When Do Babies Start Sleeping Through the Night: An Age-by-Age Timeline

Birth to 8 weeks: the bottleneck is the clock, and it does not exist yet

Newborns do not have a circadian rhythm. That is the whole story of the first two months, and understanding it saves an enormous amount of pointless effort.

A baby in the first weeks distributes sleep more or less evenly across the twenty-four hours, in blocks governed by hunger and by a sleep architecture that has not yet differentiated into adult-like stages. The day–night rhythm most of us take for granted is not present at birth; it establishes gradually across roughly the first two to three months as the circadian system matures and light exposure begins entraining it. Before that entrainment happens, there is no biological night for your baby to sleep through. There is just a series of feed-sleep-wake cycles that happen to be inconvenient for half of them.

Stomach capacity is the second constraint. It is small, and breast milk in particular digests quickly, so frequent feeding is not a habit or a preference — it is the mechanism by which a newborn gets enough calories. Long stretches without feeding in the first weeks are not a goal, and if your baby is not yet back to birth weight, night feeds may be specifically necessary. That is a conversation with your pediatrician, not a schedule decision.

So nothing on this page applies yet. What does apply:

The American Academy of Sleep Medicine's pediatric sleep-duration consensus deliberately makes no recommendation below four months of age — the panel judged that the evidence and the variation did not support one. Anyone giving you a firm target for a newborn is going beyond what the consensus was willing to state.

The one thing worth doing now is starting the circadian input, which costs nothing and takes no discipline. Bright light and normal household noise during the day; dim, dull, boring nights. You are not training anything. You are feeding the clock the signal it needs to set itself.

• Get outside, or at least near a bright window, in the morning — daylight is the strongest input to circadian development
• Keep daytime feeds social and lit; keep night feeds dim, quiet and transactional
• Do not attempt to stretch night feeds until your pediatrician confirms weight gain is on track
• Do not expect or chase a schedule — newborn sleep is not yet organised enough to have one
• Follow AAP safe sleep guidance from night one: back to sleep, firm flat surface, nothing else in the space

8 weeks to 4 months: a promising stretch, then the rewiring

This is the period that fools people, and it is worth knowing about in advance so it does not land as a personal failure.

Somewhere in the second and third months, many babies produce a genuinely encouraging stretch. The circadian rhythm is establishing, night sleep starts clumping together, and a baby who was waking every two hours starts producing a four, five, or six-hour block at the front of the night. Under the definition used in a great deal of published research — five uninterrupted hours — a meaningful share of babies technically qualify as sleeping through the night in this window. It feels like the corner has been turned.

Then, for a lot of families, it comes apart around three to four months.

The reason is developmental and it is permanent. Newborn sleep is organised into active and quiet states and babies enter sleep through the active state, which is part of why a young newborn can be fed to sleep, carried, transferred and stay down. Around three to four months, sleep architecture matures: distinct non-REM stages emerge, the baby starts entering sleep through a light stage rather than dropping straight into a deep one, and cycles settle into a pattern of surfacing toward near-waking every forty to sixty minutes or so.

The practical effect is that the baby now notices. They surface at the end of a cycle, register that they are somewhere other than where they fell asleep and that whatever was happening at the time has stopped, and they call for it back. Nothing has broken. The bottleneck simply moved: it is no longer the clock, and it is not yet the stomach. It is now the third thing — the ability to get back to sleep unaided.

This is the change usually labelled the four-month sleep regression, which our page on baby sleep regression ages and stages argues is a badly chosen name for what is really a one-way maturation. The acute disruption typically settles over two to four weeks. The underlying architecture does not revert.

4 to 6 months: the mechanics finally allow it

From roughly four months, all three prerequisites are at least plausibly available for the first time, and this is where consolidated night sleep becomes genuinely possible rather than accidental.

The circadian rhythm is established. Stomach capacity has grown substantially, and while many babies at this age still genuinely need one or more night feeds — and there is no single age at which they universally stop — the physical ceiling on how long a baby can go is much higher than it was. The AASM consensus now applies, recommending 12 to 16 hours of sleep per 24 hours including naps for infants aged 4 to 12 months, and a night that is starting to look like a night is compatible with that.

So the bottleneck is almost always the third factor: what happens at the end of a sleep cycle. And the useful distinction here is one researchers make between signalling and self-soothing. Every baby wakes between sleep cycles — objective measurement using video and actigraphy consistently records more awakenings than parents report, because a baby who surfaces, stirs and goes back down without a sound never registers with anyone. The babies described as "sleeping through" are usually not the ones who stopped waking. They are the ones who stopped calling.

