How Much Sleep Does a Baby Need by Age

Twelve to sixteen hours per 24 for a baby of four to twelve months, eleven to fourteen for a one- or two-year-old, ten to thirteen for a three- to five-year-old — naps counted inside the total, not added on top. Those are the American Academy of Sleep Medicine's consensus figures, and the American Academy of Pediatrics supports them. Under four months there is no figure at all, and that is a decision rather than a gap. Notice the width before anything else: four hours separates the two ends of the youngest band, and a baby at either end is inside the same recommendation.

This is the one page on this site allowed to print published numbers rather than bands compiled from them, because for once somebody published some. What follows the table is what those numbers are, what they are not, and why two respectable bodies reading the same literature end up a couple of hours apart.

The published totals, side by side

Two columns, deliberately. Averaging them would produce a tidier number that neither body publishes, and the gap between them is the most useful thing on the page.

The width is the finding

A four-hour band at four to twelve months is not the panel hedging. It is what the evidence supports, and the AAP says on the same page as the figures that sleep needs vary from one person to another. Twelve hours and sixteen hours are both correct answers to the same question, asked about two different babies of the same age.

This is why the tidy charts elsewhere are worse than useless rather than merely imprecise. A single figure implies a target, a target implies a shortfall, and a shortfall implies you are doing something about it at two in the morning. A band implies what is actually true: a range of normal, inside which the thing to watch is your baby rather than the arithmetic.

It also means the honest way to use this page is subtraction, not addition. If your week lands inside the band, there is nothing to fix even if it looks nothing like the sample day you saw somewhere. If it lands well outside it in either direction, that is worth noticing — not because a number has been missed, but because it usually means the day has a shape problem rather than a sleep problem.

Naps are counted in, not added on

The AAP is explicit: the numbers are total sleep hours in a 24-hour period, so a child who still naps has those naps counted towards the total rather than stacked on top of a night.

That sounds like bookkeeping until a nap goes. Take an hour and a half of daytime sleep out of a toddler's day, keep bedtime where it was, and the same total now has to come out of the night — which it will not, because a night does not stretch to order. The child ends up short by roughly the length of the nap she lost, and the shortfall shows up as a fight at bedtime, an early morning, or both. This is the entire argument for bringing bedtime forward when a nap disappears, and it is why the two-to-one nap transition page insists on an earlier bedtime rather than a later one. The nap transition checker will tell you to wait as readily as it tells you to change anything, and waiting is usually the better answer.

The blank under four months, and why it is blank

The consensus statement makes no recommendation for infants younger than four months, and it says why: the range of normal variation in the duration and the patterns of sleep at that age is too wide for a figure to mean anything. That is a panel refusing to publish something it could easily have published, which is rarer and more useful than most of what gets printed about newborns.

What exists instead is description rather than target. The AAP writes that newborns sleep about 16 to 17 hours a day but may only manage one or two hours at a time, and that regular sleep cycles do not appear until about four months. The NHS puts a newborn's total at around 18 hours and adds the sentence that ought to be on the front of every newborn book: your baby will have her own pattern of waking and sleeping, and it is unlikely to be the same as other babies you know.

So anybody handing you an exact total for a six-week-old has invented it. If the day feels inside out at that age rather than merely short, the day and night confusion page is the more useful one — at that stage the problem is almost never the total, it is where the total is landing.

Where the two publishers disagree

Read the table across and the columns do not line up. The NHS puts babies of six to twelve months at around 15 hours; the AASM band for the same months runs 12 to 16. After the first birthday the NHS says around 12 to 15 hours and the consensus says 11 to 14. They overlap, they do not agree, and neither is wrong.

That disagreement is the best available evidence that the honest unit here is a band rather than a number. Two national-scale bodies, reading the same literature, arriving a couple of hours apart — after which a website prints "a 9-month-old needs 14 hours" and adds a graphic. Print both, and the reader can see the size of the uncertainty for herself.

Counting your own week, which beats both

A total for your baby is worth more than either column, and it takes about five days to get. Write down when she goes to sleep and when she wakes, naps included, and add it up — not for one day, because one day is noise, but for four or five. Most families are surprised twice: by how much a short nap they had discounted was contributing, and by how far the total moves between a Tuesday and a Saturday. A printed day sheet is enough, and one hand can fill it in.

Once you have a real total, the day becomes arithmetic rather than guesswork. Subtract the night, divide what is left across the naps she still takes, and you have the awake stretches — which is exactly how the bands on the wake windows by age page were compiled, and why they are wide. The wake window calculator does that division for you from an age, and the sleep schedule generator stacks the result into a sample day so you can see where the total is going. Neither prints a single clock time, because neither knows one.

What the total does not tell you

It says nothing about how the sleep is arranged, and arrangement is what parents are usually asking about. A baby can hit fourteen hours in five-hour pieces or in one-hour pieces. Night waking is ordinary well past the first birthday and no month attaches to the end of it, so a total inside the band and a night broken three times are perfectly compatible — the split nights and early morning waking pages deal with the arrangement rather than the amount.

It also says nothing about where the sleep happens, which matters more than any figure on this page. Every sleep, day or night, however the total looked that week: babies go down on their back, on a firm flat mattress, in a cot clear of blankets, pillows, bumpers and soft toys, in your room for at least the first six months. The safe sleep guide carries that in full, with the guidance each part of it comes from.

When the number is not the question

A total below the band for one week is a week. A baby who is not gaining weight, feeding poorly, unusually hard to rouse or unusually sleepy, who cannot be settled at all, or whose cry does not sound like her usual one is not a chart problem — and neither is a worry that will not go away, whatever the arithmetic says. That is a conversation with your pediatrician, GP or health visitor, and five days of written notes make it a shorter one. This page is general information, not medical advice, and it cannot see your baby.