Belgium scored 490 in science, 480 in mathematics and 466 in reading in PISA 2025. Science was broadly unchanged from 2022, while mathematics and reading moved lower. Across the longer comparable series, all three domains remain below Belgium’s earlier PISA levels.
Belgium’s PISA 2025 Results
Belgium’s national result combines students from the Flemish, French and German-speaking Communities.
Belgium performed above the OECD average in science and mathematics. Its reading mean was five score points above the OECD average numerically, but the difference was not statistically distinguishable. Computational problem solving, introduced as part of PISA 2025’s Learning in the Digital World assessment, was also above the OECD average.
Belgium and the OECD Average in PISA 2025
Reported mean scores and arithmetic differences from the corresponding OECD average.
| Domain | Belgium | OECD average | Difference | OECD comparison |
|---|---|---|---|---|
| Science | 490 | 482 | +8 | Above OECD average |
| Mathematics | 480 | 463 | +17 | Above OECD average |
| Reading | 466 | 461 | +5 | Not statistically distinguishable from OECD average |
| Computational problem solving | 508 | 500 | +8 | Above OECD average |
The difference column is calculated from the published rounded means. A numerical gap alone does not establish a statistically detectable difference.
Belgium’s PISA Score Trend Since 2000
Reading has a comparable trend series from 2000, mathematics from 2003 and science from 2006.
Belgium’s Three Core PISA Scores Have Moved Lower Over the Long Run
Mean scores for each comparable PISA cycle, 2000–2025
The vertical scale begins at 450 rather than zero so changes within the PISA score range remain readable. Reading, mathematics and science are separate scales and should not be combined into an overall PISA score.
- Mathematics fell from 529 in 2003 to 480 in 2025, a 49-point difference based on the published rounded means.
- Reading reached 509 in 2012 and was 466 in 2025, 43 points below that observed high.
- Science moved from 510 in 2006 to 490 in 2025. Its latest-cycle movement was much smaller than the declines in mathematics and reading.
Belgium PISA Scores by Cycle
Official OECD mean-score series used for long-term comparison.
Scroll horizontally to view all columns.
| PISA cycle | Mathematics | Reading | Science |
|---|---|---|---|
| 2000 | — | 507 | — |
| 2003 | 529 | 507 | — |
| 2006 | 520 | 501 | 510 |
| 2009 | 515 | 506 | 507 |
| 2012 | 515 | 509 | 505 |
| 2015 | 507 | 499 | 502 |
| 2018 | 508 | 493 | 499 |
| 2022 | 489 | 479 | 491 |
| 2025 | 480 | 466 | 490 |
Belgium participated in every listed cycle. The dashes mark years outside the OECD’s comparable long-term trend starting point for that domain: mathematics begins in 2003 and science in 2006.
The 2022–2025 Movement Was Uneven Across Subjects
OECD short-term change estimates use the underlying unrounded score estimates rather than subtraction of the displayed whole-number means.
Reading and Mathematics Fell Again; Science Was Broadly Stable
Official score-point change from PISA 2022 to PISA 2025
OECD reports the mathematics and reading decreases as statistically detectable. The one-point science change is not statistically distinguishable from no change.
The latest cycle did not extend the long-run decline equally across subjects. Science held close to its 2022 level, while reading and mathematics moved lower. That separation matters because a single description of Belgium’s overall PISA direction would hide different subject trajectories.
The displayed mathematics means are rounded to whole score points. OECD calculates cycle-to-cycle change from the underlying unrounded estimates, so the official 2022–2025 change can differ slightly from simple subtraction of the rounded values shown in a table.
Absolute Scores Fell, but Belgium Remained Above the OECD Average in Two Domains
Belgium’s lower scores do not automatically imply the same change in its position relative to the OECD benchmark. The OECD averages also moved lower between 2022 and 2025. In 2022, Belgium scored 489 in mathematics against an OECD average of 472, 479 in reading against 476, and 491 in science against 485.
Belgium’s mathematics gap over the OECD average was about the same in 2022 and 2025 when calculated from the published whole-number means: +17 points in both cycles. Science’s displayed gap widened from +6 to +8, while reading moved from +3 to +5. These descriptive gaps should not be treated as precise rankings or as evidence that every small difference is statistically detectable.
