Singapore Education Reforms in 2026: Fairer School Access, Smaller Classes and AI—But Will Students Learn More Independently?

Education news

Oct 16

Singapore Education Reforms in 2026: Fairer School Access, Smaller Classes and AI—But Will Students Learn More Independently

Singapore is entering another important phase of education reform. Changes to Primary 1 registration, plans to test smaller classes and lighter non-teaching duties, and expanded artificial intelligence literacy all suggest a common ambition: a system that is more inclusive, more responsive and better equipped for the future. Yet a difficult question remains. Will the reforms change the day-to-day experience of learning, or simply reorganise how pupils enter schools, sit examinations and use technology?

Imagine a hypothetical family in Singapore planning their child's first school application. The parents compare distance, admission phases and a school's reputation. Their child, meanwhile, is much more interested in whether classmates will be kind, whether lessons will spark curiosity, and whether making a mistake will feel safe. At another school, a teacher is planning a lesson that must serve pupils with different strengths while also managing administrative tasks. Neither family nor teacher wants a slogan. Both want a system that works.

The latest announcements deserve careful attention—not because Singapore has suddenly abandoned standards, but because they reveal an ongoing attempt to reconcile academic expectations with equity, teacher capacity, individual strengths and learner agency. The next challenge is to connect policy design with what happens in the classroom.

What has actually changed in Singapore's education system?

As of 16 October 2026, several measures are at different stages. They should not be described as if they have all been implemented.

Primary 1 admissions: On 10 September, the Ministry of Education announced changes to the 2027 registration exercise, for children entering Primary 1 in 2028. Most schools will reserve 60 Phase 2C places instead of 40. Twelve specified schools will operate a two-track system within that phase. On 7 October, Education Minister Desmond Lee confirmed there would be no transition period. See CNA's original report and October parliamentary follow-up.
Classroom redesign: At the 16 September MOE Work Plan Seminar, the ministry announced that six schools would pilot ideas from 2027, including the potential benefits of smaller classes and a different distribution of teachers' non-classroom responsibilities. This is a pilot, not a nationwide class-size reduction. See CNA's 16 September report.
AI literacy: MOE's 2026 announcements state that the updated Code for Fun programme, with stronger AI skills, will be available to all schools in 2027. Its AI-in-Education Framework, updated on 10 September 2026, emphasises agency, inclusivity, fairness and safety.
Secondary pathways: The new Singapore-Cambridge Secondary Education Certificate examination begins in 2027, under Full Subject-Based Banding. A new post-secondary admissions exercise is planned for 2028. These are separate changes with separate implementation dates. See SEAB's SEC explanation and MOE's 2026 announcements.

The reforms are related, but not identical. Admissions is about opportunity to enter a school. Class design is about the capacity to teach. Examinations shape incentives. AI alters how learners access and produce knowledge. A strong education strategy must connect all four.

Primary 1 registration in 2027: opening doors without pretending competition disappears

The most immediate political debate concerns Primary 1 registration. From the exercise held in 2027, most primary schools will set aside 60 places for Phase 2C, up from 40. This phase is available to children without a prior connection to a particular school. Schools with fewer than 180 Primary 1 places retain the 40-place reservation, so the rule is not universal.

At 12 selected schools located in areas with relatively little public housing close by, Phase 2C places will be divided equally into two tracks: children living within 2 km and those living beyond 2 km. Within either track, there is no additional home-to-school distance priority. Citizen applicants continue to receive priority under the applicable rules, and balloting may still be necessary when demand exceeds places. These details matter more than the simplistic headline that distance no longer counts.

The ministry's stated objective is to broaden access and encourage children from different social backgrounds to learn together. In a parliamentary exchange on 7 October, the minister said the changes would start without a transition period, despite concerns from families that had already made residential decisions under existing expectations. The government indicated that implementation would be monitored and reviewed.

Both the rationale and the criticism deserve consideration. A school can be publicly funded yet difficult for some families to access because of its location and the cumulative advantages of prior connections. Increasing open places may widen opportunity. At the same time, a new admission rule can impose unexpected costs on parents who have organised housing, commuting and childcare around the old system.

There is also a deeper problem that no allocation formula can completely solve: when parents perceive a small number of schools as uniquely desirable, competition does not disappear simply because the admissions mechanism changes. It may move to housing decisions, preparatory activities, or alternative school choices.

