A-Level Bell Curve Singapore - H2 Grading, Distinction Rates & What Marks You Need (2026)
Study guide/
A-Level Bell Curve Singapore - H2 Grading, Distinction Rates & What Marks You Need (2026)
In one line
MOE officially states that A-Level grades are criterion-referenced (not curved), but statistical moderation is applied to ensure year-on-year comparability.
Key points
In practice, the raw mark needed for an A grade varies each year.
Estimated A-grade cut-offs from school reports and tuition data: H2 Maths 70–78, H2 Chemistry 72–78, H2 Physics 70–76, H2 Biology 72–78 (out of 100).
Planning a revision session? Use our study places near me map to find libraries, community study rooms, and late-night spots.
Read in layers
1 second
Read the summary above.
10 seconds
Scan the first few sections below.
100 seconds
Jump into the section that matches your decision.
Quick reading map
Concrete example: using an estimated boundary
What MOE and SEAB actually say
Estimated distinction rates by subject
Q: How does the A-Level bell curve work in Singapore? A: Technically, there is no "bell curve" - MOE and SEAB use criterion-referenced grading with statistical moderation. In practice, grade boundaries shift each year based on paper difficulty and cohort performance, which is why students and parents experience it as a bell curve. This guide covers what we know about grade boundaries, estimated distinction rates, and how moderation works.
TL;DR MOE officially states that A-Level grades are criterion-referenced (not curved), but statistical moderation is applied to ensure year-on-year comparability. In practice, the raw mark needed for an A grade varies each year. Estimated A-grade cut-offs from school reports and tuition data: H2 Maths 70–78, H2 Chemistry 72–78, H2 Physics 70–76, H2 Biology 72–78 (out of 100). These are estimates, not official figures - SEAB does not publish grade boundaries.
If you have...
Read this first
1 second
SEAB does not publish A-Level grade boundaries.
10 seconds
Check criterion-referenced grading, statistical moderation, estimated A ranges, distinction rates, prelim difficulty, subject differences, raw marks, and safer score targets.
100 seconds
Do not plan around a guessed boundary. Use estimates only to set a buffer, then aim above the danger zone.
Concrete example
If an estimated A range is 70 to 78, target 80 or higher rather than betting on 70.
Best next step
Convert prelim mistakes into topic targets instead of trying to game moderation.
Why your school prelim grade is not your A-Level grade
What to do with the information
Does the bell curve help or hurt you?
Concrete example: using an estimated boundary
If an estimated A range is 70 to 78, do not aim for exactly 70. Treat 70 as a danger zone and 80+ as a safer revision target, because the actual boundary depends on paper difficulty and cohort performance.
What MOE and SEAB actually say
SEAB's official position is that A-Level grades are criterion-referenced: each grade (A, B, C, D, E, S, U) represents a defined standard of achievement, not a percentile rank. In theory, if every student achieves the required standard, every student could receive an A.
In practice, SEAB applies statistical moderation to maintain comparability across years. If a paper is unusually difficult, the grade boundaries are adjusted downward so that students are not penalised for a harder exam. Conversely, if a paper is easier, boundaries may shift upward.
This moderation process is what students and parents experience as a "bell curve" - even though it is technically different from a pure norm-referenced curve.
Key point: SEAB does not publish grade boundaries, percentile data, or distinction rates. All figures in this guide are estimates based on school-level data, tuition centre reports, and community analysis.
Estimated distinction rates by subject
These are community-sourced estimates, not official MOE/SEAB data. They reflect the percentage of candidates receiving a grade of A.
Subject
Estimated National A Rate
Top JC A Rate
Mid-Tier JC A Rate
H2 Mathematics (9758)
~50%
80–90%
30–45%
H2 Physics (9478)
~40–45%
60–80%
20–35%
H2 Chemistry (9476)
~38%
60–75%
25–35%
H2 Biology (9477)
~40%
55–70%
25–35%
H2 Economics (9570)
~45%
70–85%
30–40%
H1 General Paper (8881)
~25–30%
40–55%
15–25%
Important caveats:
The H2 Maths distinction rate is the highest partly because many strong students take it, not because the subject is easiest.
School-level distinction rates vary enormously - a 50% national rate masks a range from 15% to 90% across different JCs.
These estimates are based on pre-2026 data (9749 Physics, 9729 Chemistry, 9744 Biology). The new 9478, 9476, and 9477 syllabi may shift these rates.
Estimated grade boundaries (raw marks for each grade)
These are rough estimates compiled from school prelim statistics, tuition centre data, and community analysis. They assume a total of 100 marks across papers.