Proficiency in PISA 2025
Mean scores describe the center of a performance distribution. Proficiency shares show how students are distributed below the baseline and at the upper end of each scale.
Belgium’s 2025 Proficiency Shares
Rounded percentages reported in the OECD Belgium country note.
Scroll horizontally to view all columns.
| Domain | Below Level 2 | Level 2 or higher | Levels 5–6 | OECD Level 2+ share | OECD Levels 5–6 share |
|---|---|---|---|---|---|
| Science | 23% | 77% | 7% | 74% | 7% |
| Mathematics | 28% | 72% | 10% | 65% | 8% |
| Reading | 29% | 71% | 6% | 69% | 6% |
Below Level 2 and Level 2-or-higher shares are complementary rounded values. A system mean should not be interpreted as the proficiency level of an individual student.
Belgium’s mathematics mean remained above the OECD average while more than one-quarter of students were below Level 2. The same mean score can coexist with substantial variation across students, which is why proficiency shares add information that a country mean alone cannot provide.
The lower-performing share changed more in mathematics and reading than in science
Compared with PISA 2015, OECD reports that Belgium’s share of students below Level 2 increased by 8 percentage points in mathematics and 9 percentage points in reading. The corresponding science change was not statistically detectable. This distributional pattern is consistent with the stronger long-term movement visible in mathematics and reading.
Belgium’s PISA Participation and 2025 Coverage
Belgium has taken part in every regular PISA cycle since the program began in 2000. The normal three-year schedule was disrupted after 2018, with the next assessment administered in 2022 rather than 2021.
The school response rate in the Flemish Community did not fully meet the PISA standard. OECD conducted further analyses and judged the data fit for reporting. Belgium’s published national estimate combines its three education communities.
How to Read Belgium’s Long-Term PISA Series
- Reading trend
- The OECD comparable reading series begins with PISA 2000. Belgium has a reported reading mean for every regular cycle from 2000 through 2025.
- Mathematics trend
- The comparable mathematics trend begins in 2003, when mathematics was first established as the major assessment domain for the long-term scale used here.
- Science trend
- The comparable science trend begins in 2006. Earlier PISA participation does not extend the same science trend scale backward to 2000 or 2003.
- Mean score
- A mean score summarizes the estimated performance distribution for the assessed population. It is not a pass mark and does not describe the skills of every student.
- OECD average
- The OECD average is calculated for OECD members included in the relevant PISA calculation. It should not be interpreted as the average of every participating country and economy worldwide.
- Score-based comparisons
- Small differences in mean scores can fall within sampling uncertainty. PISA therefore does not support treating every one-point ordering as an exact ordinal rank.
Data Sources and Measurement
The principal reference for Belgium’s 2025 mean scores, short-term changes, proficiency shares, sample coverage and national interpretation. OECD: PISA 2025 Results (Volume I) — Belgium
International tables and OECD averages for science, mathematics, reading and computational problem solving. OECD: Student Performance in PISA 2025
OECD’s Belgium trend tables provide the pre-2025 sequence for reading, mathematics and science, including cycle-specific uncertainty. OECD: PISA 2022 Performance Trends
OECD’s interactive trend tool accompanies the PISA 2025 country note and allows cycle-by-cycle comparison of performance estimates. OECD PISA Performance Trends
The OECD participant record confirms Belgium’s participation in 2000, 2003, 2006, 2009, 2012, 2015, 2018, 2022 and 2025. OECD PISA Participant: Belgium
The historical table uses OECD’s comparable domain series: reading from 2000, mathematics from 2003 and science from 2006. Scores are rounded to whole points for display. Derived differences based on those rounded values can differ slightly from OECD change estimates calculated from unrounded underlying estimates.
PISA estimates population performance from samples of 15-year-old students rather than testing every student. Score comparisons therefore carry statistical uncertainty. The 2025 assessment placed science at the center of the assessment, while mathematics and reading were minor domains. Computational problem solving was reported separately and is not merged with the three established subject scales.