91 Countries, One Verdict: Is PISA 2025 Fair?

A Small Country With a Large Score and a New Test to Pass

Estonia carries a peculiar burden: a population smaller than many European cities is expected to model what the continent's best education looks like. The pressure was visible long before 10:30 AM on September 8, 2026, when OECD Director Andreas Schleicher and Minister Kristina Kallas stood before a Tallinn audience to release results that 91 countries had been waiting for. Estonia leads all EU member states in PISA scores — a fact that transforms every new cycle from a measurement into a verdict.

This cycle, however, the test itself changed. PISA 2025 does not simply ask what students know; it asks whether they are ready for futures that no previous generation navigated. The first of three published volumes carries the title "Future-Ready Students" — and that phrase functions less as a compliment than as a diagnostic question directed squarely at every education ministry in the ranking. Are your students self-directed when the algorithm is unreliable, the environment is destabilizing, and the teacher is not in the room?

For Estonia, this reframing is both an opportunity and an exposure. Approximately 760,000 students across 91 participating countries and economic regions sat the assessment, with 7,600 Estonian students born in 2009 representing the national cohort from 206 schools. Their results will either confirm that Estonia's educational advantage is structural and replicable, or reveal that high scores in previous cycles reflected a specific window of stability that is now closing. The release date is not an administrative calendar event. It is a stress test of a model the entire European Education Area has been quietly borrowing from.

What PISA 2025 Actually Measures — and Why the Scope Has Shifted

PISA has never been a neutral instrument. Each triennial cycle quietly renegotiates what counts as educated, and the 2025 iteration makes that renegotiation explicit. Science — designated as the primary domain for this cycle — moves beyond the traditional reading-mathematics axis, signaling that the OECD views scientific literacy as the foundational competency for navigating a technologically fractured world. The shift is deliberate, structural, and consequential.

The numbers behind this cycle give those claims geopolitical weight. Ninety-one countries and economic regions participated, a record high that transforms PISA from a predominantly Western benchmarking exercise into something closer to a genuinely global instrument of educational measurement. Approximately 760,000 students sat the assessment — a figure large enough to generate statistically robust cross-border correlation across vastly different school systems, from Singapore to Rwanda.

What the 2025 cycle tests has also expanded inward, toward student behavior. A new domain titled "Learning in the Digital World" measures whether students can direct their own problem-solving without institutional scaffolding. For the first time, data on student use of artificial intelligence in schoolwork was collected, acknowledging that the test environment can no longer be meaningfully separated from the tools students carry into it. The concept of "Anthropocene agency" — student capacity to reason and act on human-environment interdependence — further extends the assessment's scope into territory that no standardized test has systematically mapped before.

Estonia's cohort of 7,600 students, drawn from 206 schools, completed their assessments between March 31 and May 9, 2025. That narrow empirical window is now the factual foundation upon which every subsequent policy claim about European educational leadership will rest.

The Digital Autonomy Frontier: AI, Agency, and the Self-Regulation Gap

There is a hard distinction worth making here: the new "Learning in the Digital World" domain does not measure how much screen time a student accumulates. It measures whether a student can solve an unfamiliar problem independently, without a teacher standing over them and without an algorithm doing the thinking for them. That distinction — between consumption and agency — is the operative one for any policymaker reading these results.

The 2025 cycle introduces two genuinely new dimensions. First, "agency in the Anthropocene" evaluates how students reason through complex environmental decisions where there is no clean answer and multiple stakeholders hold competing claims. This is not an ecology quiz. It is a behavioral diagnostic of whether education systems are producing citizens capable of navigating systemic ambiguity. Second, and for the first time in PISA history, the study collected data on student use of AI in schoolwork, creating a baseline against which every future cycle will be measured.

Gunda Tire, Harno's coordinator for international assessments, has named the diagnostic challenge plainly: the results will reveal whether students are genuinely self-directed in digital environments or whether they surrender to what she frames as an "I can't be bothered" factor. The self-regulation gap — the distance between a student who uses AI as a thinking partner and one who uses it as a substitute for thought — may prove to be the defining educational fault line of this decade. Institutions that conflate digital access with digital competence will find the data unforgiving. The question for Estonian education leadership is whether the system's well-documented structural strengths are producing self-directed learners or merely well-drilled ones.

