Which Grade 9 Is Hardest? The 2026 GCSE Raw-Mark Ranking
AQA’s June 2026 boundaries show a 26-point gap in the marks needed for a grade 9 — from Biology’s 66.5% to Panjabi’s 92.5%.
Book a Free ConsultationAQA’s June 2026 boundaries show a 26-point gap in the marks needed for a grade 9 — from Biology’s 66.5% to Panjabi’s 92.5%.
Book a Free ConsultationA GCSE grade 9 costs a different share of the total marks in every subject: AQA’s June 2026 boundaries show Biology needed the smallest share, 133 of 200 marks (66.5%), while Panjabi needed the largest, 222 of 240 (92.5%). That is a 26-percentage-point spread between the least and most demanding subject on the same exam board, in the same series — a gap no single-subject explainer or attainment-rate list makes visible in one number.
Every figure in this ranking is a raw-mark share: the grade 9 boundary AQA published in its June 2026 subject grade boundaries document for a subject, divided by that paper’s maximum mark. AQA sets a boundary only once marking is substantially complete, so it can see how the whole cohort actually performed on that year’s specific paper — the mechanics of how and why a boundary moves are covered in full in our GCSE grade boundaries explainer; this page does not repeat that ground.
A raw-mark share is not a difficulty score. It shows where the boundary had to sit for the cohort that actually sat that paper — a high share means the marks were clustered high enough that the boundary had to sit high to keep roughly the same proportion of students at grade 9 as in previous years. It does not mean the paper itself asked harder questions, and it does not tell you how hard the subject is to learn from nothing.
Three entries in the table below need that caveat attached directly. Panjabi, Urdu and Chinese (Spoken Mandarin) are marked with ✱ because all three are community-language GCSEs, sat by comparatively small, self-selected entries that include many candidates who already speak the language at home — a very different population from the near-universal Mathematics or English Language cohort. A cohort weighted that way scores higher overall on the same paper, which pushes the boundary itself higher. Read Panjabi’s 92.5% as ‘this cohort’s marks clustered high’, not as ‘the hardest GCSE’.
Most published ‘hardest GCSE’ lists use a different measure entirely: the attainment rate, meaning what share of entries achieved a grade 9. That measure answers a different question — how many people cleared the bar — and it is shaped by who chooses to sit the subject as much as by anything else. Section 6 below sets the two measures side by side using Ofqual’s and FFT Education Datalab’s own summer 2026 figures, because a subject can disagree with itself across the two measures.
The table below ranks AQA GCSE titles by the share of total raw marks their June 2026 grade 9 boundary represented. Tiered subjects (Biology, Chemistry, Physics, French, German, Spanish and others) show the Higher tier figure, since grade 9 is only awarded on Higher tier. Combined Science is deliberately excluded from this table and discussed separately in the next section, because its top grade is a double award (9-9) marked out of 400 or 420, not a single-subject grade on the same scale as the rows below.
| Subject (AQA, June 2026) | Grade 9 boundary | Share of total marks |
|---|---|---|
| Panjabi ✱ | 222 of 240 | 92.5% |
| Mathematics | 219 of 240 | 91.25% |
| Urdu ✱ | 219 of 240 | 91.25% |
| Design and Technology | 179 of 200 | 89.5% |
| French | 179 of 200 | 89.5% |
| Computer Science | 160 of 180 | 88.89% |
| English Literature | 142 of 160 | 88.75% |
| Engineering | 177 of 200 | 88.5% |
| Chinese (Spoken Mandarin) ✱ | 212 of 240 | 88.33% |
| Citizenship Studies | 141 of 160 | 88.12% |
| German | 176 of 200 | 88% |
| Spanish | 175 of 200 | 87.5% |
| Food Preparation and Nutrition | 172 of 200 | 86% |
| Economics | 136 of 160 | 85% |
| Psychology | 168 of 200 | 84% |
| Art and Design (all six titles) | 403 of 480 | 83.96% |
| Drama | 167 of 200 | 83.5% |
| Dance | 333 of 400 | 83.25% |
| Physical Education | 216 of 260 | 83.08% |
| Music | 198 of 240 | 82.5% |
| Sociology | 164 of 200 | 82% |
| Geography | 204 of 252 | 80.95% |
| Physics | 157 of 200 | 78.5% |
| Chemistry | 155 of 200 | 77.5% |
| Media Studies | 184 of 240 | 76.67% |
| English Language | 122 of 160 | 76.25% |
| Business | 133 of 180 | 73.89% |
| Statistics | 114 of 160 | 71.25% |
| Biology | 133 of 200 | 66.5% |
✱ Community-language GCSE with a comparatively small, self-selected entry — see the caveat above before reading this as a difficulty ranking.
Trying to work out what a specific boundary means for your child’s next set of GCSEs? A published share tells you where the whole cohort landed, not which topics cost your own child the marks that mattered.
