CEM vs GL Assessment: which provider sets your child's 11+ for 2027 entry

Five of the six pages ranking for this question still tell parents to buy CEM-style practice books. Not one of the admissions authorities checked for this page now sits a CEM 11-plus. Below is the provider for each area, named on the council's or school's own page, the four different pass-mark systems in use, and the switch to Quest Assessments already contracted for the exam held in September 2027.

Is the CEM 11+ still a live option for 2027 entry?

Most families reach this question with a practice book already open, and the cover says CEM. The position for September 2027 entry is blunter than the comparison articles suggest: of the admissions authorities checked for this page, not one now sits a CEM 11-plus. Buckinghamshire, Gloucestershire and the Slough Consortium — three of the areas routinely listed as CEM territory in those articles — each name GL Assessment as their provider on their own admissions pages. Bexley names a different provider again. Reading School names a third. Essex sets its own papers and always has. The two-provider choice the whole search results page is built around has quietly stopped describing anything real.

Warwickshire County Council is the clearest primary record of how the change happened, because it announced the move itself and gave the reason: "This is the result of a joint procurement exercise with The Schools of King Edward VI Foundation, which took place when the previous contract with CEM came to an end." A contract ended and a new one was tendered. The council was also careful to say the new papers would cover the same ground: "The exams will test the same four competencies of Maths, English, Verbal and Non-Verbal Reasoning as the CEM papers have in previous years." It added that GL was "already used by other authorities, including Buckinghamshire, Kent and Lincolnshire".

Two honest caveats before the detail. First, this page only claims what an admissions authority or a member school of a consortium states on its own site. Coventry and Wiltshire were not checked, so nothing here should be read as a statement about either, and if you live in either place the only safe source is the authority's own admissions page. Second, "GL has won" is itself too neat a story. The useful question for the next two admission rounds is not CEM or GL. It is GL, or Quest Assessments, or FSCE, or a consortium that writes its own papers — and in Northern Ireland, a single assessment that GL sets and marks for every academically selective school in the country.

That distinction is practical, not pedantic. A family in Bexley and a family in Buckinghamshire are preparing for tests that differ in provider, in what is weighted, in the scale the score is reported on and in what counts as a pass. Generic eleven plus preparation that treats the country as one exam leaves both of them guessing at the thing that decides the outcome. The rest of this page names the provider area by area with the page it came from, sets four real pass-mark systems side by side, and gives the admissions numbers that one council publishes and almost nobody else does.

What happened to CEM, and what does Cambridge sell now?

"CEM was abolished" is the version you will read most often, and it is wrong in a way that matters if you are trying to judge whether the material on your kitchen table is worth anything. Cambridge still sells an entrance assessment. What ended was the paper CEM 11-plus supplied to state grammar consortia. The surviving product, Cambridge Select Insight, is described on Cambridge's own entrance assessments page as "Suitable for entry into Years 7, 8, 9 and 10", with an average duration of 60 minutes, delivered as a digital download, and covering "Numerical, Verbal and Non-verbal ability".

The detail that settles it for grammar school parents is which variant Cambridge points at grammar schools, because only one is described that way and it is not a Year 7 test: "Cambridge Select Insight: Mid-Year Transfer is designed explicitly as a late-transfer test for Grammar schools and is suitable for entry into years 8, 9 and 10." A late-transfer test is what a school uses when a place comes free part way through a key stage. It is a useful thing to know exists if you are hoping to move a child into a grammar school after Year 7, but it has nothing to do with the September assessment that decides Year 7 entry.

Cambridge is also explicit that nothing on a shop shelf carries its endorsement: "To maintain the fairness of our assessments, we do not provide any commercially available practice materials (including practice or past papers) nor endorse any other commercially available resources or tuition services." Read that alongside the advice on most comparison pages, which is to buy CEM-style practice books and drill the style. There is no published official paper for those books to be modelled on, and the organisation whose name sits on the cover does not release one. The style being practised is a publisher's reconstruction, not a specification.

None of which means you should bin a CEM book in a panic. Verbal reasoning is verbal reasoning, comprehension is comprehension, and a child who works through good material gets better at the underlying skills regardless of the logo. What you should stop doing is treating the book as format rehearsal. Format rehearsal is only worth the time when you know the format, and the format is set by whichever provider your admissions authority has actually contracted. That is a question with a checkable answer, and it is the one the next section answers.

