within eight hours of bedtime, and ten if your sleep is already thin, light, or broken. For an 11 p.m. bedtime that’s a 3 p.m. line; my own landed at noon after a genuinely miserable month of adjustment, and I’d make the trade again without hesitating. The math is the math. Your bedtime coffee is being served at lunch, once you run the decay curve honestly.
If eight hours sounds impossible, taper instead of going cold turkey: halve the afternoon cup for two days, then drop it. The first stretch without it can carry a headache and a mood the color of wet cardboard, and the withdrawal generally runs two to nine days. Both can be true: the adjustment is miserable, and the sleep on the other side is worth it. I still look at the 3 p.m. espresso menu with real grief. I’d rather have the sleep.
Two honest caveats before you restructure your whole afternoon. Some people metabolize caffeine fast and genuinely aren’t affected, which you can’t know from how falling asleep feels, only from a proper experiment. And the pop-under-the-tent feeling of the afternoon coffee is real, so the swap matters: the post on how long before bed should I stop drinking coffee covers the full cutoff plan, and hidden sources of caffeine covers the cups you’re not counting.
The two-week experiment that finds your line
One boring experiment, two weeks, one change. Bedtime minus eight hours, no caffeine after the line, everything else deliberately untouched: same dinner, same lights, same bedtime. Write down the time of your last dose each day, because memory rounds. By the end of two weeks you have something no article, including this one, can tell you: whether the cutoff moved your nights, and how much.
Two results count as wins. The obvious one: second-half sleep locks in, the 2 to 3 a.m. surfacing quiets, mornings arrive less like a dugout. The less obvious one: nothing changes, which means caffeine wasn’t your thief. That’s a legitimate verdict and it saves you years of pointless cutoff guilt. Alex, from the book, blamed everything except his 4 p.m. latte for a season, and the case only closed when the latte moved to 12:30 and the 2 a.m. surfacing stopped with it. His text to his wife ran four words: “It was the latte.”
While you’re logging, count the milligrams, not the cups. Tea runs 40 to 120, cola 30 to 40, dark chocolate 25 to 70, decaf 2 to 15, some painkillers 65 and up. A normal Tuesday of mine, back when I believed I’d gone caffeine-free after lunch, totaled 50 to 100 milligrams from tea, chocolate, and decaf, all technically after the line. “The cutoff doesn’t count sources. It counts milligrams.” The experiment is only as honest as the tally.
Frequently asked questions
Does the 5-6 hour rule for coffee actually work?
It’s the half-life math, not the whole story. A typical adult clears half a dose in five to six hours, so a 4 p.m. coffee still leaves a quarter of it in your blood at 2 a.m. The rule is a floor, and people with already-thin sleep do better with eight or more hours.
When should I stop drinking coffee for good sleep?
Start with bedtime minus eight hours, and treat ten as the target if your sleep is already light or broken. The two-week experiment, changing nothing else, is what finds your personal line.
Does decaf, tea, or chocolate count against the cutoff?
Yes. The cutoff doesn’t count sources. It counts milligrams. Tea, cola, dark chocolate, decaf, and some painkillers all add to the daily total after your line without a single cup of coffee.
What if cutting late coffee doesn’t change anything?
Then caffeine wasn’t your thief, and that’s a real result. Some people’s sleep doesn’t move with the cutoff, and the experiment exists to find that out rather than to assume it.
Will I get withdrawal headaches?
Possibly, for two to nine days, along with a flat mood. Tapering the last cups over two or three days softens it. Both things are true: the adjustment can be grim, and most people who stick the two weeks keep the cutoff voluntarily.
The chapter behind this post lives in Sleep, Finally. Tonight’s version: find your bedtime, count back eight hours, and write the line somewhere you’ll see it tomorrow at 2:59.
This guide is educational, not medical advice. If sleep stays broken after the experiment, or caffeine interacts with a condition or medication, a doctor is the right door, no shame in that.