Book a Free Consultation Message us on WhatsAppBiology’s Higher tier grade 9 boundary in June 2026 was 133 out of 200 marks — 66.5%, the lowest share in this entire dataset, single or double award. Chemistry and Physics, sat by many of the same students on the same exam board in the same series, needed meaningfully more: Chemistry’s Higher tier grade 9 was 155 of 200 (77.5%) and Physics’s was 157 of 200 (78.5%). That is a 12-percentage-point spread across three subjects that a school timetable, and most parents, treat as one interchangeable block: ‘the sciences’.
Combined Science complicates the picture rather than resolving it, because its top grade is not a single 9 — it is a double award, 9-9, marked across six papers rather than two. AQA runs two Combined Science routes. Trilogy’s 9-9 boundary was 298 out of 420 marks (70.95%); Synergy’s 9-9 boundary was 267 out of 400 (66.75%). Both sit above standalone Biology’s 66.5%, and both sit below standalone Chemistry and Physics taken separately. Because a double-award boundary is not directly comparable to a single-subject boundary on the same axis, neither figure appears in the main ranking table above — they are reported here, on their own terms, instead.
The practical reading: on raw marks alone, Biology is the most forgiving of AQA’s three separate sciences, and both combined-science routes ask for a smaller share of the total mark than either Chemistry or Physics sat on their own. FFT Education Datalab’s analysis of the same results day supports the same shape from a different angle — Biology’s attainment rate at the top grades barely moved in 2026, up just 0.1 percentage points, while Chemistry rose 1.0 points and Physics rose 0.9 points, so the two measures agree that Biology sits apart from its two sister sciences this series.
Some GCSE titles do not have one grade-9 mark at all. AQA’s GCSE History (specification 8145) is built from a Period Study, a Wider World Depth Study and a British Depth Study, and centres choose which combination to teach — AQA publishes a separate grade boundary for every combination, 153 option codes in the June 2026 document alone. Their grade 9 raw-mark share ranges from 68.45% (option 8145BA, 115 of 168 marks) to 77.38% (option 8145QL, 130 of 168) — a spread of almost 9 percentage points inside a single GCSE title, before any comparison to another subject at all.
Religious Studies shows the same pattern on a smaller scale. Across AQA’s Short Course (8061) and its two full-course routes (8062 and 8063), there are 34 separate option combinations in the June 2026 document. Their grade 9 share ranges from 79.41% (Short Course Option E, 81 of 102 marks) to 87.25% (Short Course Option B, 89 of 102) — an 8-point spread that a single ‘RS grade 9%’ headline figure erases completely.
The consequence for a parent comparing schools or specifications: a figure like ‘AQA History needs 73% for a 9’ is only ever true for one specific option code, and the true range in June 2026 ran from 68.45% to 77.38%. Always check the option code your child’s centre actually teaches before treating any single History or Religious Studies percentage as universal.
Every figure so far is AQA’s, and the board matters. OCR’s GCSE Mathematics Higher tier (J560), marked out of 300 rather than AQA’s 240, set its June 2026 grade 9 boundary at 264 marks — 88.0%, against AQA’s 91.25% for the same subject in the same series. That is a 3.25-percentage-point gap for a subject every Year 11 in England sits, on papers set to the same national standard but not the same raw-mark share.
OCR’s own GCSE Physics specifications disagree with each other by even more. OCR offers two Physics routes: Physics A, branded Gateway Science, and Physics B, branded Twenty First Century Science. Physics A’s Higher tier grade 9 in June 2026 was 152 of 180 marks (84.44%); Physics B’s was only 141 of 180 (78.33%) — a 6-point gap between two GCSEs with the same subject name, the same exam board, the same maximum mark and the same series, distinguished only by specification.
The pattern holds across every comparison we checked: the subject name alone never fixes the answer to ‘how hard is a 9?’ The board, the specification and, for History and Religious Studies, the option code all move the number independently of the subject.
A raw-mark share and an attainment rate answer different questions, and they can point in opposite directions for the same subject. A raw-mark share asks: what fraction of the available marks did this year’s boundary require? An attainment rate asks: what fraction of entries actually reached grade 9? A subject sat almost entirely by strong, self-selecting candidates can have both a high raw-mark share and a high attainment rate at once — the two measures are not opposites, but they are not the same question, and a page that only publishes one of them is answering half of what ‘hardest’ can mean.
Ofqual’s official summer 2026 release is one clean anchor for the attainment side: only 0.2% of the 551,950 sixteen-year-olds who sat seven or more GCSEs achieved a grade 9 in every single one, 0.3% of girls and 0.2% of boys. That figure says nothing about any one subject’s raw-mark share — it measures how rare it is to clear the bar everywhere at once.
FFT Education Datalab’s results-day analysis shows how the two measures can move independently within a single series. Nationally, the proportion of GCSE entries graded 7 or above rose from 23.0% to 23.1% between summer 2025 and summer 2026 — a 0.1-point shift. Mathematics moved by more: its grade-7-or-above rate rose from 20.3% to 21.2% after Ofqual adjusted higher-grade Maths boundaries down, using evidence from the 2026 National Reference Test that this year’s cohort was performing very slightly above 2017 levels at grades 9-7 and 9-5. A falling raw-mark boundary and a rising attainment rate happened in the same subject, in the same series, for the same reason — which is exactly why the two measures need to be read separately, not folded into one number.
