Australia remained above the OECD average in science, mathematics and reading in PISA 2025, but the long-term record is uneven. Mathematics and reading reached their lowest reported Australian means, while science stayed close to its 2015 level after a much flatter decade.
Australia’s PISA 2025 scores
The latest cycle separates current level from longer-run direction.
Australia was above the 2025 OECD average by 27 points in science, 15 in mathematics and 30 in reading, calculated from the published rounded means. That relative position does not indicate improvement over time: mathematics and reading fell again from 2022, while science changed little.
The long-term record splits into three different subject stories
Reading has a comparable trend from 2000, mathematics from 2003 and science from 2006.
Australia PISA mean scores across comparable cycles
The latest reading and mathematics results extend long declines; science has been much flatter since 2015.
The vertical scale starts at 450 to make cycle-to-cycle movement readable; it does not imply a zero-score baseline. Blank early points mark years before the official long-term comparison baseline for that domain.
Australia PISA scores by cycle, 2000–2025
Published mean scores for every comparable core-domain observation used in the trend.
Scroll horizontally to view all columns.
| PISA cycle | Reading | Mathematics | Science |
|---|---|---|---|
| 2000 | 528 | — | — |
| 2003 | 525 | 524 | — |
| 2006 | 513 | 520 | 527 |
| 2009 | 515 | 514 | 527 |
| 2012 | 512 | 504 | 521 |
| 2015 | 503 | 494 | 510 |
| 2018 | 503 | 491 | 503 |
| 2022 | 498 | 487 | 507 |
| 2025 | 491 | 478 | 509 |
An em dash means that the domain had not yet reached its official long-term trend baseline: reading was established in 2000, mathematics in 2003 and science in 2006.
- Reading fell 37 points from 528 in 2000 to 491 in 2025. The path was not continuous: there was a small rise in 2009 and no rounded-score change between 2015 and 2018.
- Mathematics shows the largest first-to-latest decline: 524 in 2003 to 478 in 2025, a 46-point difference from the published rounded means.
- Science fell from 527 in 2006 to 509 in 2025, but the recent pattern differs from the other domains: 510 in 2015, 503 in 2018, 507 in 2022 and 509 in 2025.
Mathematics and reading reached new lows; science did not
Australia’s mathematics decline was already established well before the 2022 assessment. The mean moved from 524 in 2003 to 504 in 2012 and 491 in 2018, before reaching 478 in 2025. Reading also weakened over a long period, from 528 in 2000 to 503 in both 2015 and 2018, then to 498 in 2022 and 491 in 2025.
Science follows a different trajectory. Its earlier high was 527 in both 2006 and 2009, but the series has remained close to the 500–510 range since 2015. OECD reporting for Australia describes science performance as stable for more than a decade and close in 2025 to the 2015 result.
PISA scales are anchored to the first cycle in which each domain was a major focus: reading in 2000, mathematics in 2003 and science in 2006. A 2000–2025 page can therefore cover Australia’s full PISA participation history without treating mathematics or science as having comparable trend scores before their official baselines.
2025 proficiency shows where the mean-score pattern sits in the student distribution
Level 2 is the standard baseline used for the core PISA domains; Levels 5 and 6 identify top performers.
Scroll horizontally to view all columns.
| Domain | Australia Level 2+ | Australia below Level 2 | Australia Levels 5–6 | OECD Level 2+ | OECD Levels 5–6 |
|---|---|---|---|---|---|
| Science | 82% | 18% | 12% | 74% | 7% |
| Mathematics | 71% | 29% | 10% | 65% | 8% |
| Reading | 77% | 23% | 11% | 69% | 6% |
Below-Level-2 shares are the complements of the rounded Level-2-or-higher shares shown here. Percentage totals may reflect rounding.
Mathematics had the largest low-performing share of the three Australian core domains in 2025. OECD trend reporting also shows that, compared with 2015, the share below Level 2 rose by 7 percentage points in mathematics and 5 points in reading, while science showed no statistically detectable change in that low-performing share.
The 2025 gaps are strongly domain-specific
| Domain | Boys minus girls | 2025 pattern |
|---|---|---|
| Science | +1 point | The mean gender difference was not statistically distinguishable from zero. |
| Mathematics | +16 points | Boys had the higher mean. |
| Reading | −31 points | Girls had the higher mean. |
Socio-economic background also remains associated with science performance. Students in Australia’s top national quarter of the PISA economic, social and cultural status index scored 80 points above those in the bottom quarter in science; the OECD comparison gap was about 85 points. Socio-economic status accounted for 9% of the variation in Australian science performance, compared with 12% across OECD countries.
The 2015–2025 science gap between Australia’s advantaged and disadvantaged quarters narrowed. OECD reporting links that narrowing to lower performance among advantaged students while performance among disadvantaged students remained stable, rather than to a broad rise at the lower end.
PISA 2025 added a separate digital problem-solving measure
Computational problem solving was reported through PISA 2025’s Learning in the Digital World assessment. It is a separate scale from mathematics, reading and science and does not have a 2000–2025 Australian trend series, so it should not be inserted into the historical core-domain chart as if it were a fourth long-running subject.
Source years, measurement and comparability
The 2025 mean scores, proficiency shares, socio-economic results and cycle comparisons come from the OECD country note and PISA 2025 Volume I tables. Australia’s 2025 sample included 14,658 students in 766 schools, representing about 275,400 15-year-olds. The OECD reports that Australia met PISA 2025 data-quality standards.
Australia participated in every PISA cycle from 2000 through 2025. OECD trend tables and the performance-trends dashboard provide the historical means used to compare the core domains across valid cycles.
OECD PISA participant: Australia
OECD PISA performance trends
The OECD sets long-term comparison baselines at 2000 for reading, 2003 for mathematics and 2006 for science. Official cycle-change figures can differ by one point from subtraction of displayed rounded means because OECD comparisons use underlying estimates before rounding.
PISA 2022 was delayed from 2021. Australia’s student response rate fell to 76%, below the technical target, and the OECD stated that a small residual upward bias could not be ruled out. The 2022 point should therefore be read with extra caution. That issue does not create the earlier decline visible in reading and mathematics, which was already present across multiple pre-2022 cycles.