Sweden’s PISA record has three clear phases: a broad decline through 2012, a recovery through 2018, and a renewed fall in reading and mathematics by 2025. Science moved within a narrower band. In PISA 2025, all three Swedish mean scores were close to their respective OECD averages.
Sweden’s PISA 2025 scores
The latest cycle places Sweden close to the OECD average in science, reading and mathematics, while the movement since 2022 differs sharply by subject.
2025 subject results and change since 2022
Mean PISA score, OECD comparison and OECD-published change from the previous cycle.
Scroll horizontally to view all columns.
| Subject | Sweden 2025 | OECD average | Sweden minus OECD | 2022–2025 change | OECD change |
|---|---|---|---|---|---|
| Science | 485 | 482 | +3 | −8 | −3 |
| Reading | 466 | 461 | +5 | −21 | −14 |
| Mathematics | 464 | 463 | +1 | −18 | −9 |
OECD-published cycle changes use the underlying estimates rather than subtraction of rounded whole-number means. For science, for example, the displayed means round to 494 in 2022 and 485 in 2025, while the OECD reports the underlying change as −8 points.
The level and the direction tell different stories. Sweden was only 1–5 points above the corresponding OECD averages in 2025, but reading and mathematics fell more from 2022 than the OECD averages did. The 8-point science change was not statistically significant according to the OECD. The OECD also describes all three 2025 results as among the lowest observed for Sweden and reports each 2025 mean as statistically below its 2018 level.
A 25-year pattern of decline, recovery and reversal
Reading provides the longest comparable series, followed by mathematics and science. Sweden participated in every PISA cycle from 2000 through 2025.
Sweden’s PISA subject trends, 2000–2025
Reported mean scores across every comparable cycle for reading, mathematics and science.
The vertical scale begins at 440 rather than zero so movement within the PISA score range remains readable without using an unusually narrow axis. Mathematics begins in 2003 and science in 2006 because those are the standard long-term comparison baselines for those domains.
Sweden’s reported mean scores by PISA cycle
The table preserves the complete comparable subject series rather than interpolating missing early-domain values.
Scroll horizontally to view all columns.
| PISA cycle | Reading | Mathematics | Science |
|---|---|---|---|
| 2000 | 516 | — | — |
| 2003 | 514 | 509 | — |
| 2006 | 507 | 502 | 503 |
| 2009 | 497 | 494 | 495 |
| 2012 | 483 | 478 | 485 |
| 2015 | 500 | 494 | 493 |
| 2018 | 506 | 502 | 499 |
| 2022 | 487 | 482 | 494 |
| 2025 | 466 | 464 | 485 |
Reading has a comparable trend from 2000, mathematics from 2003 and science from 2006. An em dash marks a cycle before the subject’s long-term comparison baseline; it does not represent a zero score or non-participation.
The 2012–2018 recovery covered all three subjects: using the rounded published means, reading rose 23 points, mathematics 24 and science 14. The direction then reversed. By 2025, reading was 17 points below its 2012 mean, mathematics 14 points below, and science had returned to the same rounded mean of 485.
The 2012–2018 recovery did not hold through 2025
Comparing two adjacent phases separates the mid-2010s rebound from the later reversal.
Recovery versus reversal by subject
Score-point change calculated from the rounded means shown in the historical table.
These are descriptive differences calculated from rounded OECD-published mean scores. They are not substitutes for OECD tests of statistical uncertainty.
Reading and mathematics gave back more between 2018 and 2025 than they gained between 2012 and 2018. Science followed the same directional pattern but within a smaller range. OECD’s own 2025 country note describes the post-2018 trajectory as reversing the gains seen from 2012 to 2018.
How the subject trajectories differ
Reading: the largest long-run decline
Reading fell from 516 in 2000 to 466 in 2025, a 50-point difference across the full comparable series. The path was not continuous: the mean reached 483 in 2012, recovered to 506 in 2018, then fell to 487 in 2022 and 466 in 2025. The 2018–2025 difference is 40 points. For the latest cycle alone, the OECD reports a 21-point decline from 2022, compared with a 14-point decline in the OECD average.
Mathematics: recovery to 2018 followed by a lower 2025 mean
The comparable mathematics series starts at 509 in 2003 and ends at 464 in 2025, a 45-point difference. Mathematics reached 478 in 2012 and returned to 502 by 2018 before falling to 482 in 2022 and 464 in 2025. The 2018–2025 difference is 38 points. From 2022 to 2025, the OECD reports a decline of 18 points for Sweden, twice the 9-point OECD-average decline.
