GCSE Grade Boundaries Explained: How They’re Set in 2026

A GCSE grade boundary is the minimum raw mark an exam board sets, after marking is complete, for a student to be awarded a particular grade in a specific series. In the June 2026 series, AQA set the grade 9 boundary for GCSE Mathematics Higher tier at 219 out of 240 marks (91.25%), while Eduqas set the same grade 9 boundary at 185 out of 240 (77.1%) — a gap that reflects paper design, not a lower standard at Eduqas.

What Is a GCSE Grade Boundary, Exactly?

A grade boundary is the lowest total raw mark a student can achieve and still be awarded a given grade. Every GCSE paper is marked first, without reference to grades at all — examiners apply the mark scheme and produce a raw total for each candidate. Only once that marking is substantially complete does an exam board convert those raw totals into grades, by drawing a line: everything at or above the line for grade 7 is a 7, everything below it but at or above the grade 6 line is a 6, and so on down to grade 1 and then U (ungraded).

Because GCSEs are linear qualifications, all of a subject's papers are added together into one overall mark before any grade boundary is applied. AQA's GCSE Mathematics (8300) Foundation tier has a maximum of 240 marks across three papers. If the grade 4 boundary is 154, the arithmetic a student needs to clear it is simple: 154 ÷ 240 = 64.2% of the total mark. A student who scores 153 gets a grade 3; a student who scores 154 gets a grade 4. There is no partial credit for being close, which is exactly why a one-mark shortfall matters so much to the family experiencing it, even though it changes nothing about how the system as a whole is designed.

Why Do Grade Boundaries Move Every Series?

Government guidance on GCSE and A level grading, published jointly by the Department for Education and Ofqual, explains the mechanism directly: “If an exam is easier than in previous years, the grade boundaries for that paper will be higher. If it is harder, the grade boundaries will be lower.” The difficulty of a set of exam questions varies from year to year even when boards try to hold the level of demand constant, because it is genuinely impossible to know in advance exactly how a cohort of students will find a paper.

The process that sets the final numbers is not purely statistical. Exam boards first analyse mark data to identify potential boundaries, comparing this year's cohort with previous years. Then, in the government's own words, “senior examiners then scrutinise samples of work from students who have achieved marks on and around these potential grade boundaries. This is compared with work of students from the previous year who were on the same grade boundaries.” Only once that scrutiny confirms the standard of work at a potential boundary is appropriate do senior examiners recommend the final mark.

The Boundary Did Not Move Because of a Quota

The most common misreading of this whole system is that a moving boundary means an exam board is managing a target pass rate — that a subject “got harder” because too many students were about to pass, or that boards quietly ration top grades. That is backwards, and the Department for Education and Ofqual say so in explicit terms: “There are no quotas for the number of students who must pass or fail a subject each year, and no cap on the number of students that can get each grade.”

The purpose of moving the boundary is the opposite of a quota. Official guidance states it plainly: “Grade boundaries are set with the aim of maintaining standards over time, not to ensure that the same proportion of students get a particular grade.” The reason grade proportions look stable most years is that student attainment is itself broadly stable most years — the stability is an observed outcome, not an engineered one. And the same logic applies across exam boards, not just across years: “Where grade boundaries increase between one year and the next, or are higher with one board than another, it does not mean it is more difficult to achieve a grade. Rather, it reflects that that exam paper was easier.” A high boundary is evidence of an easy paper, not a hard grade to reach.

How the National Reference Test Moved This Year’s Maths Boundaries

Every February and March, a representative sample of Year 11 pupils at more than 300 schools in England sit the National Reference Test (NRT) — a fixed-format test, running since 2017, used to check whether GCSE English language and maths attainment is genuinely drifting over the long term rather than just fluctuating within ordinary year-to-year noise. The NRT itself does not set anyone's grade; it exists purely to sanity-check the wider system.

FFT Education Datalab, an independent research organisation that publishes results-day analysis of the national exam data, reported that the 2026 NRT showed a long-term rise in attainment at grades 9-7 and 9-5 in maths compared with the 2017 baseline. Ofqual acted on that evidence: boundaries for the higher grades in GCSE maths were shifted down slightly for the June 2026 series, so more pupils reached the top grades without the standard of work actually falling.

FFT quantified the scale of that shift in its own analysis, and it is worth quoting exactly rather than paraphrasing: “In practice this is a relatively small change: 21.2% of entries in maths received a grade 7 or above this year, compared with 20.3% last year. That's an increase of 0.8 percentage points.” Those are FFT's own figures and FFT's own stated delta — worth flagging, because 21.2 minus 20.3, taking the two rounded percentages alone, comes to 0.9, not 0.8. FFT's 0.8 is drawn from their underlying, unrounded data, which is a normal rounding difference rather than an error on anyone's part. Nationally, the effect was smaller still: FFT's analysis of the full June 2026 results shows the overall proportion of GCSE entries graded 7 or above rose from 23.0% to 23.1%, and grade 4 or above rose from 70.5% to 70.7%.

