TOEFL Score Guide (2025)
How the 0–120 scale works, section-by-section scoring methods, score percentiles, and university requirements — everything you need to understand your TOEFL score.
Last updated: 2025 · 12 min read
How TOEFL Is Scored (0–120 Total)
The TOEFL iBT is scored on a scale of 0–120. This total is the sum of four section scores, each on a 0–30 scale: Reading, Listening, Speaking, and Writing. There are no penalties for wrong answers — your raw score is the number of correct responses (or rubric points), converted to a scaled score.
Overall score benchmarks
ETS provides score recipients with both your scaled scores and performance descriptors that explain what each score range indicates about your academic English ability.
Reading Section Scoring (0–30)
The Reading section consists of 2 passages with 10 questions each — 20 questions total. Most questions are worth 1 point. The exception is the Prose Summary question at the end of each passage, which is worth up to 2 points (you earn 1 point for selecting 2 of the 3 correct statements, and 2 points for all 3 correct). This gives a maximum raw score of 22 points, which ETS converts to the 0–30 scale.
Raw score to scaled score conversion (approximate)
| Raw Score | Scaled Score (0–30) | Approximate % Correct |
|---|---|---|
| 22 | 30 | 100% |
| 21 | 29 | 95% |
| 20 | 28 | 91% |
| 19 | 26 | 86% |
| 18 | 24 | 82% |
| 17 | 23 | 77% |
| 16 | 21 | 73% |
| 15 | 20 | 68% |
| 14 | 18 | 64% |
| 12 | 15 | 55% |
| 10 | 12 | 45% |
| 8 | 9 | 36% |
| 5 | 5 | 23% |
| 2 | 2 | 9% |
| 0 | 0 | 0% |
Conversion tables are approximate — ETS applies equating across test forms to account for slight differences in difficulty. Your actual conversion may vary by 1–2 points.
Listening Section Scoring (0–30)
The Listening section contains 2 academic lectures and 1 campus conversation, totaling 17–28 questions (the exact count varies slightly between test forms). Almost all questions are worth 1 point each. ETS scales the raw score to 0–30.
Raw score to scaled score conversion (approximate)
| Raw Score (of ~28) | Scaled Score (0–30) | Approximate % Correct |
|---|---|---|
| 28 | 30 | 100% |
| 27 | 29 | 96% |
| 26 | 28 | 93% |
| 25 | 26 | 89% |
| 24 | 25 | 86% |
| 22 | 23 | 79% |
| 20 | 20 | 71% |
| 18 | 18 | 64% |
| 15 | 14 | 54% |
| 12 | 11 | 43% |
| 9 | 7 | 32% |
| 6 | 4 | 21% |
| 3 | 2 | 11% |
| 0 | 0 | 0% |
Speaking Section Scoring (0–30)
The Speaking section has 4 tasks: 1 independent (Task 1) and 3 integrated (Tasks 2–4). Each task is scored by trained human raters (and increasingly by ETS's automated scoring engine, SpeechRater) on a 0–4 rubric. Your scores are averaged and then converted to the 0–30 scale.
The 0–4 rubric for each Speaking task
Average task score → scaled score conversion (approximate)
| Avg Task Score (0–4) | Scaled Score (0–30) |
|---|---|
| 4.00 | 30 |
| 3.75 | 28 |
| 3.50 | 26 |
| 3.25 | 24 |
| 3.00 | 22 |
| 2.75 | 20 |
| 2.50 | 18 |
| 2.25 | 16 |
| 2.00 | 14 |
| 1.75 | 11 |
| 1.50 | 9 |
| 1.00 | 5 |
| 0.50 | 2 |
| 0.00 | 0 |
Raters evaluate three dimensions: Delivery (pace, pronunciation, fluency), Language Use (grammar, vocabulary, sentence variety), and Topic Development (relevance, coherence, how well you address the task). All four tasks are weighted equally.
Writing Section Scoring (0–30)
The Writing section has 2 tasks. Each is scored on a 0–5 rubric. The two task scores are averaged and converted to the 0–30 scale. ETS uses a combination of human raters and its automated e-rater engine, with the human score taking precedence when the two diverge significantly.
The 0–5 rubric for each Writing task
Average task score → scaled score conversion (approximate)
| Avg Task Score (0–5) | Scaled Score (0–30) |
|---|---|
| 5.00 | 30 |
| 4.75 | 29 |
| 4.50 | 27 |
| 4.25 | 25 |
| 4.00 | 24 |
| 3.75 | 22 |
| 3.50 | 20 |
| 3.25 | 18 |
| 3.00 | 17 |
| 2.75 | 15 |
| 2.50 | 14 |
| 2.00 | 11 |
| 1.50 | 7 |
| 1.00 | 4 |
| 0.00 | 0 |
For the Integrated Writing task, raters primarily assess how accurately and completely you summarize the lecture and connect it to the reading. For the Academic Discussion task, raters also evaluate your ability to express and support a personal opinion clearly within the academic conversation.
