U.S. PISA 2025 Results
Official PISA 2025 results for U.S. 15-year-old students, covering science, reading and mathematics scores, OECD comparisons, change from 2022, proficiency distributions, the environmental science subscale, and digitalisation and AI context.
U.S. PISA 2025 Scores
U.S. Scores Compared with the OECD Average
Mean PISA score by core domain, 2025
PISA reports each subject on its own achievement scale. Cross-subject scores should not be combined into a single average.
OECD PISA 2025 Results, Volume I.
PISA 2025 U.S. Results by Subject
| OECD Comparison | ||||
|---|---|---|---|---|
| Science | 502 | 482 | +20 | Higher than the OECD average |
| Reading | 490 | 461 | +29 | Higher than the OECD average |
| Mathematics | 463 | 463 | 0 | Not measurably different from the OECD average |
U.S. Performance Was Uneven Across Subjects
- The U.S. comparison with the OECD average was stronger in science and reading than in mathematics. U.S. students scored above the OECD average in science and reading, while mathematics performance was statistically similar to the OECD average.
- The three domains describe different parts of the U.S. performance profile. Science was the major domain in 2025, while reading and mathematics were minor domains, so the separate scores should not be treated as components of one combined achievement measure.
- The 2025 cross-country comparison and the change over time answer different questions. Reading remained above the OECD average in 2025 even though U.S. reading performance declined from the previous PISA cycle.
Change from PISA 2022 to PISA 2025
U.S. mean scores across the two most recent cycles
Rounded score differences are shown in the table. Formal cycle comparisons account for sampling error and PISA linking error.
OECD PISA 2025 and NCES PISA 2022 results.
| Cycle Comparison | ||||
|---|---|---|---|---|
| Science | 499 | 502 | +3 | No measurable change |
| Reading | 504 | 490 | −14 | Lower in 2025 |
| Mathematics | 465 | 463 | −2 | No measurable change |
Reading Was the Only Core Domain with a Measurable 2022–2025 Decline
- Science and mathematics moved only slightly between the two cycles, and the official comparisons do not identify those changes as measurable.
- Reading followed a different pattern. The decline from 2022 occurred even though the 2025 U.S. reading score remained above the OECD average for the same cycle.
- OECD also reports that U.S. reading and mathematics results in 2025 were below their 2018 levels, while science has shown comparatively stable performance across PISA cycles.
Proficiency Levels in Science, Mathematics and Reading
PISA Level 2 is the baseline proficiency threshold used in the core domains. Levels 5 and 6 identify top performers.
| Science | 78% | 74% | 13% | 7% |
| Mathematics | 65% | 65% | 8% | 8% |
| Reading | 74% | 69% | 13% | 6% |
Baseline and Top-Performer Shares Show Different Parts of the Distribution
- Science and reading combined relatively high baseline attainment with larger U.S. top-performer shares than the corresponding OECD averages.
- Mathematics had a different distribution. The U.S. shares reaching Level 2 and Levels 5–6 matched the OECD averages after rounding.
- The Level 2 threshold and the Level 5–6 share measure different portions of the score distribution. A higher top-performer share does not by itself indicate that fewer students are below the baseline threshold.
U.S. PISA 2025 Environmental Science Performance
PISA 2025 introduced an environmental science scale built from a subset of science assessment items. It reports performance within environmental science and should remain separate from the overall PISA science score.
| Measurement Scope | ||||
|---|---|---|---|---|
| Environmental science | 500 | 480 | +20 | Science subscale based on selected environmental science items |
| Overall science | 502 | 482 | +20 | Full PISA 2025 science domain |
Environmental Science Proficiency
Level 2 is the baseline on the environmental science scale. U.S. students below Level 2 accounted for 21.7% of the assessed population, leaving 78.3% at Level 2 or higher. OECD reporting places the corresponding OECD averages at 26% below Level 2 and about 74% at Level 2 or higher.
OECD PISA 2025 Database, Table I.B1.5.2
The Environmental Science Result Adds a Separate View of Science Performance
- The U.S. environmental science mean was 500, compared with an OECD average of 480. The U.S. value was therefore numerically above the OECD average on this subscale.
- The environmental science score and the overall science score answer related but different questions. Environmental science uses selected science items tied to environmental and ecological content, while the overall science result uses the broader science assessment.
- The U.S. environmental science result appears in the international table between Finland and Hong Kong (China) when systems are ordered by mean score. An exact ordinal rank is not used because score uncertainty and the U.S. sampling caution matter when comparing nearby systems.