The same 2026 NRT round found lower attainment in English language at every grade boundary compared with 2017 — but Ofqual did not adjust English language boundaries, judging the difference not large enough to act on.

GCSE Maths Grade Boundaries, Summer 2026: AQA, Edexcel, OCR and Eduqas

Mathematics is assessed at two tiers — Foundation, which awards grades 5 down to 1, and Higher, which awards grades 9 down to 3 — so a Foundation grade 5 and a Higher grade 5 sit on completely different mark scales and cannot be compared directly. The tables below give the published boundaries for each board's June 2026 series, taken directly from each board's own grade-boundary document.

Foundation tier

Read OCR's Foundation grade 4 boundary carefully: 133 out of 300 is 44.3%, well below AQA's 154 out of 240 (64.2%) for the same nominal grade. That is not OCR being more generous. OCR's Foundation maths paper is marked out of 300 across three papers rather than 240, so the two percentages are not measuring the same thing at all — only the standard of work behind each board's own boundary is calibrated to be equivalent, never the raw percentage.

Higher tier

Notice how far the Higher tier grade 4 boundary sits below half marks at every single board — from Eduqas's 11.7% up to AQA's 26.25%. That is normal and expected, not evidence of an easy grade to reach: the Higher tier paper assesses content up to and including the hardest topics on the specification, so even a student heading for a bare pass has to attempt questions built for students aiming at grade 9. For families weighing which tier to enter, our guide to Foundation versus Higher tier for a resit works through the decision in more detail, including who actually makes the final call.

Combined Science follows the same logic but awards two grades at once. AQA's Combined Science: Trilogy is marked out of 420 across six papers. In the June 2026 Higher tier series, the 7-7 boundary sat at 228 out of 420 (54.3%), and the 4-4 boundary sat at 111 out of 420 (26.4%) — a double-award qualification that counts as two GCSEs, built from exactly the same marking and boundary-setting process as a single-award subject.

GCSE English Language Grade Boundaries, Summer 2026

English Language has no tiers — every candidate sits the same papers and is eligible for grades 9 down to 1. That makes it a cleaner subject for a straight cross-board comparison, with one caveat: Eduqas marks the qualification out of 200, while AQA, Edexcel and OCR all mark it out of 160, so the raw marks in the table below are not directly comparable across that one row without converting to a percentage first.

Working the arithmetic for one row: Edexcel's grade 4 boundary in English Language was 87 out of 160 marks. 87 ÷ 160 = 54.4%, comfortably the highest grade 4 percentage of the four boards for this subject, against AQA's and OCR's 46.9% and Eduqas's 44.5%. None of these numbers tell you which paper was “hardest” in any meaningful sense; they tell you what each board's senior examiners judged the standard of work at each boundary to require, on that board's own paper, in that series.

Why Do Two Boards Set Very Different Boundaries for the Same Grade?

The clearest illustration on this page is GCSE Mathematics Higher tier, grade 9, June 2026 series. AQA's boundary was 219 out of 240 marks, or 91.25% of the total. Eduqas's boundary, on the identical 240-mark scale, was 185 out of 240, or 77.1%. That is a gap of 34 marks and roughly 14 percentage points for what both boards call, and certify as, the same grade 9.

The instinct this invites — that Eduqas must be “easier” for grade 9, or that a grade 9 from Eduqas is worth less — is exactly the misreading official guidance warns against. AQA and Eduqas write entirely separate papers, with different questions, different balances of question types and different overall demand, even though both must cover the same specification content and assessment objectives. A paper that turns out more demanding produces a lower boundary for the same standard of work; a paper that turns out less demanding produces a higher one. Government guidance is explicit that a lower boundary at one board “reflects that the paper was harder, not that it is easier to get a particular grade” there. Universities, colleges and employers treat a 9 from any board as the same grade, because the senior-examiner scrutiny process described earlier exists precisely to make that true, whatever the raw mark required to reach it.

The same pattern holds for English Language, just less dramatically: a 20-percentage-point spread across Higher tier maths grade 9 narrows to roughly a 10-point spread across grade 4 in English Language. Our AQA vs Edexcel vs OCR comparison sets out approximate boundary ranges across recent series if you want a wider view than this single series.

What to Do If You’re One or Two Marks Below a Boundary

The most direct option is a review of marking, requested through your school or college rather than by you directly. This is a genuine re-mark against the mark scheme by another examiner, not a one-way appeal — every board states explicitly that there is no grade protection, so a mark can move down as well as up, or not change at all. Our separate guide, is a remark worth it, works through Ofqual's own published data on how often grades actually change on review, board by board, and the exact 2026 deadlines and costs.

If a review is not the right route — because the gap is too large, or because the deadline has passed — the more useful question is what the missed grade actually blocks, and by when it needs resolving. Our guide to what to do if you missed your GCSE grades sets out the resit calendar, sixth form appeal routes and the realistic timeline for each. Where the answer is targeted preparation for a November resit or the next academic year, our GCSE tuition pages set out how our specialist tutors structure that work around the specific topics that cost the most marks.