Grade
H2 Maths
H2 Physics
H2 Chemistry
H2 Biology
A
70–78
70–76
72–78
72–78
B
60–69
60–69
62–71
62–71
C
55–59
55–59
55–61
55–61
D
50–54
50–54
50–54
50–54
E
45–49
45–49
45–49
45–49
S (sub-pass)
40–44
40–44
40–44
40–44
U (ungraded)
<40
<40
<40
<40
Critical warning: These are estimates, not official figures. The actual cut-off varies each year based on paper difficulty and cohort performance. Do not plan your revision around hitting a specific mark - aim as high as possible.
Why your school prelim grade is not your A-Level grade
JC preliminary exams ("prelims") are set by each school independently and are typically harder than the actual A-Levels. This is intentional - schools want to push students before the final exam.
Common patterns:
A prelim C/D often converts to an A-Level A/B.
A prelim U/S can convert to an A-Level C/D.
Top JC prelims are significantly harder than neighbourhood JC prelims, so a C at RJC may indicate stronger absolute ability than a B at a neighbourhood JC.
Do not panic about prelim grades. Use them as diagnostic tools to identify weak topics, not as predictions of A-Level performance. Many students report improving by 1–2 grades between prelims and the actual exam, though SEAB does not publish official data on this.
Does the bell curve help or hurt you?
The moderation process can work for or against you depending on the year:
If the paper is unusually hard: Grade boundaries drop, and more students receive A grades than their raw marks would suggest. This helps everyone proportionally.
If the paper is unusually easy: Grade boundaries rise, and fewer students receive A grades than expected. This means a raw 75% might only earn a B in an easy year.
In a typical year: The moderation effect is small (±2–3 marks on the A boundary), and grade boundaries are relatively stable.
The practical implication: you cannot "game" the bell curve. Focus on maximising your own raw score rather than worrying about how hard the paper will be for others.
Subject-specific grading insights
H2 Mathematics
The highest national distinction rate (~50%) reflects both strong student selection and relatively predictable exam formats.
The A boundary is stable at roughly 70–75 raw marks in most years.
Students who lose marks typically lose them on the final 2–3 marks of long questions (the "show" or "prove" steps that require clean algebraic manipulation).
H2 Chemistry
The lowest distinction rate (~38%) among the major H2 Sciences.
Organic chemistry marks are the most variable - students either know the reactions or they do not.
The A boundary for H2 Chemistry is slightly higher than Physics because the cohort-wide performance tends to cluster more tightly.
H2 Physics
Distinction rate estimated at 40–45%.
Paper 4 (Practical) marks can be highly variable - students who have not practised spreadsheet analysis can lose 10+ marks here.
The new 9478 syllabus may shift grade boundaries as students adapt to the spreadsheet requirement.
H2 Biology
Distinction rate estimated at ~40%.
Essay marks (Paper 3) are subjective and can vary between markers, though SEAB standardises marking.
Students who can connect molecular-level detail to macro-level biological outcomes score disproportionately well.
FAQ
Is there really a bell curve for A-Levels in Singapore?
Not technically. MOE uses criterion-referenced grading with statistical moderation. However, the effect is similar to a bell curve in practice: grade boundaries shift based on paper difficulty and cohort performance, and a fixed proportion of students typically receives each grade band.
What marks do I need for A in H2 Maths?
Estimated 70–78 raw marks out of 100 (combined across Papers 1 and 2). This varies by year - an unusually hard paper may lower the boundary to ~68, while an easy paper may raise it to ~80. Aim for 75+ to be safe.
Is 70 marks enough for A in H2 Chemistry?
Possibly, but it is close to the borderline. Estimated A boundaries for H2 Chemistry are 72–78. In a year with a harder paper, 70 may be enough; in an easier year, it may only earn a B. Do not rely on borderline scores - aim for 80+ to be confident.
Why is my prelim grade so much worse than my school CA grades?
Prelim papers are designed to be harder than the A-Level to push students. Many schools deliberately set prelim papers 10–15% harder than the expected A-Level difficulty. A prelim D often converts to an A-Level B with proper revision in the final 2–3 months.
Do SEAB examiners deliberately set papers to fail a certain percentage?
No. SEAB sets papers to test the syllabus at appropriate difficulty levels. However, moderation ensures that year-on-year grade distributions remain broadly comparable. This means that if a paper is unexpectedly easy, the A boundary will be higher - but this is a quality-control mechanism, not a deliberate attempt to fail students.
Does the bell curve affect H1 GP differently?
H1 GP has a lower distinction rate (~25–30%) than most H2 subjects. This reflects both the broader student base (every JC student sits GP) and the subjective nature of essay marking. The moderation process applies similarly, but the content type (essays, comprehension) means that mark distributions tend to be wider.
