Field note
How much sleep do I actually need, by age
Everyone has heard eight hours. Almost nobody has heard where it came from, which is a shame, because the actual guidelines are ranges, they differ by age, and two respected bodies phrase them differently. Here are the real numbers and what they do and do not entitle you to conclude about yourself.
Third-year medical student at Manipal Tata Medical College, writing about sleep research. A student, not yet a licensed clinician — so everything here is sourced, and none of it is medical advice. Why you should or should not trust me →
Published 2026-08-15 · Last reviewed 2026-08-15 by Chaitanya Banerjee
Researched against primary sources, drafted with AI assistance, then reviewed and approved by the named author. How we write · Corrections log
If you are an adult, aim for 7 hours or more on a regular basis. That is the American Academy of Sleep Medicine's position, and it is a floor rather than a target — the National Sleep Foundation's equivalent guidance gives 7 to 9 hours for ages 18 to 64 and 7 to 8 for 65 and over. Teenagers aged 13 to 18 need 8 to 10 hours, and school-age children 6 to 12 need 9 to 12. These are population recommendations expressed as ranges on purpose, because individual need genuinely varies. What they are not is a diagnosis: consistently feeling well on 7 hours 15 minutes does not mean you are sleep deprived, and consistently feeling terrible on 8 hours does not mean you slept enough.
- Adults: 7+ hours (AASM), 7–9 (NSF). The two agree more than the phrasing suggests.
- Teens 13–18: 8–10 hours. Most get far less, for structural reasons.
- Eight hours is a midpoint, not a law. It was never a biological constant.
- Quality and regularity sit alongside quantity. Hours alone do not settle it.
The numbers, by age
Two bodies publish the recommendations most often quoted in English-language guidance: the American Academy of Sleep Medicine, whose paediatric figures came out of a formal consensus process, and the National Sleep Foundation. They are close but not identical, and it is worth seeing both rather than one dressed up as the single truth.
| Age | AASM | National Sleep Foundation |
|---|---|---|
| Newborn (0–3 months) | Not specified | 14–17 hours |
| Infant (4–11 months) | 12–16 hours (4–12 mo) | 12–15 hours |
| Toddler (1–2 years) | 11–14 hours | 11–14 hours |
| Preschool (3–5 years) | 10–13 hours | 10–13 hours |
| School age (6–12 years) | 9–12 hours | 9–11 hours (6–13) |
| Teen (13–18 years) | 8–10 hours | 8–10 hours (14–17) |
| Adult (18–64) | 7 or more hours | 7–9 hours |
| Older adult (65+) | 7 or more hours | 7–8 hours |
Note the shape of the AASM adult recommendation. It is "7 hours or more", not "7 to 9". That phrasing is deliberate: the evidence for harm below 7 hours is considerably stronger than the evidence for harm above 9, so they set a floor rather than a band. Sleeping the recommended amount regularly is associated with better attention, behaviour, learning, memory, emotional regulation, and mental and physical health.
The calculator on this site uses these bands to flag results. Set the age selector and it will mark a schedule as in range, short, or long against the appropriate row.
Why they are ranges and not a number
Because the underlying quantity varies between people, and the committees knew it.
A recommendation range is a statement about a population, produced by looking at what sleep durations are associated with good outcomes across large groups. It is not a prescription fitted to you. Two adults can both be sleeping optimally at 7 hours and 9 hours respectively, and the guideline accommodates both by not pretending to distinguish them.
This has a consequence people routinely get backwards. The range is not a target to hit the middle of. If you land at 7 hours 20 minutes and feel well, you are inside the recommendation and there is nothing to fix. Chasing the midpoint because it sounds more virtuous is not evidence-based; it is numerology.
The distinction that matters
"How much sleep do I need" and "how much sleep is associated with good outcomes in populations like mine" are different questions. The guidelines answer the second. Only you, over weeks of observation, can answer the first.
Where "eight hours" actually came from
Nowhere in particular, which is the interesting part.
Eight is the round midpoint of the adult range, and it has the additional convenience of dividing a day into three tidy blocks — eight hours of work, eight of rest, eight of what you will — a slogan from nineteenth-century labour organising rather than from sleep science. It was a demand about the working day that got retrofitted into a biological claim.
This is the same failure mode as the 90-minute sleep cycle: a range gets published, the midpoint gets quoted, the uncertainty gets discarded, and a generation later the midpoint is repeated as a constant. Once you notice the pattern you see it everywhere in health advice.
None of which makes eight hours a bad amount of sleep. It sits comfortably inside every adult recommendation. It is just not a law, and treating it as one produces a specific and common harm: people who sleep 7 hours 10 minutes and feel fine concluding they have a problem, and lying awake worrying about it.
The people who genuinely need less
They exist, they are rare, and you are probably not one of them.