Which provider sets your area's test for 2027 entry?

Every cell below comes from the admissions authority's own page, or from a member school of the consortium that runs the test. None of it comes from a tuition blog or a papers publisher. Where an authority publishes the structure of its papers, it is given in the authority's words. Where it does not, the cell is left empty rather than filled with something plausible — an inferred cell in a table like this is worse than a gap, because it looks exactly as authoritative as the sourced ones.

Essex is the row that does the most damage to the CEM-or-GL framing, because the Consortium of Selective Schools in Essex sits neither. It sets its own papers, and it does not test reasoning at all: "Your child will take two separate tests, one in English and one in Mathematics." The arrangement is long-standing rather than a recent reform — "The CSSE was formed in 1993 as a voluntary association" — and its purpose is to spare families a circuit of separate exams, since "the same selective test is utilised by ten CSSE selective schools and saves children, whose parents express a preference for more than one selective school, having to take multiple tests." A parent in Chelmsford drilling non-verbal reasoning grids is preparing for a paper their child will never sit.

Reading School breaks the other received rule, the one that says the test is sat in September of Year 6. Its published testing date for day and boarding applicants is Thursday 16th July, which falls in the summer term of Year 5, more than a year before the place is taken up. That is a different preparation calendar entirely, and a family who assumes an autumn exam loses a full term without noticing. September of Year 6 is the pattern in Buckinghamshire and Bexley; it is not a national rule and should never be assumed. The same caution applies across the wider set of admissions tests used by selective and independent schools, where the calendar is set school by school.

Which areas are changing provider for the next exam?

Two live switches are already contracted, and neither appears on any of the six pages currently ranking for this question. The Slough Consortium publishes both on its member schools' admissions pages. For the exam held in September 2026: "The Slough Consortium of Grammar Schools has contracted with GL Assessment as the test provider for the 11+ examination to be held in September 2026." For the exam a year later: "The Slough Consortium of Grammar Schools has contracted with Quest Assessments as the test provider for the 11+ examination to be held in September 2027."

Read those two sentences together and the practical consequence is sharp. Two children in the same Slough primary school, one year apart, will sit tests set by two different organisations. Any parent of the younger one who reads a comparison page today and settles into GL-format drilling is preparing for the test their neighbour's child takes, not the one their own child takes. Nothing about that is hidden — it is on the consortium's own admissions pages — but you have to know to look, and no aggregator page tells you to look.

Bexley is the second Quest area, and it is already there for 2027 entry: "Quest Assessments is the Bexley Selection Test provider." The council pairs that with a line worth reading twice if you are applying across a borough boundary: "Achieving a selective standard in the Kent Test or any other selection test will not be considered for the Bexley Selection Test." A Kent pass is a Kent pass. It buys nothing in Bexley, and a family targeting schools on both sides of that line is looking at two separate tests, two separate registrations and two separate results.

Slough also publishes something most authorities keep vague, which is what its threshold means in population terms: "The eligibility score will continue to be set at 111 on a standardised scale. Pupils who score 111 or above will be in the top 35% of the cohort sitting the examination and will be considered eligible for consideration for a grammar school place should their parents apply on their behalf." Note the wording the consortium chose — eligible for consideration. Clearing 111 puts a child into the pool. It does not put them into a school.

What is actually inside a GL paper?

GL is the default answer across much of England now, and it says so plainly on its schools-facing page: "We are the largest provider of 11+ tests to schools in England and deliver over 300,000 papers every year." It also claims a quality accreditation no rival advertises: "We are the only ISO 9001 accredited admissions test provider." Take that for what it is — a statement about process control and auditing, not evidence that any individual child's paper is fairer or harder. It is an argument aimed at procurement teams, and it is probably part of why councils keep choosing it.

The most repeated error about GL is that its test covers four clearly defined papers. What GL actually offers in four areas is practice material for parents: "We offer practice papers in four specific areas: verbal reasoning, non-verbal reasoning, English and maths, and there are three books (packs) available for each area as well as a bundle of all the books." That is a consumer range. It is not a specification of what your local authority commissioned, and treating one as the other is how families end up surprised on the day.

