Average scores, recent changes, and long-term trends
International standing. Latin America remains far behind the OECD in all three subjects, even among the region's highest-performing countries.
Learning trends. Declines in reading and mathematics span both recent years and the past decade. Science shows a more favorable trend.
Differences across countries. Performance levels and trends do not always align. Some countries with low scores are improving, while others with higher scores are declining.
Latin America remains far below the OECD
Average scores in reading, mathematics, and science, 2025
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.1-I.B1.2a.3 and I.B2.1-I.B2.3. LAC average: simple average of the 13 participating countries.
The region lost ground in reading, but not in science
Change in average scores, 2022-2025
Reading
Mathematics
Science
Source: OECD, PISA 2025; Argentina's Ministry of Education, PISA 2022. LAC: simple average of the 12 countries with comparable results.
Better performance did not mean greater improvement
Average scores in 2025 and change since 2022
Reading
Mathematics
Science
Source: OECD, PISA 2025; Argentina's Ministry of Education, PISA 2022. LAC: simple average of 13 countries for performance and 12 for change.
The region declined in reading and mathematics
Change in average scores, 2015-2025
Reading
Mathematics
Science
Source: OECD, PISA 2015 and PISA 2025. LAC: simple average of the 8 countries with comparable results.
Countries followed very different trajectories
Average scores in 2025 and change since 2015
Reading
Mathematics
Science
Source: OECD, PISA 2015 and PISA 2025. LAC: simple average of the 8 countries with comparable results.
2Proficiency levels
Students at the lowest and highest levels and changes since 2022
Basic skills. Most students in the region do not reach the minimum level, especially in mathematics. Since 2022, this problem has worsened mainly in reading.
Advanced skills. Very few students reach the highest levels, even in the countries with the best results. Recent changes are small and show no expansion in high performance.
Most students do not reach the minimum proficiency level
Percentage of students below Level 2, 2025
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.10-I.B1.2a.12 and I.B2.9-I.B2.11. LAC: simple average of the 13 participating countries.
Low performance rose mainly in reading
Change in the percentage of students below Level 2, 2022-2025
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.33-I.B1.2a.35; CABA, Córdoba, and Mendoza: author's calculations using PISA 2022 and 2025 microdata. Changes in percentage points; LAC: simple average of the 12 comparable countries.
High performance remains exceptional
Percentage of students at Level 5 or 6, 2025
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.10-I.B1.2a.12 and I.B2.9-I.B2.11. LAC: simple average of the 13 participating countries.
High performance has barely changed since 2022
Change in the percentage of students at Level 5 or 6, 2022-2025
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.33-I.B1.2a.35; CABA, Córdoba, and Mendoza: author's calculations using PISA 2022 and 2025 microdata. Changes in percentage points; LAC: simple average of the 12 comparable countries.
3Performance gaps
Performance distribution, socioeconomic gaps, and resilient students
Inequality within countries. National averages coexist with wide differences among students. Recent trends vary by country and subject.
Socioeconomic gaps. Advantaged students perform better throughout the region. The narrowing gap in reading reflects a larger decline among these students.
Disadvantaged students with strong results. Few rank among the highest-performing students in their country. Their share increased mainly in reading, although this measure reflects a relative position.
Performance gaps remain wide
Average scores and scores at the lower and upper quartiles, 2025
StudentsAverage
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.1-I.B1.2a.3 and I.B2.1-I.B2.3. LAC: simple average of the 13 participating countries.
Performance gaps changed unevenly
Change in the gap between the lower and upper quartiles, 2022-2025
StudentsGap
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.39-I.B1.2a.41. CABA, Córdoba, and Mendoza: author's calculations using PISA 2022 and 2025 microdata; LAC: simple average of the 12 comparable countries.
Socioeconomic gaps span the entire region
Average scores by socioeconomic quartile, 2025
StudentsAverage
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2b.2-I.B1.2b.4 and I.B2.29-I.B2.31. LAC: simple average of the 12 countries with data; Costa Rica does not report results by socioeconomic quartile.
The socioeconomic gap fell in reading, not science
Change in the gap between socioeconomic quartiles, 2022-2025
StudentsGap
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2b.22-I.B1.2b.24. CABA, Córdoba, and Mendoza: author's calculations using PISA 2022 and 2025 microdata; LAC: simple average of the 11 comparable countries.
One in ten disadvantaged students is resilient
Percentage of the bottom socioeconomic quartile reaching the top performance quartile, 2025
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2b.2-I.B1.2b.4 and I.B2.29-I.B2.31. LAC: simple average of the 12 countries with data; Costa Rica does not report results.
Resilienceincreased mainly in reading
Change in the percentage of resilient students, 2022-2025
Reading
Mathematics
Science
Source: OECD, PISA 2022, Tables I.B1.4.3-I.B1.4.5, and PISA 2025, Tables I.B1.2b.2-I.B1.2b.4; CABA, Córdoba, and Mendoza: author's calculations using PISA 2022 and 2025 microdata. Changes in percentage points; LAC: simple average of the 11 comparable countries.
