PISA Scores Comparison: Math, Reading, Science by Country

PISA compares how well 15-year-old students can apply knowledge and skills in mathematics, reading, and science to real-life problems. [OECD] The latest published core-domain results are from PISA 2025, released on September 8, 2026. More than 760,000 students in 91 countries and economies participated. B-S-J-Z (China) recorded the highest mean scores in mathematics at 612 and science at 597, while Singapore led reading at 535. [OECD]

PISA 2025 Scores by Country and Economy

The OECD averages in PISA 2025 were 463 points in mathematics, 461 in reading, and 482 in science. These are separate subject scales, not percentages. The table below reports official mean scores rounded to whole points for the 91 participating countries and economies. Uzbekistan has no published mathematics or reading mean in the OECD snapshot, so those cells are shown as data not available rather than estimated. [OECD]

Data: OECD PISA 2025 Database, Tables I.B1.2a.1, I.B1.2a.2, and I.B1.2a.3. [OECD] Rows follow the OECD snapshot order, which is sorted by science performance; this is not an official combined ranking. An asterisk (*) marks Albania, Canada, the Netherlands, New Zealand, Norway, and the United States, where OECD advises caution because one or more PISA sampling standards were not met. “Data not available” is used for Uzbekistan’s mathematics and reading means because the OECD snapshot reports those values as missing.

The three-domain table shows why one league-table position can conceal useful detail. A system can occupy a different place in mathematics, reading, and science, and nearby whole-number scores may not represent clear population differences. In PISA 2025, 18 countries and economies performed above the OECD average in all three core domains. [OECD]

Mathematics Has The Widest Top-To-Bottom Range

B-S-J-Z (China) recorded the highest mathematics mean in PISA 2025 at 612 points, followed by Singapore at 563 and Macao (China) at 549. Across all participating education systems, OECD reported a 301-point gap between the highest and lowest mathematics means. The full-participant gaps were 280 points in science and 234 in reading. [OECD]

The upper mathematics group remains concentrated among several East Asian education systems. Estonia is the highest-scoring European Union member in this table at 508, while Switzerland records 499 and the United Kingdom 488. The OECD mathematics average is 463. Equal rounded scores, such as 522 for Korea and Hong Kong (China), should be treated as tied at the displayed precision rather than forced into separate whole-number ranks.

Reading Changes The Order Of The Leading Systems

Reading produces a different order. Singapore records the highest mean at 535 points, followed by B-S-J-Z (China) at 527, Chinese Taipei at 508, and Japan at 503. Macao (China) and Korea both round to 501, while Ireland records 500 and Estonia 499. [OECD]

The reading results also show why rounded means are not exact rank boundaries. Macao (China) and Korea both display as 501, and several systems are separated by only a few points. OECD comparisons use the underlying estimates and their uncertainty, not only the displayed whole-number score. [OECD]

Science Was The Focal Domain In PISA 2025

Science was the main assessment domain in PISA 2025. B-S-J-Z (China) recorded 597 points, followed by Singapore at 560, Macao (China) at 541, Chinese Taipei at 540, and Japan at 538. Estonia, at 527, was the highest-scoring European Union member in the science table. [OECD]

Across all 91 participating systems, the distance between the highest and lowest science means was 280 points. Among OECD countries alone, the science range was 124 points. These ranges describe differences among education systems in the 2025 assessment; they should not be converted mechanically into percentages of curriculum mastered or into a single measure of school quality. [OECD]

There Is No Official Overall PISA Score

Mathematics, reading, and science scores should be kept as three separate measures. Performance across the domains is reported on separate scales. A higher numerical value in one domain than another does not, by itself, show that a student group is stronger in that subject. Adding the three scores also does not create an official PISA total. [OECD]

This distinction matters in country comparisons. It is valid to compare Japan’s mathematics score of 525 with another system’s mathematics score, or its reading score of 503 with another system’s reading score. It is not valid to conclude that Japan is “better at science than reading” simply because its science mean of 538 is numerically higher than its reading mean. The domains have separate reporting scales.

Mean Scores and Proficiency Levels Describe Different Parts Of Performance

A country mean summarizes the central level of performance, but it does not show how students are distributed across PISA proficiency levels. OECD treats Level 2 as the baseline proficiency level in each core domain. In PISA 2025, 65% of students across OECD countries reached Level 2 or higher in mathematics, 69% in reading, and 74% in science. [OECD]

The upper end of the distribution is smaller. Across OECD countries, about 8% of students reached Level 5 or 6 in mathematics, 6% reached the top levels in reading, and 7% did so in science. Mean scores and proficiency shares therefore answer different questions: the mean shows the overall position of a system, while the proficiency distribution shows how many students fall below the baseline or reach advanced levels. [OECD]

Small Score Gaps Do Not Always Create A Clear Rank Order

PISA scores are estimates based on samples of students. Each mean has a standard error, and OECD uses confidence intervals and statistical tests when comparing systems. A displayed difference of one, two, or three points can therefore be too small to support a clear ordering, even when a sorted table places one system above another. [OECD]

This is visible near the top of the 2025 tables. New Zealand and Australia both round to 509 in science; Korea and Hong Kong (China) both round to 522 in mathematics; Macao (China) and Korea both round to 501 in reading. Even where rounded scores differ by a point or two, the underlying uncertainty still matters.