Familial natural short sleep is an inherited trait in which people sleep roughly 4 to 6 hours without the adverse effects that sleep restriction normally produces. It has been traced to variants in several genes: DEC2 (also called BHLHE41), ADRB1, NPSR1 and GRM1, catalogued as distinct familial short sleep types. Carriers are not disciplined; they are differently built.
Two things follow, and the second is the one that gets ignored. First, "some people function on five hours" is true. Second, it is not evidence that you are one of them. The defining feature of the trait is the absence of consequences, and the defining feature of ordinary sleep restriction is that people are demonstrably impaired while reporting that they feel fine — self-assessment is exactly the instrument that fails under sleep loss. Feeling okay on six hours is what the trait looks like and also what being wrong about it looks like.
Reliable prevalence figures for the trait are not well established, but the studies describing it work with individual families and small cohorts, which tells you something about how uncommon it is. Base rates should make you sceptical of a self-diagnosis here.
How to find your own number
The classic method is to sleep without an alarm for a stretch long enough that accumulated debt clears, then see where your natural duration settles. That means a holiday, roughly two weeks, and no early commitments — which is why almost nobody does it.
The practical version, which I think is more useful because people will actually do it:
- Fix your wake time first, not your bedtime. Wake time is usually externally imposed anyway, and anchoring it makes everything downstream measurable.
- Work backwards to a bedtime that gives 7.5 hours. The calculator does this in one step. Hold it for a fortnight, weekends included.
- Rate mornings before caffeine, out of ten, in a notes app. Two seconds a day. After coffee you are rating the coffee.
- Then adjust in 30-minute steps, not 5. Give each duration a full week. Night-to-night variation is large enough to drown any smaller signal.
- Judge on the average, never on one bad Tuesday. A single rough morning tells you about yesterday, not about your sleep need.
What you are looking for is the shortest duration at which your mornings are consistently unremarkable. Unremarkable is the goal. Feeling spectacular on waking is not a realistic target for most people, and chasing it is a good way to become anxious about sleep — which is itself one of the more reliable ways to sleep badly.
What the hours figure leaves out
Duration is the headline number and the easiest to measure, which is exactly why it gets over-weighted.
The most striking finding I came across while building this site is that consistency may matter more. A UK Biobank analysis of 60,977 participants, using over 10 million hours of accelerometer data, found that sleep regularity — the day-to-day consistency of when you sleep and wake — predicted all-cause mortality more strongly than sleep duration did. Participants in the more regular quintiles had a 20% to 48% lower risk of all-cause mortality than the least regular. The authors also found the regularity measure contained information beyond the variability in sleep duration, while the reverse was not true.
That reframes the practical advice. Eight hours on weekdays and eleven at the weekend averages out to a respectable number and is, on this evidence, not the same thing as sleeping well.
I will admit this landed with me personally before I read the paper. There was a period where I was getting roughly eight hours a night but at completely inconsistent times — late one night, early the next, repaying the debt at weekends. On paper it was enough sleep. It felt awful. A later stretch where I slept closer to six hours but went to bed and woke within about the same half-hour every day felt substantially better, despite being two hours shorter.
I want to be careful about what that does and does not license, because it is one person's uncontrolled experience and the obvious misreading is dangerous. It is not evidence that six hours is enough, and I am not recommending it — when I later measured properly my own number turned out to be seven, and eight is better still. Six hours sits below the AASM's seven-or-more, my own six-hour stretch was a symptom of a timetable rather than a plan, and the well-documented problem with sleep restriction is precisely that people feel fine while measurably impaired. What the experience did was make the regularity finding legible to me. The right conclusion is not "sleep less on a schedule". It is "if you are going to fix one thing, fixing when may buy you more than adding an hour to how long."
Sleep quality and continuity matter too, and neither shows up in a duration figure. Eight hours broken into six fragments is not eight hours in any functional sense. If your time in bed is adequate and you are still sleepy through the day, duration is not your problem and a calculator will not find it — that is a conversation to have with a doctor.
Questions people actually ask
How much sleep do adults need?
The AASM recommends 7 hours or more per night on a regular basis. The National Sleep Foundation gives 7 to 9 hours for ages 18 to 64, and 7 to 8 for 65 and over.
How much sleep do teenagers need?
8 to 10 hours per 24 hours for ages 13 to 18. Most teenagers get considerably less, largely because adolescent circadian timing shifts later while school start times do not.
Is 6 hours of sleep enough?
For almost everyone, no — it is below the 7-or-more the AASM recommends for adults. The rare exception is familial natural short sleep, an inherited trait linked to variants in genes including DEC2/BHLHE41 and ADRB1, and it is not something you can diagnose in yourself by feeling alright.
Do I really need exactly 8 hours of sleep?
No. Eight is a convenient midpoint of the adult range, not a biological requirement. If you feel consistently well at 7 hours 15 minutes, you are inside the guideline.
Can I catch up on sleep at the weekend?