Gloucestershire shows the difference. Pate's Grammar School confirms the provider — "The test provider for the entrance test for entry into Year 7 in September 2027 will be GL Assessment" — and then describes a structure that is not four papers at all: "In one paper, children are assessed on their verbal skills (including comprehension, vocabulary and verbal reasoning). The other test paper assesses non-verbal reasoning and mathematical knowledge, skills and understanding. Each test paper will take around one hour and there will be a break between the papers." Buckinghamshire is built the same way: "Children sit two test papers, each about 60 minutes long, with a break between." Same provider, two papers, four skills bundled into them.

The correction that changes what you actually do at home is about scope. GL states that its English and maths admissions papers assess "areas taught up to the end of Year 5 (P6 in Northern Ireland)". Not the whole of Key Stage 2, and not the Year 6 curriculum a child is starting in the same month they sit the test. A child taking the paper in September of Year 6 is being examined on the year they have just finished. The implication is unglamorous and reliable: consolidate Year 5 arithmetic, vocabulary and comprehension until they are automatic, rather than racing ahead into content the paper will not ask about. If the gaps are in a specific subject rather than in test technique, our subject pages set out where that work usually starts.

Four areas, four scoring systems that do not compare

Almost every page on this subject says the pass mark sits somewhere between 111 and 121. That range is not wrong so much as it is a category error: it treats four genuinely different scoring models as points on one line. Two of the four systems below do not produce a number in that range at all, and one of them does not produce a single number in the first place. If you are trying to work out what your child needs, the model matters more than the threshold.

Buckinghamshire is the simple case: "grammar schools admit children who score at least 121 on the Secondary Transfer Test." One score, one line, published in advance. Slough is almost as simple at 111, with the useful addition that it tells you what the number means in rank terms — the top 35% of the cohort sitting the exam. Between those two, the familiar range holds. Then it stops holding.

Kent reports a set, not a number: "Your child will get 3 standardised scores, one for English, one for maths and one for reasoning, and a total (aggregate) score." The standard applied is a pair of conditions: "To be given a grammar school assessment, children needed a total score of 332 or more, with no single score lower than 108. Test scores range from 69 to 141. The highest possible total score is 423." The second condition is the one that catches families out. A child can clear 332 comfortably on the strength of one outstanding paper and still fall short because a single score sits below 108. In Kent, breadth beats a spike. Three solid papers are worth more than two brilliant ones and a weak one, which is the opposite of the instinct to pour revision into a child's best subject.

Bexley is different again, because it weights before it totals: "50% from the verbal ability score 25% from the numerical ability score 25% from the non-verbal ability score". Verbal ability is worth as much as numerical and non-verbal combined, which should reorder any Bexley revision plan on its own. And the resulting figure is not on the scale most parents assume: "the average total weighted age-standardised score is 200, with approximately two-thirds of candidates scoring between 170 and 230". A parent who hears a score of 190 and panics because they are thinking of 100 as average has misread the scale, not the child.

Kent adds one more component that no score captures. There is a writing exercise, and the council is careful about its status: "There will also be a writing exercise which will not be marked but may be used by a local headteacher panel as part of the headteacher assessment stage of the process. 40 minutes will be allowed for the writing task, including 10 minutes planning time." Unmarked but read. It carries no points and it can still matter, which is an uncomfortable thing to plan around — and a good argument for a child having written something under a 40 minute clock well before the day.

How competitive is it, really?

Most admissions authorities publish thresholds and stay quiet about outcomes. Bexley publishes both, which makes it the one place in this comparison where you can see the real shape of the funnel. The council states: "Only about 30% of children achieve a high enough mark to be able to apply to grammar school. Last year, 5,866 pupils sat the test, and 2070 were deemed selective." Those two published counts do not quite agree with the headline percentage: 2070 of 5,866 is 35.3%. Read the 30% as the council's rounded rule of thumb and the 35.3% as what its own two figures for that year actually produce.

The more revealing gate is the second one. Bexley states that "a total of 800 places will be available across these schools in September 2027" and, in the same breath, that "achieving the selective standard does not guarantee a grammar school place unless your child is among the top 180 highest scorers". So the selective standard is a qualification to apply, not an offer. Set the top 180 against the 2070 children who reached the selective standard and that inner band is 8.7% of them. Two thresholds, and most parents only know about the first.

Scale differs enormously between areas, and it changes what a threshold feels like. Warwickshire noted "there were over 3,500 Warwickshire children who sat the 11+ exam" when it moved provider. Bexley's 5,866 sit within a single London borough. Northern Ireland runs one assessment for the whole country, with "around 13,000 pupils" sitting each year. A percentile threshold in a cohort that size is a far more stable thing than the same percentile in a small county, where one unusually strong primary school cohort can move the distribution.