What Changed From PISA 2022 To PISA 2025

For the OECD average used in the short-term comparison, mean performance changed by −9 points in mathematics, −14 points in reading, and −3 points in science from PISA 2022 to PISA 2025. OECD describes science performance as broadly stable over this period, while reading and mathematics moved lower. The short-term OECD comparison excludes Costa Rica, Luxembourg, and Spain because comparable change estimates were not available for those systems. [OECD]

Score trends and rank movement are not the same thing. A system can move upward or downward in a sorted table because its own score changed, because other systems changed, because the participant group changed, or through a mixture of these. For time comparisons, the score change for the same domain is more informative than a simple change in rank.

Historical PISA 2022 Scores by Country and Economy

PISA 2022 remains useful as the previous cycle for historical comparison. Mathematics was the focal domain in that assessment, and the OECD averages were 472 in mathematics, 476 in reading, and 485 in science. The table below preserves the earlier country and economy results so changes can be checked against the same published 2022 values. [OECD]

Data: OECD PISA 2022 Database, Tables I.B1.2.1, I.B1.2.2, and I.B1.2.3. [OECD] An asterisk (*) marks systems for which OECD advises care because one or more sampling standards were not met. The double asterisk on Viet Nam applies to reading: OECD advises extra care when comparing that reading estimate internationally because the linkage to the international reading scale was not strong enough.

What A PISA Score Represents

PISA is designed around the ability of 15-year-olds to use knowledge and skills in practical contexts. It does not test whether every participating education system teaches the same national curriculum. OECD describes the assessment as measuring how students use reading, mathematics, and science knowledge to handle real-life tasks and problems. [OECD]

• Mathematics: mathematical reasoning and the use of mathematical concepts, procedures, facts, and tools in varied situations.
• Reading: understanding, using, evaluating, reflecting on, and engaging with texts.
• Science: using scientific knowledge, evidence, and reasoning to explain phenomena, evaluate information, and interpret investigations or data.

The target population is age-based rather than grade-based. Fifteen-year-olds can be enrolled in different grades across education systems, so the common age range improves comparability. PISA should therefore be read as a measure of tested learning outcomes among the defined 15-year-old population, not as a direct measure of primary education, university quality, school funding, or every part of an education system.

Why Some Entries Are Countries, Economies, Or Regions

OECD reports PISA results for participating countries and economies, and some entries represent defined jurisdictions or regions rather than a whole sovereign state. PISA 2025 includes B-S-J-Z (China), Macao (China), Hong Kong (China), Chinese Taipei, Dushanbe (Tajikistan), Kurdistan Region (Iraq), and “Ukrainian regions (17 of 27)” as reported entities. B-S-J-Z refers to Beijing, Shanghai, Jiangsu, and Zhejiang. Keeping the OECD labels preserves the units used in the published dataset. [OECD]

PISA 2025 Also Includes Computational Problem Solving

The 2025 cycle assessed science, reading, mathematics, and computational problem solving. The OECD average in computational problem solving was 500 points. This measure is reported separately from the three core-domain scores and should not be added to them to form a combined PISA total. [OECD]

More than 760,000 students took part in PISA 2025, representing about 33 million 15-year-olds in the schools of the 91 participating countries and economies. Science was the focal domain, while reading and mathematics remained part of the core comparison. [OECD]

Questions About PISA Country Scores

Which Country Has The Highest PISA Score?

There is no official combined PISA score. In PISA 2025, B-S-J-Z (China) had the highest mean in mathematics at 612 and science at 597. Singapore had the highest reading mean at 535. [OECD]

What Is The OECD Average PISA Score?

For PISA 2025, OECD-country means were 463 in mathematics, 461 in reading, and 482 in science. The three values belong to separate subject scales and should not be averaged into an official OECD overall score. [OECD]

Is A PISA Score A Percentage?

No. A PISA score is a scaled score point estimate, not the percentage of questions answered correctly and not the percentage of a curriculum mastered. Proficiency levels add another form of interpretation by describing the types of tasks students at different points on each subject scale can typically handle. [OECD]

What Does Level 2 Mean In PISA?

OECD uses Level 2 as the baseline proficiency level in mathematics, reading, and science. The task descriptions differ by domain, so Level 2 is not one identical skill across all subjects. In PISA 2025, 65% of OECD students reached Level 2 or higher in mathematics, 69% in reading, and 74% in science. [OECD]

Can Two Countries With Similar Scores Have The Same Statistical Standing?

Yes. PISA means are sample estimates, so two nearby scores may not be statistically distinguishable. OECD publishes standard errors and confidence intervals for country and economy estimates. A one-place difference in a sorted ranking can therefore suggest more precision than the estimates support. [OECD]

When Were PISA 2025 Country Scores Released?

The first PISA 2025 results were released on September 8, 2026. They are now the newest published core-domain country and economy results, replacing PISA 2022 as the current reference cycle for mathematics, reading, and science comparisons. [OECD]