OECD PISA 2025 · released 8 September 2026
Where 15-year-olds gained ground, and where they lost it
PISA 2025 tested about 760 000 students across 91 education systems. Each nationally covered system is shaded by how far its mean score moved — blue for a rise, red for a fall — against the previous cycle, or across the past decade. These are changes in system averages, not the progress of the same students. Score levels appear in the table below.
Science, change from PISA 2022 to PISA 2025
Select a system for details below. + / − to zoom; drag to pan when zoomed. Ctrl/⌘ + scroll to zoom.Choropleth world map. Every value shown here is also listed in the data table below.
Biggest gains
Ten largest increases
Biggest losses
Ten largest declines
Every participating system
PISA 2025 means and selected-period changes. Click a column to sort.Scroll sideways for all subjects. Select a column heading to sort; select it again to reverse the order.
How to read this
2022 → 2025 is the difference between the two most recent cycles. Decade trend is the OECD’s average decennial trend: a linear trend fitted across the PISA cycles from 2015 to 2025 and expressed as score points per decade, so it is not simply the 2015–2025 gap.
The OECD flags a change as statistically significant when it is distinguishable from zero at the 95% level. Systems whose change is not significant are hatched on the map and marked “NS” in the table — a non-significant estimate does not establish that performance was unchanged.
Regional samples use locator circles instead of shading their parent country: Beijing, Shanghai, Jiangsu and Zhejiang; 17 of Ukraine’s 27 regions; Dushanbe; and Iraqi Kurdistan. Select or focus a marker for coverage notes. Cyprus covers the area under the effective control of the Republic of Cyprus.
An asterisk marks a sampling caution. Albania’s trends are excluded under OECD reporting guidance. Viet Nam’s trends are unavailable after a change in test instruments. Uzbekistan’s reading and mathematics are withheld. Guatemala and Paraguay’s test-mode changes add uncertainty beyond the reported standard errors. See the Reader’s Guide, pp. 16–20.
Source. OECD (2026), PISA 2025 Results (Volume I): Future-Ready Students. Mean scores from Tables I.B1.2a.1–4; changes, trend coefficients and standard errors from Tables I.B1.2a.36–38. Means were cross-checked with Table I.1; chapter-table changes and significance flags were checked against the annex. Comparisons are included separately for each subject.
PISA 2025 Results (Volume I) · PISA 2025 Database
OECD annex workbook · Chapter workbook · Download data with cell references · Audit findings
OECD changes use OECD average-35, the published fixed comparison group, rather than subtracting changing-membership OECD averages. “NS” means not statistically significant at 5%. Short-term tests use published difference standard errors; decade tests use published p-values. Displayed confidence intervals are approximate 95% intervals (estimate ± 1.96 × SE). Rankings order point estimates, not statistically established differences between systems.
Country outlines: Natural Earth, 1:50m, via world-atlas. Equal Earth projection — an equal-area world, so no country is inflated by where it sits.
Verified against source files on 8 September 2026; figures rounded to one decimal place. Source snapshots and SHA-256 checksums accompany the build. OECD revisions require a new audit.
This is an adaptation of an original work by the OECD. The opinions expressed and arguments employed in this adaptation should not be reported as representing the official views of the OECD or of its Member countries.