Questionnaire coverage: The environmental science mean and proficiency results come from cognitive assessment data. Environmental awareness, environmental attitudes, reported activities and similar indicators rely on questionnaire responses. More than 40% of responding U.S. students were not administered the student questionnaire in PISA 2025, so those questionnaire-based U.S. indicators require an additional coverage caution and are not included here.
Digitalisation and AI Context
PISA 2025 also reported how students described AI use, digital-device conditions and school rules. These questionnaire indicators provide context for the learning environment and are separate from achievement scores.
| Used AI at least weekly to help them learn | 37.3% | 45.5% |
| Learned to assess AI-generated information in school lessons | 56.3% | 62.6% |
| Reported no distraction from digital devices | 34.1% | 39.2% |
| Attended schools where cell phones were not allowed | 31.3% | 49.5% |
| Attended schools with subject-specific guidelines on digital devices | 89.3% | 71.2% |
U.S. Digital Learning Conditions Differed Across the Five Indicators
- The U.S. shares were lower than the OECD averages for weekly AI use for learning, learning to assess AI-generated information, reporting no digital distraction and attending schools where cell phones were not allowed.
- The U.S. share was higher for students attending schools with subject-specific guidelines on digital-device use.
- These percentages describe reported learning conditions and school practices. They should not be interpreted as PISA performance measures.
U.S. questionnaire coverage caution
These five indicators rely on questionnaire responses. More than 40% of responding U.S. students were not administered the student questionnaire in PISA 2025, and OECD calls for particular caution when interpreting U.S. questionnaire results because of this coverage issue. The percentages should therefore be used as contextual indicators with that limitation in view.
Learning in the Digital World
PISA 2025 also included the new Learning in the Digital World assessment. U.S. students recorded a mean computational problem-solving score of 513, compared with an OECD average of 500. This domain uses a separate scale and should not be treated as a fourth core PISA subject.
Additional PISA 2025 Results Still Forthcoming
The September 8, 2026 release covered science, mathematics, reading and computational problem solving. OECD states that additional reporting on self-directed learning in the digital environment, also described in its release schedule as self-regulated learning, will be published in 2027. Because the United States participated in Learning in the Digital World, additional U.S.-relevant PISA 2025 analysis remains forthcoming.
OECD also plans 2027 results for the optional English as a foreign language assessment. NCES indicates that the United States did not participate in that optional assessment, so that release is not expected to produce a U.S. foreign-language score.
Data Sources, Measurement and Comparability
Official 2025 results
The initial PISA 2025 results were released on September 8, 2026. Science was the major PISA domain, with mathematics and reading assessed as minor core domains.
OECD PISA 2025 Results: United States
OECD PISA 2025 Results, Volume I
U.S. methodology
PISA targets students aged 15 years and 3 months through 16 years and 2 months who are enrolled in grade 7 or higher. U.S. testing was conducted from April 9 through June 2, 2025.
Previous-cycle comparison
The 2022 U.S. mean scores used for the two-cycle comparison are 499 in science, 504 in reading and 465 in mathematics. Cycle comparisons require more than subtracting rounded scores because PISA incorporates sampling uncertainty and linking error.
Environmental science subscale
The environmental science scale uses a subset of PISA 2025 science items related to environmental and ecological content. OECD PISA 2025 Database Table I.B1.5.1 reports mean performance. Table I.B1.5.2 reports the environmental science proficiency distribution, including the below-Level-2 and Level-2-or-higher shares. Related tables I.B1.5.3 through I.B1.5.5 provide gender and student- or school-characteristic breakdowns.
Digitalisation and AI questionnaire indicators
The digitalisation and AI comparison uses the U.S. and OECD-average percentages reported in the PISA 2025 executive-summary snapshot. These measures come from questionnaire data rather than the cognitive assessment, so the U.S. questionnaire coverage limitation applies.
Score and proficiency measurement
PISA science, reading and mathematics results are reported on separate 0–1,000 scales using item-response models and plausible values. Level 2 is the baseline threshold for the three core domains. The environmental science subscale has its own seven-level proficiency distribution, with Level 2 used as its baseline.
U.S. 2025 sampling caution
The United States did not meet every PISA 2025 sampling standard. The weighted school participation rate after replacement was 53.9%, the weighted student participation rate was 75.7%, and the overall student exclusion rate was 6.87%. OECD states that school non-participation and student non-response could affect the U.S. estimates and that the direction of any resulting bias cannot be determined from the available information.
The U.S. assessment included 4,625 participating students in 144 schools. More than 40% of responding U.S. students were not administered the student questionnaire, limiting reporting based on questionnaire variables. Core achievement results therefore require the sampling caution above, while many questionnaire-based subgroup analyses face an additional coverage limitation.