A constructive response therefore has two levels. The first is fairer access rules, which require transparent implementation. The second is to make the everyday educational experience strong across schools, so that a child's prospects depend less on securing one coveted address and more on what they can actually learn wherever they enrol.

Six-school pilot in 2027: a chance to rethink teaching, not just class size

The September announcement about smaller classes is arguably more consequential for what happens after admission. MOE intends to test classroom redesign ideas in six schools from 2027 and explore where lower pupil numbers could make the most difference. It will also examine whether teachers can concentrate on two rather than all three of the following non-teaching duty areas: form teachership, co-curricular activities and committee work.

This is worth welcoming as an experiment, but it needs a precise interpretation. Smaller classes are not themselves a teaching method. A lesson delivered to 20 passive pupils can reproduce many of the same problems as a lesson delivered to 35. Conversely, thoughtfully structured discussion, targeted feedback and formative checks can make learning more responsive even before class sizes change.

The real educational opportunity is the time and attention teachers may gain. With manageable groups and fewer competing duties, teachers could notice misconceptions earlier, talk to quieter learners, provide feedback while it is still useful and adapt explanations to individual progress. But those benefits are not automatic; staffing, training, lesson design and implementation all influence the result.

How should the pilot be judged? Looking only at examination scores would miss its stated classroom purpose. Schools should also examine whether more pupils participate, whether teachers can act on diagnostic evidence, whether feedback becomes more timely, and whether students become better at explaining their own thinking. Teachers' workloads and well-being should be measured rather than assumed to improve.

For Learning Coach, this raises a practical question: what would a teacher do differently with an extra five minutes of meaningful attention per learner? More talk from the teacher is one option. A better option may be to ask the learner to identify a difficulty, test a strategy and explain the next step. That is where the idea of the teacher as coach becomes useful.

Beyond streaming: new certificates, flexible subject levels and the risk of new labels

Singapore's move toward Full Subject-Based Banding is another part of the same conversation about individual strengths. The 2027 Singapore-Cambridge Secondary Education Certificate, or SEC, will replace the previous separate N- and O-Level examination structures. Students will sit subjects at G1, G2 or G3 levels as applicable, and their certificate will reflect the subjects and levels they took. SEAB says the examination standards themselves are not being lowered by this change.

The distinction is important. A learner may need greater challenge in mathematics and more support in language learning, or vice versa. Treating academic ability as one fixed category can obscure these differences. A subject-based approach recognises that strengths are multidimensional and that development is not always linear.

Yet administrative flexibility will not eliminate the social meaning attached to levels. If a student internalises a G1, G2 or G3 placement as a judgment of personal worth, the new system could reproduce old assumptions in a new vocabulary. Schools must therefore communicate levels as current learning pathways, not permanent identities. Opportunities to review placement and support progress matter as much as the labels.

From 2028, MOE also plans to introduce a new post-secondary admissions exercise alongside the SEC results cycle. The promise of more coherent pathways will need to be tested against students' understanding of their choices, the quality of counselling and how smoothly transitions work in practice.

Artificial intelligence in Singapore schools: learning with a tool or delegating the learning?

Singapore is not treating AI literacy as a short-lived enrichment activity. Its 2026 policy announcements describe progression through the curriculum, co-curricular experiences and self-directed resources. The updated Code for Fun programme is expected to be available to all schools in 2027. Guidance on cyber wellness includes questioning information generated by AI and recognising deepfakes.

The government's AI-in-Education Framework highlights four principles: agency, inclusivity, fairness and safety. Those principles create a sensible test for any classroom AI tool. Does it help a student think, or make thinking unnecessary? Does it support pupils with different needs? Can teachers inspect errors and biases? Are privacy and age-appropriate safeguards in place?

MOE's guidance on generative AI also distinguishes useful support from premature dependence, noting the value of foundational concepts and thinking skills. That is an important corrective to the idea that more AI use is always better.

Consider an example in a secondary science lesson. One pupil asks an AI assistant to generate the answer to a question about heat transfer, then copies it. Another first proposes an explanation, asks AI for a counterexample, checks the response against a classroom experiment and rewrites the explanation in their own words. Both have used technology. Only the second has clearly shown evidence of reasoning.

A useful sequence is: think first, consult second, verify third, explain fourth. A learner should be able to say what the AI contributed, what they rejected, and why the final answer makes sense. Teachers can adapt this sequence by age and task, but the principle should remain: technology must not conceal a loss of understanding.