It is worth being clear about what these numbers do not show. Nothing in them says the tests have got harder. They say that in the areas that publish figures, more children reach the published standard than there are places behind it, and that authorities then apply a second rule — rank, distance, or a school's own oversubscription criteria — to decide who actually gets in. The preparation question that follows is not "how do we scrape the threshold". In an area with a rank-based second gate, scraping the threshold is not where the places are.

Why Northern Ireland is now one test, not two

If you have read that Northern Ireland runs two transfer tests, AQE and PPTC, that is out of date, and this page said it too until this update. There is one assessment, run by SEAG, and its outcome is accepted everywhere that selects: "SEAG provides an Entrance Assessment, held in November each year, for pupils who are in their final year of primary school. All 63 post-primary schools in N. Ireland which include academic selection within their Year 8 admissions criteria accept the SEAG Entrance Assessment Outcomes."

The provider question has a familiar answer. SEAG runs the process and GL does the assessment work: "We liaise with primary schools and we also liaise with GL Assessment who set and mark the Entrance Assessment." SEAG is also plain about what it replaced, and the description is a good measure of how much changed: "For 14 years, prior to the formation of SEAG, P7 pupils, whose parents were considering academically selective post-primary schools, needed to sit two different types of Entrance Assessment on four consecutive Saturdays to cover all 63 potential schools." Four Saturdays and two test types, replaced by one route.

The content is the part that most surprises families moving across from England, because there are no reasoning papers at all: "The SEAG Specification assesses knowledge and understanding in English (or Gaeilge) and Mathematics. It is based entirely upon the content and syllabus of the statutory Northern Ireland Curriculum for Language and Literacy and statutory Northern Ireland Curriculum for Mathematics and Numeracy at Key Stage 2." Entirely upon the statutory curriculum. That makes preparation a matter of curriculum security and timed accuracy rather than puzzle technique, and it leaves most of the England-focused practice market close to irrelevant.

On dates, SEAG's key dates page gives the first paper as "Saturday, 14th November 2026 Assessment Paper 1". That is the one date read directly from the source for this update, so it is the only one published here; the second paper and the outcomes date should be taken from SEAG's own key dates page rather than from any guide, including this one. Around 13,000 pupils sit the assessment each year, which makes it one of the largest single selective cohorts in the United Kingdom and, unusually, one governed by a single set of rules.

What does this mean for preparation now?

The instinct this whole search term encourages — identify the board, buy the matching books, drill the style — is the one to let go of. Provider is a variable. Slough changes provider between one exam and the next. Bexley moved. Warwickshire moved. What does not move is the underlying set of skills: reading quickly and accurately, arithmetic that is automatic rather than reconstructed, spotting a pattern under time pressure, and finishing a paper. Every provider on this page tests some combination of those, and none of them tests a book's house style.

There are also areas where matching practice material legitimately does not exist, and it is worth knowing which so you do not spend the autumn hunting for it. Reading School states it flatly: "Please note: there are no commercially available practice test papers for the FSCE Ltd tests." FSCE describes itself as a test written by the schools that use it — "bespoke and accessible 11+ and In-Year tests, written by Grammar Schools, for Grammar Schools". Cambridge, as quoted above, publishes nothing and endorses nothing. Anything sold as practice for those tests is somebody's guess at the format.

What replaces format-matching is fairly ordinary and quite specific. Work to the scope the provider actually publishes: for GL, that means content taught up to the end of Year 5, consolidated until it is quick. Build stamina for the real shape of the day — in Buckinghamshire and Gloucestershire that is two papers of roughly an hour with a break between, not a string of ten minute exercises. Where the scoring model rewards breadth, as Kent's per-paper floor of 108 does, protect the weakest paper rather than polishing the strongest. Where verbal work carries half the weight, as it does in Bexley, spend the time there.

Then check your own area yourself, once, properly, on the admissions authority's page rather than on a summary of it. The provider, the date, the number of papers, the threshold and the second gate behind the threshold are all published, and they are the five things that determine what the coming year looks like. Families weighing selective state entry against independent options usually need the same exercise for pre-tests and thirteen plus entry, where the timetable runs earlier still.

Frequently Asked Questions