Spanish shows the same disconnect from the other direction: FFT records the largest rise in top-grade attainment of any GCSE subject in summer 2026, from 26.0% to 27.8%, with German and French also rising by 0.7 and 0.6 points. None of that attainment-rate movement is visible in this page’s raw-mark ranking, because a raw-mark share is fixed once the boundary is set for that series — it does not track how the following year’s cohort performs. Treat the ranking table above as the answer to ‘where was the boundary set’, and this section as the answer to ‘how many students got there’; they are not the same question, and neither one alone is the full answer to ‘which grade 9 is hardest’.
By raw-mark share, AQA’s Panjabi (222 of 240 marks, 92.5%) and Mathematics and Urdu (219 of 240, 91.25% each) needed the largest share of available marks for a grade 9 in the June 2026 series. But Panjabi and Urdu are community-language GCSEs with far smaller, more skewed entries than Maths, so a high share there reflects who sits the exam as much as how hard it is — read the caveat above the ranking table before treating this as a straight difficulty order.
Biology: AQA’s Higher tier grade 9 boundary in June 2026 was 133 out of 200 marks, 66.5% — the lowest raw-mark share of any AQA subject in this ranking, single or double award. Both combined-science double awards sat above it: Synergy at 267 of 400 (66.75%) for a 9-9, and Trilogy at 298 of 420 (70.95%) for a 9-9. Chemistry (77.5%) and Physics (78.5%) both needed a meaningfully higher share than Biology on the same board and series.
Extremely, once you look across a student’s whole set of GCSEs rather than one subject. Ofqual’s official summer 2026 release records that only 0.2% of the 551,950 sixteen-year-olds who sat seven or more GCSEs achieved a grade 9 in every single one — 0.3% of girls and 0.2% of boys. That is a different question from any single subject’s raw-mark share: it measures how many students clear the bar everywhere at once, not where the bar sits in one subject.
By raw-mark share, yes, clearly. AQA’s June 2026 boundaries put Mathematics Higher tier’s grade 9 at 219 of 240 marks (91.25%), against English Literature’s 142 of 160 (88.75%) and English Language’s 122 of 160 (76.25%) — a 15-percentage-point gap between Maths and English Language specifically. English Language’s lower share is one reason attainment-rate lists and raw-mark lists can disagree: the two measures are not answering the same question.
Yes. AQA’s GCSE Mathematics Higher tier needed 219 of 240 marks (91.25%) for a 9 in June 2026; OCR’s Mathematics Higher tier, marked out of 300, needed 264 (88.0%) — a 3.25-point gap for the same subject and series. OCR’s own two GCSE Physics specifications disagree with each other too: Physics A (Gateway Science) needed 152 of 180 (84.44%), Physics B (Twenty First Century Science) needed only 141 of 180 (78.33%). The subject name alone does not fix the answer; the board and the specification do.
Because both subjects are built from optional units, and AQA publishes a separate boundary for every combination. GCSE History (8145) has 153 different option codes in the June 2026 document, and their grade 9 raw-mark share ranges from 68.45% (option 8145BA) to 77.38% (option 8145QL) — nearly a 9-point spread inside one subject. Religious Studies spans 34 option combinations across its Short Course and two full-course routes, from 79.41% to 87.25%. A single ‘GCSE History grade 9%’ figure does not exist without naming the option code.
By raw-mark share for the top double grade, yes, on this evidence. AQA’s Combined Science: Trilogy needed 298 of 420 marks (70.95%) for a 9-9, and Combined Science: Synergy needed 267 of 400 (66.75%) — both are double-award GCSEs graded as a pair, not a single subject grade, so they are not directly the same object as Biology’s single 133-of-200 (66.5%) grade 9. Even allowing for that difference, neither combined-science route needed a higher raw-mark share than Chemistry (77.5%) or Physics (78.5%) taken separately.
No, and this is the single most important caveat in this data. A raw-mark share shows where that year’s grade boundary had to sit for the cohort who actually sat the paper — it does not measure how difficult the content is to learn from nothing. Panjabi (92.5%) and Urdu (91.25%) top this ranking partly because both are community-language GCSEs, sat by comparatively small, self-selected entries that include many candidates who already speak the language at home; a cohort like that scores higher overall, which pushes the boundary itself higher. Read a high share as ‘this cohort’s marks clustered high’, not as ‘hardest subject’.
A published boundary tells you what the whole national cohort needed. It does not tell you which topics cost your own child the marks that mattered.
Leading Tuition is rated 4.8 out of 5 on Trustpilot from 58 reviews, and 91% of our students achieve their desired grades. See also our guide to getting a grade 9 in GCSE Maths and our GCSE tuition service.
Book a Free Consultation Message us on WhatsApp