What that means for what you do: the highest-leverage variable is how similar the room is at 2am to how it was at 7pm. A baby who fell asleep being rocked in a lit room and surfaces alone in a dark one has experienced a change and reasonably wants it explained. A baby who fell asleep in a dark, quiet, unchanging room surfaces into the same conditions and has less to object to.

This is the age at which a lot of families start considering a more formal approach, and it is a genuinely open decision with reasonable people on both sides — our page on how to sleep train a baby lays out the methods and the evidence, including the case for not doing it. Either way, the schedule work matters more than most people expect. Wake windows that have grown but not been adjusted produce an overtired baby who wakes more, not less, and our wake windows for a 4-month-old page has the current numbers.

• Keep the room dark enough that the baby cannot see across it — light is the single biggest lever at this age
• Aim to make the falling-asleep conditions ones that will still be true at 2am
• Adjust wake windows as they grow; a schedule that fit at ten weeks does not fit at twenty
• Do not eliminate night feeds on a schedule from the internet — that is a pediatrician conversation, and it varies by baby
• Expect an unsettled fortnight around any new gross motor skill, and let it pass

6 to 12 months: the window where it usually happens, and the things that interrupt it

If you had to name a single span in which most babies begin having consolidated nights, this is it. It is also the span in which the largest number of things arrive to interrupt the process, which is why it rarely feels like steady progress.

On prevalence: in the cohort studied by Pennestri and colleagues, fewer than half of infants slept eight consecutive hours at six months, while a clear majority managed six — and both figures rose by twelve months without either becoming universal. That is the realistic shape. Longer stretches become common in this window. Uninterrupted nights do not become the norm.

What interrupts it, roughly in order of how often it is the actual cause:

Schedule drift is first and it is badly underrated. Between six and twelve months a baby goes from three naps to two, wake windows lengthen substantially, and total day sleep needs to come down. A schedule that is not adjusted produces too much day sleep, a late bedtime, an overtired baby, or all three — and the symptoms look exactly like a sleep problem. If night sleep deteriorates and no new skill has appeared, look at the clock times before anything else. Our 3 to 2 nap transition page covers the change that usually lands in this window.

Motor development is second. Babies rehearse new skills in their sleep, so the week your baby learns to pull to stand is the week they will pull to stand in the crib at 1am and then be genuinely stuck. Practising the reverse of the skill on the floor during the day shortens this considerably.

Separation awareness is third. Object permanence matures through the second half of the first year, and once your baby understands that you continue to exist elsewhere, your leaving becomes a meaningful event rather than a neutral one. Protest at bedtime and on wakings commonly sharpens for a few weeks around this.

And fourth, prosaically: teething, first illnesses as maternal antibodies wane, solids and the digestive adjustment, and travel. All real, all temporary, all frequently blamed for disruption that outlasts them.

Past 12 months: what it means when it still has not happened

Some children reach a year, and eighteen months, and beyond, still waking regularly. If that is you, the most useful thing to know is that this is a documented pattern rather than a failure state.

Weinraub and colleagues followed a large cohort of infants from six to thirty-six months and, rather than finding one population curve, identified distinct trajectories: a majority whose waking declined steadily through the first year, and a substantial group — roughly a third — who continued waking regularly, declining much more slowly and persisting well into the second and third years. That group is large enough that a persistently waking eighteen-month-old is common, not exceptional.

What predicted membership in that group was largely constitutional and contextual — sex, feeding, temperament and family factors — rather than anything resembling parental error. The trajectory research is observational and cannot establish cause, but it consistently fails to find the thing parents most fear, which is that they did this.

There is also a reassuring negative finding worth carrying with you. Pennestri and colleagues looked directly at whether uninterrupted sleep was associated with infants' mental and psychomotor development and found no significant association. What they did find was an association with maternal mood — which is the honest version of this: the cost of a waking toddler falls mainly on the parents, not on the child.

That is not a reason to accept it if it is grinding you down. It is a reason to be clear about whose problem you are solving. Chronic sleep deprivation in adults is a genuine health issue, and persistent low mood after birth deserves clinical attention rather than endurance.

Two things are worth escalating rather than waiting out. First, if waking is accompanied by snoring, mouth-breathing, pauses in breathing, unusual restlessness, poor growth, or signs of pain, that warrants a pediatrician visit — sleep-disordered breathing, reflux and food-protein intolerance all present as unexplained night waking and none of them resolve with schedule work. Second, if you have been running on fragmented sleep for months, that is itself a reason to ask for help, from your clinician or from whoever else is available.

For the schedule side of the second year, the 2 to 1 nap transition is usually the dominant variable, and our homepage has the full schedule-by-age picture if you want to see where your current one sits.