4Technology and learning
Access, use, distraction, school policies, and artificial intelligence
Access at home. Device availability still lags behind the OECD and is lower among the most disadvantaged students.
Conditions in schools. Resource shortages are concentrated in the most disadvantaged schools, and reported capacity to integrate technology has declined.
Use at school. Use for learning fell and use for leisure rose. Intensive use is associated with lower performance.
Use at home. Leisure predominates, and use for learning declined. Moderate use is associated with better results; social media use grew the most.
Distraction and regulation. Distraction is associated with less learning. Cellphone bans show no learning advantage, but are associated with less distraction.
Artificial intelligence. Use is common, with relatively small social gaps. Greater use is associated with more motivation and curiosity, but not always with better performance.
In LAC, few students have access to technology at home
Percentage of students with devices available at home, as reported by students, 2025
Computers
Laptops
Source: author's calculations using OECD PISA 2025 microdata. LAC: simple average of the 11 countries with data; Costa Rica and El Salvador did not administer these questions.
The most disadvantaged students have less access to technology
Percentage with devices at home by student socioeconomic status, 2025
Students
Computers
Laptops
Source: author's calculations using OECD PISA 2025 microdata. Lines connect the percentages of students in the bottom and top quartiles of the socioeconomic index in each country or jurisdiction. LAC: simple average of 11 countries; Costa Rica and El Salvador did not administer these questions. Lighter shades indicate differences that are not statistically significant.
Schools have few devices per student
Devices available for educational purposes per 10 students, as reported by principals, 2025
Computers
Tablets
Source: OECD, PISA 2025, Tables I.B1.4.1 and I.B1.4.4. LAC: simple average of the 13 participating countries. Public microdata do not allow exact estimates for subnational jurisdictions.
Principals say limited resources constrain technology use
Percentage of students in schools with insufficient resources or adequate Internet, as reported by principals, 2025
Resource shortages limit teaching
Internet speed is adequate
Source: OECD, PISA 2025, Tables I.B1.4.7, I.B1.4.14, I.B2.60, and I.B2.61. LAC: simple average of the 13 participating countries.
The most disadvantaged schools report more limitations
Percentage of students by school socioeconomic status, as reported by principals, 2025
Schools
Resource shortages limit teaching
Internet speed is inadequate
Source: OECD, PISA 2025, Table I.B1.4.7 and author's calculations using PISA 2025 microdata. Lines connect the percentages for schools in the bottom and top socioeconomic quartiles. LAC: simple average of countries with data (12 in the left panel and 11 in the right). Costa Rica lacks data by school socioeconomic status; El Salvador also lacks data for the right panel.
Many schools lack capacity to integrate technology
Percentage of students in schools whose principals report sufficient capacity, 2025
Teacher skills
Professional resources
Time to prepare lessons
Source: OECD, PISA 2025, Table I.B1.4.17; author's estimates for CABA, Córdoba, and Mendoza using PISA 2025 microdata. LAC: simple average of the same 12 countries included in the 2022-2025 comparison.
School capacity to integrate technology declined
Change in the percentage of students in schools whose principals report sufficient capacity, 2022-2025
Teacher skills
Professional resources
Time to prepare lessons
Source: OECD, PISA 2025, Table I.B1.4.17; author's estimates for CABA, Córdoba, and Mendoza using PISA 2022 and 2025 microdata. LAC: simple average of the 12 countries with comparable data. Ecuador lacks comparable data for 2022.
At school, technology is used for learning and leisure
Daily hours of digital device use at school, as reported by students, 2025
For learning at school
For leisure at school
Source: OECD, PISA 2025, Tables I.B1.4.21 and I.B2.62. LAC: simple average of the 13 participating countries.
Those who use technology mostperform worse
Average science scores by hours of use reported by students, 2025
OECDLAC
For learning at school
For leisure at school
Source: OECD, PISA 2025, Figures I.4.5-I.4.6 and Tables I.B1.4.21, I.B1.4.33, and I.B1.4.35. LAC: simple average of the 12 countries with data.
Use for learning fell and use for leisure rose
Change in the percentage using devices more than one hour per day, as reported by students, 2022-2025
For learning
For leisure
Source: OECD, PISA 2025, Figure I.4.3 and Table I.B1.4.30. CABA, Córdoba, and Mendoza: author's calculations using PISA 2022 and 2025 microdata. LAC: simple average of the 10 countries with comparable data.
At home, technology is used more for leisure
Daily hours of digital device use before and after school, as reported by students, 2025
For learning at home
For leisure at home
Source: OECD, PISA 2025, student microdata (ST326Q02JA and ST326Q05JA), weighted means. Weekdays. LAC: simple average of the 13 participating countries.
Moderate use at home is associated with better results
Average science scores by hours of use reported by students, 2025
OECDLAC
For learning at home
For leisure at home
Source: author's calculations using PISA 2025 microdata (ST326Q02JA and ST326Q05JA), use before and after school on weekdays. Weighted averages of the 10 science plausible values, without adjustment for socioeconomic status. LAC: simple average of 13 countries; OECD: 37 countries with data. Percentages below the axis refer to the OECD.
