U.S. Student Performance in PISA 2025: Math, Reading, Science & Problem Solving
PISA 2025 reports average U.S. scores of 502 in science, 490 in reading, 463 in mathematics, and 513 in computational problem solving. The United States scored above the OECD average in science, reading, and computational problem solving and close to the OECD average in mathematics.
The results describe nationally weighted performance among 15-year-old students. Science was the major assessment domain in 2025; reading and mathematics were minor domains. Computational problem solving was assessed as part of Learning in the Digital World.
PISA 2025 U.S. Score Snapshot
Mean scores are reported on the PISA scale. OECD comparison values are the 2025 OECD averages published with the same assessment cycle.
U.S. and OECD Average PISA Scores, 2025
Science, reading, mathematics, and computational problem-solving mean scores.
The chart uses a limited score range so differences remain visible; exact values are listed in the table. PISA score points are not percentages.
PISA 2025 Results by Subject
The 2022–2025 change column uses OECD-reported changes calculated from underlying unrounded estimates. A reported change can differ slightly from subtraction of the displayed rounded means. Computational problem solving has no comparable 2022 PISA score.
| Subject | U.S. 2025 | OECD 2025 | Difference | U.S. 2022 | 2022–25 Change | OECD Comparison |
|---|---|---|---|---|---|---|
| Science | 502 | 482 | +20 | 499 | +2 | Higher than OECD average |
| Reading | 490 | 461 | +29 | 504 | −14 | Higher than OECD average |
| Mathematics | 463 | 463 | 0 | 465 | −2 | Close to OECD average |
| Computational problem solving | 513 | 500 | +13 | Not assessed on this scale | Not available | Higher than OECD average |
The computational problem-solving scale was introduced in PISA 2025. OECD released the score on September 8, 2026, while its analyses of students’ self-directed learning in digital settings are scheduled for 2027.
Proficiency and U.S. Performance Over Time
Proficiency shares add information that national mean scores do not show, while the historical series places the 2025 results in prior PISA cycles.
Baseline and Top-Performer Shares
Level 2 is the OECD baseline proficiency threshold in science, reading, and mathematics. Top performers reached Level 5 or 6. Computational problem solving uses Level 3 as its baseline and is not included in this three-subject table. Percentages are rounded to whole numbers as reported in the OECD U.S. country note.
| Subject | U.S. Below L2 | U.S. L2+ | OECD L2+ | U.S. Top | OECD Top |
|---|---|---|---|---|---|
| Science | 22% | 78% | 74% | 13% | 7% |
| Reading | 26% | 74% | 69% | 13% | 6% |
| Mathematics | 35% | 65% | 65% | 8% | 8% |
Below-Level-2 shares are the complements of the rounded Level-2-or-higher percentages, so totals may reflect rounding.
In science, about 65% of U.S. students performed at Levels 2–4, calculated from the rounded 78% at Level 2 or higher minus the 13% at Levels 5–6 (about 67% across OECD countries). Level 2 includes explaining familiar scientific phenomena and evaluating simple evidence. At Levels 5–6, students can apply scientific knowledge to unfamiliar situations.
U.S. PISA Scores Across Assessment Cycles
Comparable U.S. mean scores in science, reading, and mathematics through 2025.
Comparable mathematics trend values begin in 2003 and science values in 2006. U.S. reading results for 2006 were not reported because of a test-booklet printing error.
Historical U.S. PISA Score Table
The series uses OECD trend values. “Not comparable” marks cycles before the current subject scale can be used for the historical series; “Not reported” marks the 2006 U.S. reading exception.
| PISA Cycle | Science | Reading | Mathematics |
|---|---|---|---|
| 2000 | Not comparable | 504 | Not comparable |
| 2003 | Not comparable | 495 | 483 |
| 2006 | 489 | Not reported | 474 |
| 2009 | 502 | 500 | 487 |
| 2012 | 497 | 498 | 481 |
| 2015 | 496 | 497 | 470 |
| 2018 | 502 | 505 | 478 |
| 2022 | 499 | 504 | 465 |
| 2025 | 502 | 490 | 463 |
What the 2025 Score Pattern Shows
The score, proficiency, and trend measures answer different parts of the performance question and should be read together.
Reading moved lower in 2025
U.S. reading was 14 points lower than in 2022, while the 2025 mean remained above the OECD average. The OECD average reading change from 2022 to 2025 was also −14 points.
Science remained comparatively stable
The U.S. science mean was 502 in 2025, the same rounded value reported in 2018 and close to 499 in 2022. OECD trend reporting treats the recent U.S. science movement as broadly stable.
Mathematics stayed close to the OECD average
The U.S. and OECD averages both round to 463 in 2025. The Level-2-or-higher share was also 65% for both the United States and the OECD average.
A numeric ordering is not an exact country rank
PISA mean estimates carry sampling uncertainty. Education systems with different displayed mean scores can still have results that are not statistically distinguishable, so pairwise OECD comparisons are more precise than a simple ordered rank.
PISA and NAEP use different measurement designs
PISA compares 15-year-old students across education systems and emphasizes application of knowledge. NAEP’s national age-based long-term trends report reading and mathematics results for ages 9 and 13. The score scales should not be converted into one another.
Data Sources and Measurement
OECD, PISA 2025 Results Volume I, United States country note.
Official OECD U.S. country note
OECD, PISA 2025 Results Volume I, student performance tables and comparison notes, including computational problem solving.
Official OECD performance chapter
National Center for Education Statistics, PISA 2025 Methodology and Technical Notes.
Official NCES technical notes
OECD PISA trend tables for the United States, extended with the published 2025 results.
Official OECD historical trend tables
Science was the major domain in PISA 2025. Reading and mathematics were minor domains; computational problem solving was part of Learning in the Digital World. Results for these four areas were released on September 8, 2026. Analyses of self-directed learning are scheduled for 2027. PISA scores are scaled estimates of student proficiency, not percentages correct.
4,625 students in 144 schools participated. Domain counts were 3,583 for science, 1,900 for reading, and 1,831 for mathematics because students received different assessment forms.
Level 2 is the baseline proficiency threshold for science, reading and mathematics; the computational problem-solving baseline is Level 3. The displayed core-subject top-performer shares refer to Levels 5–6. Percentages are rounded to whole numbers from the U.S. country note.
Mean scores are displayed as whole score points. OECD change estimates use underlying unrounded values, which explains why the reported science change of +2 does not equal 502 minus 499.
U.S. weighted school participation was 45.0% before replacement and 53.9% after replacement; weighted student participation was 76%, and the overall student exclusion rate was 6.9%. These values missed PISA targets of 85% for schools, 80% for students, and no more than 5% exclusions. OECD therefore flags the U.S. results and states that bias cannot be ruled out or its direction determined.
More than 40% of responding U.S. students were not administered the student questionnaire. Questionnaire-derived results on student background, attitudes, school climate, and related context measures are therefore not used here.
The OECD average is the comparison value published for participating OECD countries in the same PISA cycle. PISA results describe the selected indicators and do not provide a complete measure of school quality, curriculum quality, or the performance of individual schools.