Science: a narrower range than reading or mathematics
Science moved from 503 in 2006 to 485 in 2025. Its rounded mean was 485 in both 2012 and 2025, with a rise to 499 in 2018 between those cycles. The OECD reports an 8-point change from 2022 to 2025 that was not statistically significant. Science was the major assessment domain in PISA 2025; reading and mathematics were minor domains.
Proficiency levels add a second view of performance
Mean scores summarize the full distribution. Proficiency shares show how many students reached the PISA baseline and how many reached the upper levels.
Level 2+ and top-performer shares in PISA 2025
Level 2 is the baseline proficiency threshold used here; top performers are students at Levels 5 or 6.
Scroll horizontally to view all columns.
| Subject | Sweden Level 2+ | OECD Level 2+ | Sweden Levels 5–6 | OECD Levels 5–6 | Change in Sweden below Level 2 since 2015 |
|---|---|---|---|---|---|
| Science | 74% | 74% | 9% | 7% | +4 percentage points |
| Mathematics | 66% | 65% | 8% | 8% | +13 percentage points |
| Reading | 69% | 69% | 8% | 6% | +13 percentage points |
The final column is the OECD-reported percentage-point change in the share scoring below Level 2 between PISA 2015 and PISA 2025.
The rise in low performers since 2015 was much larger in reading and mathematics than in science. At the same time, Sweden’s 2025 top-performer shares were equal to or above the corresponding OECD averages in all three subjects. A mean-score decline therefore does not imply that every part of the student distribution moved by the same amount.
The national mean hides wide differences within Sweden
Gender, score distribution and socio-economic status add context that cannot be read from the national mean alone.
Gender differences in PISA 2025
Mean scores for girls and boys, with the difference expressed as boys minus girls.
Scroll horizontally to view all columns.
| Subject | Girls | Boys | Boys minus girls | OECD average gap |
|---|---|---|---|---|
| Science | 491 | 479 | −12 | −1 |
| Reading | 487 | 447 | −41 | −30 |
| Mathematics | 463 | 465 | +2 | +13 |
A negative value means that girls had the higher mean score. Sweden’s 2-point mathematics gender difference was not statistically significant.
Reading has a particularly wide score spread
In reading, the distance between Sweden’s 90th- and 10th-percentile scores was 304 points in PISA 2025. The OECD also reports that this high-low gap widened between 2022 and 2025 in reading, while the corresponding gaps in mathematics and science did not change statistically. In science, Sweden’s 90–10 spread was 284 points, compared with 258 points across the OECD average.
Socio-economic differences remain visible in science
Students in Sweden’s top national quarter of socio-economic status scored 92 points above those in the bottom quarter in science; the OECD-average gap was 85 points. Socio-economic status accounted for 12% of within-country variation in Swedish science performance, the same rounded share reported for the OECD average. OECD Education GPS also reports declines from 2022 to 2025 among Sweden’s socio-economically advantaged students of 30 points in mathematics and 21 points in science, showing that, in mathematics and science, the latest weakening was not confined to the lower end of the socio-economic distribution.
Sweden scored 503 in computational problem solving, compared with an OECD average of 500. This measure belongs to PISA 2025’s Learning in the Digital World assessment and is not part of the historical reading, mathematics or science series, so it is kept outside the 2000–2025 trend chart.
Data basis and comparability
The latest scores, proficiency shares, socio-economic results, sample coverage and trend statements come from the OECD’s PISA 2025 Results, Volume I: Sweden country note and the related Volume I tables.
The historical values through 2022 are reported in the OECD’s PISA 2022 performance-trends chapter. The PISA 2025 result tables extend the comparable series through the latest cycle.
Mean scores, proficiency distributions, gender results, socio-economic differences and historical trend tables are cataloged in the OECD’s PISA 2025 Results, Volume I annex tables.
Sweden participated in PISA in 2000, 2003, 2006, 2009, 2012, 2015, 2018, 2022 and 2025, as listed on the OECD’s Sweden participant page. Additional 2025 distribution and subgroup indicators are reported in OECD Education GPS.
Sweden took part in every PISA cycle shown, but the standard long-term subject baselines differ: reading begins in 2000, mathematics in 2003 and science in 2006. Mean scores are estimates for the tested population of 15-year-old students and should not be read as an individual student’s proficiency level. Small differences between Sweden and the OECD average should also not be treated as exact ordinal rankings without accounting for sampling uncertainty.
In Sweden, 6,688 students in 206 schools completed PISA 2025. The sample represented about 115,600 15-year-olds, or an estimated 92% of the country’s total population of 15-year-olds. Science was the major domain in 2025.