ARTICLE · 11 SEPTEMBER 2026

How long should a study session be? 25, 50 or 90 minutes

There is no single right length for a sitting. There are, however, criteria. Twenty-five, fifty and ninety minutes each suit different kinds of work, and the same person needs different lengths at nine in the morning and nine at night. How to choose, how long the breaks should be, and what a day looks like.

01

The length follows the kind of work, not the subject

Assignments like "maths gets 50 minutes, languages get 25" do not hold up. Memorising formulas and wrestling with one hard problem are completely different activities within the same subject. Three things decide the length.

02

25 minutes — hard-to-start days and work made of small pieces

Twenty-five minutes is the Pomodoro length. It is not a scientific optimum but a compromise chosen because it is easy to start and not hard to sustain, and there are situations that compromise fits well.

Pomodoro 25/5 is this configuration, with a 15 minute long break after four rounds. With auto-start on, it runs itself all day.

03

50 minutes — work that needs immersion, the morning default

Fifty minutes is the length schools and universities have used for a very long time: one period, followed by a ten-minute gap. The familiarity is itself an advantage. Your body already knows the rhythm.

Focus 50/10 puts a 20 minute long break after the third session, because longer sessions need longer recovery.

04

90 minutes — exam rehearsal and deep work

Two justifications travel with the 90 minute block. One comes from sleep research: the observation that alertness rises and falls in cycles of roughly 90 to 120 minutes. Whether the night-time cycle applies to daytime work is still argued, and as evidence it is thin.

The firmer justification is exam length. Many papers run 60 to 100 minutes per section. Sitting still for 100 minutes on the day requires having sat that long beforehand. Students who only ever study in 25 minute pieces meet the second-half slump for the first time in the exam hall.

Since only one session is being timed, open the 2 hour timer and change it to 90 or 100 minutes. Deciding in advance that the last 20 minutes are for checking doubles as time-allocation practice.

05

Breaks scale with the session

SessionShort breakLong breakLong break every
25 min5 min15 min4 sessions
50 min10 min20 min3 sessions
90 min20–30 minEvery session

What you do in the break matters more than how long it is. One criterion: get away from the screen. A break spent on a phone rests neither the eyes nor the attention. Drink water, look out of a window, walk. The other failure is the five-minute break that becomes twenty once a feed is open. The break needs a visible countdown just as much as the session does.

06

A sample day

A weekday for someone a month out from an exam might look like this. Notice that the length changes with the time of day.

TimeStructureContent
09:00 – 12:0050/10 × 3Concepts and problems in the hardest subject. Immersive work in the immersive hours.
12:00 – 13:00LunchNo screen. A 30 minute timer sets the ceiling.
13:00 – 15:0025/5 × 4Memorisation and error review. Attention dips after lunch, so short and frequent.
15:00 – 15:20Long breakA walk.
15:20 – 17:0090 min + checkingTwo or three days a week: one practice paper at real length.
Evening25/5 × 2–4Light review. On bad days, stop at two.

That adds up to about six hours of sessions. Plans with more than that mostly do not survive, and a plan that fails takes the next day's plan down with it.

07

You cannot adjust what you do not record

You will not know which length fits until you have tried it for a few days, so keep a record. It does not need to be elaborate. At the end of the day write two numbers: how many sessions you did, and how many of them left you disappointed when they ended.

Mostly disappointed means lengthen; mostly relieved means shorten. A week produces your own number. It differs from anyone else's, and it shifts again as the exam approaches. If you want the raw figure for time spent actually working, run the stopwatch and pause it at every break. The gap between planned and real time is usually larger than expected.