Equity also matters. Providing access to the same software does not guarantee equal benefit. Some children have adults who help them evaluate outputs; others may encounter the technology alone. Schools need explicit AI literacy instruction, accessible alternatives and assessment tasks that reveal what learners actually understand.

The overlooked connection: a fairer school is only valuable if learners gain agency

These changes can be read as separate policy announcements. They also expose one unifying challenge: the system can create more opportunities, but students must develop the capacity to use those opportunities well.

An inclusive admissions process is important because children deserve access. A flexible subject pathway matters because strengths vary. A smaller class can create more time for feedback. AI tools can broaden access to explanations and creative possibilities. But none of these guarantees that a student can plan their own learning, recognise confusion, persist through setbacks or judge the quality of information.

This is the gap between educational provision and learning capability. Provision includes school places, timetables, devices and curricula. Capability includes goal setting, metacognition, self-regulation, help-seeking, effective practice and thoughtful transfer of knowledge. The two dimensions complement rather than replace one another.

The Learning Coach perspective is therefore not that Singapore needs to abandon its examination culture overnight, or that coaching can resolve inequity by itself. It is that reforms will be stronger when adults intentionally teach learners how to learn, not only what they need to know.

What a coaching-based classroom can actually look like

A teacher who adopts a coaching posture does not stop teaching subject knowledge. Clear explanations, worked examples and appropriate practice remain essential. Coaching adds a complementary layer: learners become active participants in understanding their needs and deciding how to make progress.

Take a hypothetical Primary 5 pupil named Mei, who consistently struggles with multi-step mathematics problems. A conventional response might be more worksheets and reminders to concentrate. A coaching-based response begins by diagnosing the precise obstacle. Can Mei identify the relevant information? Does she understand the operations? Does she lose track of intermediate steps? Is she anxious about making a mistake?

After modelling a strategy, the teacher might ask: "Which step is secure? Where did your reasoning change? What would you try differently on the next problem?" Mei could use a checklist for identifying givens, selecting operations and checking whether an answer is plausible. The teacher then observes whether the process improves on a different task.

The objective is not to make Mei dependent on a permanent stream of motivational questions. It is to transfer responsibility gradually. Eventually she should recognise the obstacle and select a strategy without adult prompting.

This idea closely matches a coaching principle described in Learning Coach's training material: the coach accompanies a learner towards a clear objective while protecting their autonomy. Coaching should build on the learner's existing strengths and avoid creating dependence on the helper. It is an educational posture, not a substitute for subject teaching, special educational support or healthcare.

Five practical changes schools can test alongside the reforms

1. Start with a learner goal, not only a curriculum objective. A curriculum may specify what content must be covered. The learner should also know what improvement they are working towards: for example, explaining a scientific claim with evidence, monitoring reading comprehension or independently planning a revision session. Goals should be concrete and revisited.

2. Use formative assessment as a conversation. After a short quiz or task, ask pupils to identify the point of confusion and compare two possible strategies. Feedback becomes more useful when the student has a role in interpreting it. The question is not merely "What mark did you get?" but "What can you do differently now?"

3. Protect independent thinking in AI-supported tasks. Before opening a chatbot, ask students to write an initial explanation or proposed solution. Afterwards, require verification against trustworthy material and an account of what changed. This makes AI use more transparent and limits unthinking copying.

4. Teach recovery from mistakes. A mistake should lead to diagnosis, correction and another attempt. Teachers can model their own revisions, discuss why a plausible answer is wrong and normalise asking for help. High expectations and psychological safety need not be opposites.

5. Track autonomy, not just attainment. Alongside scores, teachers can periodically note whether pupils start tasks with less prompting, choose appropriate strategies, seek targeted help and transfer what they learned to unfamiliar situations. These are practical observational indicators, not a claim that one score can capture independence.

None of these approaches requires inventing a new national examination. They require time, thoughtful teacher development, collaboration with families and a willingness to look beyond the most visible performance indicators.

What parents in Singapore should do now

For families approaching the 2027 Primary 1 registration exercise, the first task is informational: check the official P1 registration guidance, confirm the child's eligibility and registration year, identify which phase applies and check the specific school's current rules. Do not confuse the 2027 registration exercise with 2027 school entry; the announced changes are for admission to Primary 1 in 2028.

The second task is developmental. School choice matters, but the everyday experience of learning also depends on sleep, relationships, curiosity, routines, emotional security and the ability to ask questions. Parents can invite children to talk about how they solved something, rather than only whether they were correct. They can let a child attempt age-appropriate tasks before intervening and praise the strategy as well as the outcome.