At home, use for learning fell in almost every country
Change in the percentage using devices more than one hour per day, as reported by students, 2022-2025
For learning at home
For leisure at home
Source: author's calculations using PISA 2022 and 2025 microdata (ST326Q02JA and ST326Q05JA), use before and after school on weekdays. Changes in percentage points; weighted estimates and standard errors using 80 replicate weights. LAC: simple average of the same 10 countries with data in both years; OECD: 36 countries.
Social media dominate use at home
Percentage using devices more than one hour on a weekday, as reported by students, 2025
Social media
Entertainment
Content creation
Source: OECD, PISA 2025, Table I.B1.4.37. CABA, Córdoba, and Mendoza: author's calculations using PISA 2025 microdata. LAC: simple average of Argentina, the Dominican Republic, and Uruguay, the only countries in the region with data.
Social media useincreased the most
Change in the percentage using devices more than one hour on a weekday, as reported by students, 2022-2025
Social media
Entertainment
Content creation
Source: author's calculations using PISA 2022 and 2025 microdata. LAC: simple average of Argentina, the Dominican Republic, and Uruguay, the only countries in the region with comparable data. OECD: simple average of the 21 countries with comparable data.
Distraction is common and linked to lower performance
Device distraction: prevalence, performance association adjusted for socioeconomic status, and change, 2022-2025
Distracted students
Distraction and performance
Change, 2022-2025
Source: OECD, PISA 2025, Tables I.B1.4.43-I.B1.4.44 and I.B2.64; author's calculations using PISA 2022 and 2025 microdata. In 2022, the question referred to mathematics; in 2025, to science. LAC: simple average of countries with data. Estimates in the middle panel adjust for student and school socioeconomic status.
Many schools ban cellphones, but their effect is uncertain
Cellphone bans: prevalence, performance association adjusted for socioeconomic status, and change, 2022-2025
Cellphone bans
Bans and performance
Change, 2022-2025
Source: OECD, PISA 2025, Tables I.B1.4.47-I.B1.4.50. LAC: simple average of countries with data. Lighter bars indicate changes or differences that are not statistically significant. Estimates in the middle panel adjust for student and school socioeconomic status.
Less use and regulation are linked to less distraction
Difference in the percentage of distracted students, as reported by students and principals, 2025
Lower vs. higher use
Cellphones: banned vs. allowed
Subject policies: yes vs. no
Source: author's calculations using PISA 2025 student and school microdata. Left panel: the 25% using devices least minus the 25% using them most. Other panels: schools with the policy minus schools without it. LAC: simple average of 13 countries in the first panel and 12 in the others. Lighter bars indicate differences that are not statistically significant.
Many students use AI for schoolwork
Percentage using AI chatbots for schoolwork, as reported by students, 2025
For drafting texts
For learning
Does not use AI
Source: OECD, PISA 2025, Figures I.4.11-I.4.12 and Tables I.B1.4.53, I.B1.4.55, and I.B2.63. LAC: simple average of the 13 participating countries.
Gaps in AI use are relatively small
Percentage by socioeconomic status reported by students, 2025
Students
For drafting texts
For learning
Does not use AI
Source: author's calculations using PISA 2025 student microdata. Lines connect percentages for the bottom and top quartiles of the socioeconomic index. LAC: simple average of the 11 countries with data; Costa Rica and El Salvador lack socioeconomic index information. Lighter lines and markers indicate differences that are not statistically significant.
Performance is lower among daily AI users
Average science scores by AI use reported by students, adjusted for socioeconomic status, 2025
OECDLAC
To summarize a text
To research a new topic
For drafting texts
For learning
Source: OECD, PISA 2025, Figure I.4.13 and Table I.B1.4.61. LAC: simple average of the 12 countries with data.
More AI use, greater motivation and curiosity
Difference by frequency of AI use reported by students, 2025
Motivation (pp.)
Help-seeking (index)
Curiosity (index)
Source: author's calculations using PISA 2025 student microdata. Difference between the top and bottom quartiles of the index of AI use frequency for schoolwork. Intrinsic motivation is the percentage who agree or strongly agree that they enjoy learning new things. Other panels show differences in indices. LAC: simple average of the 13 participating countries. Lighter bars indicate differences that are not statistically significant.
Performance trends since each country's first participation
Sustaining improvements. Gains accumulated since the first PISA assessments have gone through periods of stagnation and decline. Improvements are not always sustained.
Differences across subjects. A country can follow different trajectories depending on the subject. Recent declines in reading and mathematics coexist with more favorable science trends in several countries.
The region followed different trajectories
Average scores since the first PISA participation
Reading
Mathematics
Science
Source: OECD, PISA 2025, Tables I.B1.2a.36-I.B1.2a.38 and I.B2.1-I.B2.3; Argentina's Ministry of Education, PISA 2022. LAC: simple average of countries with data in each cycle. Results from 2017 are shown in 2018.