Partially, and at a cost. Recovery sleep restores some function, but shifting your timing across the weekend reduces sleep regularity, and regularity is the measure that predicted mortality more strongly than duration in the UK Biobank analysis. A smaller, consistent schedule likely beats a large weekly swing.
Is more sleep always better?
Not necessarily. The AASM sets a floor of 7 hours rather than a band partly because the evidence around long sleep is harder to interpret — unusually long sleep is often a marker of an underlying condition rather than a cause of harm itself. If your need has recently increased a lot, that is worth mentioning to a doctor.
References
- American Academy of Sleep Medicine. Child Sleep Duration Health Advisory. The paediatric consensus figures used in the table above.
- AASM. Consensus statement methodology, Journal of Clinical Sleep Medicine, Vol. 12, No. 11, 2016.
- Sleep Education by the AASM. Healthy Sleep. Source of the adult 7-or-more-hours position.
- Windred DP et al. Sleep regularity and mortality: a prospective analysis in the UK Biobank. eLife, 2023. 60,977 participants; regularity a stronger predictor than duration.
- OMIM 612975. Short Sleep, Familial Natural, 1 (FNSS1). The DEC2/BHLHE41 entry, with cross-references to the ADRB1 and NPSR1 types.
- The impact of Mendelian sleep and circadian genetic variants in a population setting. On familial natural short sleep variants at population scale.
What this looked like on me
I ran the two-week method described above on myself, over six weeks in early 2026: wake time fixed, mornings rated out of ten before caffeine rather than after, duration adjusted in half-hour steps rather than five-minute ones.
My number is seven hours. Seven is where I stop noticing my sleep, which I think is the right target — unremarkable rather than transcendent. Eight is noticeably better, and I would take it whenever the timetable allows, which is not often.
That figure sits inside the AASM's seven-or-more, so I cannot offer the more interesting story of an experience that contradicted the guidance. It fit. I am saying so plainly rather than manufacturing a disagreement, because “the recommendation was about right for me” is a real result and it is what most people will find.
Reconciling this with the six-hour stretch above
Earlier on this page I describe a period of roughly six hours on a consistent schedule feeling substantially better than eight hours on a chaotic one. Both are true, and together they say something more useful than either alone: regularity is necessary but not sufficient. Once the timing stopped moving around, duration reasserted itself — and measured properly, more was better, right up to eight. The six-hour stretch was never an argument for six hours. It was an argument for a fixed schedule, and the moment I had one, I wanted the hours back.
How much weight this deserves
Statistically, almost none. This is one person, unblinded, who already knew what the literature predicted and was therefore primed to notice agreement with it. It is here because a reader is entitled to know whether the person writing has actually lived any of this — and because I would rather show you my evidence and its weakness than write “in my experience” and leave you guessing what that means. Where my experience and a well-powered study disagree, believe the study.
How this was checked
A general policy is easy to write and hard to hold anyone to, so here is what was actually done for this page — including the parts that came up short.
- The AASM and NSF figures are taken from the issuing organisations directly rather than from a secondary summary, and where the two differ both are shown rather than the one that reads more cleanly.
- The familial natural short sleep genes come from the OMIM catalogue entry for FNSS1 and its cross-references, not from press coverage of the findings.
- Windred et al. was read from the eLife paper, including the specific point that the regularity index retained mortality information beyond the variability in sleep duration while the reverse did not. That asymmetry is the part that matters and it is easy to lose in summary.
- What I could not establish: a reliable prevalence figure for familial natural short sleep. Rather than estimate one, the article says the prevalence is not well established. An invented number here would be worse than the gap.
Every article here is researched against the sources listed above, drafted with AI assistance, and then reviewed and approved by me before publication. AI supplies drafting speed, not judgement: it does not choose the topics, decide what is true, or generate a single citation. No reference on this site is one I have not opened. The full policy, including why AI is used at all, is here.
What this page does not cover
Stated plainly so you are not left hunting. Where somebody answers one of these better than I can, that is where the link goes.
- What to do if you cannot get the recommended hours. For a lot of people — shift workers, parents of small children, students on clinical rotations — the constraint is not knowledge. That is a scheduling problem and this page does not solve it.
- Sleep needs in pregnancy, acute illness or athletic training. All are genuinely different and none are covered here.
- Infants under four months. Deliberately excluded. Guidance for newborns is unusually consequential and belongs with a paediatrician. AASM child sleep duration advisory →
- Whether long sleep is harmful. Unresolved. Long sleep is often a marker of an underlying condition rather than a cause of harm, and this page does not pretend to settle it.
The other half of the question, and the one the calculator is built on:
What a sleep cycle actually is →Related reading
- What a sleep cycle actually is — the stages, and where 90 minutes came from.
- Are sleep calculators accurate? — an honest audit of the tool on this site.
- The sleep calculator — set your age band and it flags short nights against these numbers.
Turn the number into a bedtime
Set your age group, enter the time you have to be up, and the calculator flags whether the schedule you are planning actually lands in range.
Open the sleep calculator