Parents should not assume that every school is identical or that selection concerns are imaginary. Instead, combine realistic research about schools with attention to the child in front of them. A prestigious name cannot replace a healthy relationship with learning.

For older children, ask a different kind of question when discussing subject levels, national examinations or post-secondary choices: "What are your strengths now, what would you like to develop, and what support would help you move forward?" That conversation recognises a growing person rather than reducing them to a score.

What Singapore should measure after 2027

By 2028 and beyond, the most useful evaluation would compare policy intentions with observable outcomes. These are questions policymakers, researchers and school leaders should ask—not results that are already established.

Admissions: Does the social mix change at the affected schools? Do access patterns improve without producing unreasonable new transport burdens? What happens to families who do not secure their first choice?
Six-school pilot: Do classroom interactions and feedback improve? Does teacher workload decline in practice? Are gains sustained and transferable to other schools with different circumstances?
AI literacy: Can pupils independently recognise unreliable outputs, protect personal information and justify their own reasoning? Are benefits distributed fairly?
Flexible pathways: Do students understand their subject-level options, feel less stigmatised and access suitable post-secondary choices? Do changes support progression without lowering meaningful standards?
Student agency: Can learners describe how they study, identify a barrier, choose a response and evaluate whether their strategy worked?

These measures need careful design and attention to context. A single before-and-after comparison, or a dramatic anecdote from one school, would not establish causation. Pilots should produce evidence before they become universal prescriptions.

What other education systems can learn from Singapore

International observers often reduce Singapore to academic rankings or exceptionally demanding examinations. The 2026 announcements reveal a more interesting story: even a high-performing system has reasons to adjust access, teaching conditions, assessment pathways and digital literacy. Excellence is not a final destination.

Other countries should resist copying individual measures without context. Singapore's admissions structure, staffing arrangements, curriculum and public institutions differ from those elsewhere. Importing an AI programme or a class-size policy without the necessary teacher development and evaluation would miss the larger lesson.

The transferable idea is more demanding: align access, high expectations, evidence-informed teaching and learner autonomy. A reform should not be judged only by how elegantly it is announced or how efficiently it is administered. It should be judged by the learning opportunities it genuinely creates.

Frequently asked questions

Will Primary 1 registration rules change in 2027?

Yes. The changes announced in September 2026 apply to the registration exercise in 2027 for children starting Primary 1 in 2028. Most schools will reserve more Phase 2C places; twelve schools will use a two-track arrangement. Families should consult the MOE registration site for the precise rules and dates.

Is Singapore reducing class sizes in every school?

No. The announcement concerns a six-school pilot starting in 2027 to explore classroom redesign and teacher duties. It would be inaccurate to describe this as an already implemented nationwide policy.

Will AI become compulsory for all pupils in 2027?

MOE says the updated Code for Fun programme will be available to all schools in 2027. That should not be misrepresented as a mandate for every child to use generative AI in every subject or assessment. Age, curriculum and responsible-use guidance remain essential.

Are the old N- and O-Level exams disappearing?

The new SEC replaces the former separate examination structures from 2027, with subjects taken at applicable G1, G2 or G3 levels. This does not mean that all academic distinctions or examination standards disappear.

Where does learning coaching fit into these changes?

A teacher-as-coach approach can complement explicit instruction through goal setting, diagnostic questioning, constructive feedback and increasing learner responsibility. The approach is not a proven solution to every policy challenge; it is one practical way to translate the ambition of greater student agency into classroom routines.

The real test is what pupils can do for themselves

Singapore's 2026 education reforms deserve to be taken seriously because they address different forms of opportunity: who gets into a school, how teachers use their time, how students move through subjects and how young people learn to work with AI. But reforms on paper do not automatically produce independent learners.

The decisive question is whether a child can leave a lesson more capable of thinking, choosing, checking and learning without constant adult direction. That is where structural reform and learning coaching can meet. Fairer access opens the door; thoughtful teaching and growing autonomy help students walk through it.

For teachers, school leaders and education professionals exploring this approach, Learning Coach's professional training pathway introduces the teacher-coach and other learning-support roles, with an emphasis on goals, ethical coaching posture, learning strategies and learner autonomy. Explore the programme and its current application requirements; applying is not the same as admission or certification.

Sources and reporting consulted

Singapore education reforms 2026

Faysal Hafidi