How the SchoolsGPT score works

Every school we can score gets one number out of ten. This page explains exactly how it is worked out, what it can tell you, and what it can't. If you disagree with a school's score after reading this, you should be able to point at the step you disagree with — that's the point of publishing it.

The short version

We take a school's published exam results and its inspection grade, work out where each one sits against every comparable school in the country, and combine them. A school exactly in the middle of the country scores 5. Nothing here is our opinion of a school — every input is an official published figure.

What goes into it

Secondary schools:

Primary schools use key stage 2 results instead:

Progress 8 carries more weight than Attainment 8 on purpose. Attainment 8 measures what pupils achieved; Progress 8 measures how much they improved compared with pupils who started from the same place. The second is much closer to what the school itself did.

Why it's a position, not a percentage

Each measure is converted to a position among comparable schools before anything is added up. Around 64% of primary pupils reach the expected standard nationally, against a much lower share of secondary pupils reaching grade 5 — so blending raw percentages would rank primaries above secondaries for no reason other than that the two bars sit at different heights. Reading each measure as a position removes that entirely.

It also means the score is relative by construction. Half of all schools are below 5, because half of all schools are below the middle. A 4 is not a failing school; it is a school with more schools above it than below.

Ofsted, under two different systems

Since November 2025 Ofsted has published report cards that grade several areas separately and give no single overall word. Older inspections still carry the familiar Outstanding / Good / Requires improvement / Inadequate. Both are mapped onto the same scale so that schools inspected under either system can sit in one list — currently 13,157 schools on an overall grade and 1,706 on a report card, a balance that shifts every month as inspections continue.

Safeguarding is deliberately left out of that average. It is closer to a pass/fail standard than a measure of quality, and averaging it in would flatter a school that is merely compliant.

An old inspection counts for less

Ofsted's routine cycle is roughly four years. A judgement from within that window counts in full; from there its weight tapers, halving by eight years. A school inspected in 2016 may be run by different people teaching a different curriculum, and its grade should not speak as loudly as one from last term.

Ofsted is also capped: after everything is weighed up it can never account for more than 40% of any school's score. One inspection, possibly years old, should not be able to decide the whole picture on its own.

Missing measures don't count against a school

A measure a school hasn't published does not drag its score down — the remaining weights are shared out across what it does have. Below a threshold, though, there is too little left to say anything honest, and we show the reason instead of a number. Of 20,138 schools in scope, 17,535 have a score; 1,932 have some data but not enough, and 671 have published nothing we can use.

The second number: schools with a similar intake

Exam results track intake heavily. A school taking many pupils from better-off homes will usually post higher results than one that doesn't, whatever either school does. So a headline score partly describes who attends a school rather than what the school adds — and that is a real limitation of the number above, not a small print caveat.

So alongside it we show the same results read against schools with a similar share of pupils eligible for free school meals. It is not a statistical model of what a school "should" achieve; it is simply a like-for-like position. Where a school's results look weak nationally but strong among comparable schools, that gap is worth more than either number on its own — and where the gap doesn't appear, that is worth knowing too.

Which schools we don't score

• Special schools and pupil referral units. Their published measures record distance from a mainstream standard, so ranking them against mainstream schools would be meaningless.
• Nurseries and 16-plus institutions, which don't sit the measures the score is built from.
• Independent schools, which are not inspected by Ofsted on the same terms and often enter pupils for qualifications these measures don't count.
• Schools in Wales, Scotland and Northern Ireland. They are inspected by Estyn, Education Scotland and ETI, and their qualifications and measures are not comparable with England's. We cover these schools; we do not pretend one number spans all four systems.
• Schools whose results don't describe them. Where a school's pupils sit qualifications the measures don't count — the European Baccalaureate, some vocational routes — its Attainment 8 describes the measure's coverage rather than the school, and we leave it out rather than treat it as a weak result.

What the score cannot tell you

• Whether your child will be happy there. No published measure covers this, and it is usually the thing that matters most.
• What the school is like now. Exam results are a year or two old by the time they are published, and an inspection is a snapshot of a few days.
• How good the teaching is in a particular subject or year group. These are whole-school figures.
• Whether you'll get a place. That's admissions, not quality — see the individual school pages for oversubscription and distance information.

Two schools a point apart are not meaningfully different. Treat the score as a way to find schools worth visiting, and the visit as the thing that decides.

Where the numbers come from

Exam results and Progress 8 come from the Department for Education's published performance data; inspection grades from Ofsted; school details from the DfE's Get Information About Schools register; free school meal eligibility from the school census. All are official published sources under the Open Government Licence. We don't collect anything from schools directly, and we don't adjust any published figure.

If a school's underlying data looks wrong to you, it needs correcting at the source — but do tell us as well, especially if you think we've combined it wrongly.