The self-regulation gap — the distance between a student who uses AI as a thinking partner and one who uses it as a substitute for thought — may prove to be the defining educational fault line of this decade.

Estonia's Structural Bets: 7,600 Students and a School Day That Starts at Nine

Seven thousand six hundred students, born in 2009, drawn from 206 schools across the country. That number carries more analytical weight than it first appears. The breadth of the cohort — spanning urban Tallinn gymnasia and rural consolidated schools alike — signals something deliberate about system-wide coverage, a sampling design built to expose cracks as much as strengths.

One of the structural interventions this cohort lived through is the 9:00 AM school start regulation, now a general requirement for Estonian general education schools. The policy intent is legible enough: align the school day with adolescent sleep biology, reduce chronic fatigue, and — implicitly — close the gap between student well-being rhetoric and the architecture of the actual school day. Whether it moved the needle for this specific cohort remains formally unmeasured. Testing ran from late March to early May 2025, precisely when the regulation was either newly in force or still embedding itself into school routines.

Then comes the rupture in the policy logic. Thirty-six schools in Tallinn applied for, and received, formal exceptions to start before 9:00 AM. This is not a marginal footnote. It reveals the persistent friction between national legislative intent and urban operational reality — transport corridors, extracurricular scheduling, parental working hours. The start-time debate is, in this sense, a proxy for a larger structural tension: who bears the cost of optimizing for student well-being when institutional convenience is competing for the same timetable slot? Estonia made the structural bet. The exceptions suggest the bet is still being negotiated.

The Global Hierarchy: Singapore, Rwanda, and Germany's Warning

A 667-point gap between Singapore and Cambodia should retire the phrase "global average" from serious educational policy discourse. Singapore's 1,679 total points against Cambodia's 1,012 do not describe a continuum; they describe two different institutional realities operating under the same assessment framework. Treating these as data points on a single spectrum produces policy conclusions that are, at best, meaningless and, at worst, actively misleading.

Germany's trajectory offers the more instructive warning for European audiences. A nation whose educational reputation was built on engineering precision and institutional rigor has recorded declining mathematics scores across recent PISA cycles. If Germany, with its resource base and federal investment capacity, cannot hold its position, then any EU member state resting on inherited credibility is already in a managed decline it has not yet diagnosed.

The more structurally significant signal, however, comes from Kigali. Rwanda is now benchmarking its 15-year-olds against students from Estonia and Germany, not against France or the United Kingdom. That is not a logistical choice; it is a declaration of reference points. Emerging economies are no longer measuring themselves against colonial-era educational powers. They are measuring themselves against systems that demonstrably work.

This reordering carries a direct implication for Estonia. The country is no longer simply a small nation punching above its weight; it has become a reference architecture for states that are choosing their futures deliberately. The question Estonian policymakers must now answer is whether the model can bear that weight, or whether external admiration will arrive precisely as internal pressures begin to compound.

What the PISA 2025 Results Demand of Estonia — and of Everyone Watching

Kristina Kallas framed the stakes precisely: these results show whether the Estonian education system has maintained its strength during a period of drastic social and technological change. That is a diagnostic question, not a victory lap. Gunda Tire's sharper formulation — PISA 2025 is a "moment of truth" — carries an implicit warning that accumulated reputation is not the same as structural durability.

The institutional pressure compounds at the European level. The European Commission treats PISA data as a tracking instrument for European Education Area 2025 goals, which transforms national scores into shared accountability. Estonia is no longer just competing against Singapore; it is carrying data obligations toward a continental policy framework that monitors convergence, inclusion, and resilience simultaneously.

The sustainability question is therefore not rhetorical. High performance achieved through 7,600 students across 206 schools is reproducible only if the conditions producing it — teacher quality, curriculum coherence, institutional trust — are actively maintained, not merely inherited. If the PISA 2025 score holds, it will not simply validate Estonia's model; it will raise the question of whether that model can survive the demographic pressures and AI-disrupted classrooms that every high-performing system must now reckon with.