TOEFL Score Percentiles
Percentile ranks show how your score compares to all TOEFL test takers worldwide over a recent multi-year period. A percentile of 75 means you scored higher than 75% of all test takers.
| Total Score | Percentile Rank | Interpretation |
|---|---|---|
| 120 | 99+ | Perfect score — top fraction of all test takers |
| 114–119 | 99 | Exceptional — top 1% globally |
| 110–113 | 98 | Outstanding — exceeds almost every university minimum |
| 105–109 | 95 | Very strong — highly competitive for elite programs |
| 100–104 | 88 | Strong — meets requirements at most top universities |
| 95–99 | 78 | Good — comfortably above most graduate minimums |
| 90–94 | 65 | Solid — meets many graduate program requirements |
| 85–89 | 53 | Above average — meets most undergraduate requirements |
| 80–84 | 42 | Average — meets common undergraduate minimums |
| 70–79 | 25 | Below average — sufficient for some programs |
| 60–69 | 12 | Low — below most university minimums |
| 50–59 | 5 | Very low — unlikely to meet admission thresholds |
Percentile data is published periodically by ETS and reflects performance across all test takers, not just applicants to specific programs. The average total TOEFL score is approximately 85–88 for test takers worldwide.
TOEFL Score Statistics
How does your score compare to test takers worldwide? The figures below are drawn from the ETS TOEFL Score Data Summary (2023) and represent approximate global averages across all test takers. Data is approximate.
Global section averages (out of 30 each)
Key facts about TOEFL test takers
Average total scores by region
| Region | Avg Total Score (/ 120) | vs. Global Avg (85) |
|---|---|---|
| Western Europe | 97 | +12 |
| North America | 93 | +8 |
| South Asia | 88 | +3 |
| Latin America | 83 | −2 |
| East Asia | 78 | −7 |
| Middle East / Africa | 76 | −9 |
Source: ETS TOEFL Score Data Summary, 2023. Regional figures are approximate and rounded. Scores reflect all test takers in the relevant period, not applicants to specific programs.
University TOEFL Score Requirements
Most universities publish minimum TOEFL scores. Reaching the minimum does not guarantee admission — it only makes your application eligible for review. Competitive applicants typically score well above the stated minimum. Requirements below are for graduate programs unless noted.
| University | Min. Total | Min. Section |
|---|---|---|
| MIT | 100 | — |
| Harvard University | 100 | — |
| Stanford University | 100 | — |
| University of Oxford | 100 | 22 each |
| University of Cambridge | 110 | 25 each |
| Columbia University | 101 | — |
| University of Chicago | 104 | 26 each |
| Yale University | 100 | — |
| Princeton University | 100 | — |
| UC Berkeley | 90 | — |
| UCLA | 87 | — |
| University of Toronto | 93 | 22 each |
| University of Melbourne | 79 | — |
| ETH Zurich | 95 | — |
| University of Edinburgh | 100 | 20 each |
| NYU | 100 | — |
| University of Michigan | 84 | — |
| Georgetown University | 100 | — |
| University of Sydney | 85 | — |
| Imperial College London | 92 | 20 each |
What Is a Good TOEFL Score?
"Good" is relative to your goal. A score that earns admission to one program may fall short at another. Here is a practical breakdown by application type.
Most four-year universities require 80+ for undergraduate admission. Competitive universities (Ivy League, Oxbridge) effectively expect 100–110. A score of 90+ is a comfortable baseline for a wide range of programs.
Graduate admissions committees scrutinize TOEFL scores more closely, especially the Speaking and Writing sub-scores. STEM programs sometimes accept 90; humanities and social science programs often require 100+. Top-ranked programs expect 105–110.
Leading business schools (Harvard Business School, Wharton, Booth) require 104–109. A score of 100 meets the minimum at most programs but is not competitive at the top tier. Strong Speaking and Writing scores carry extra weight.
Medical and nursing programs require strong English proficiency for patient safety reasons. Most US and UK medical schools require 100–110. Some licensing boards (USMLE, GMC) have their own English proficiency requirements in addition to the university minimum.
US work visa applications often use TOEFL as a supporting document; immigration authorities typically look for 80+. Professional boards (teaching, engineering, law) set their own thresholds. Check the specific body that will evaluate your application.
Section score minimums to watch
Many universities list not just a total minimum but also per-section minimums. The most commonly enforced are:
- Speaking 22–26: Common requirement for teaching assistant (TA) positions and teaching-heavy graduate programs
- Writing 22–25: Frequently required for admissions to humanities, law, journalism programs
- Reading / Listening 22+: Common baseline for academic programs at English-medium institutions
Even if your total score exceeds the minimum, a single below-threshold section score can disqualify your application at universities that enforce section minimums.
How AI Grading on FullPracticeTests Compares to Real TOEFL
FullPracticeTests uses Claude (Anthropic) to score your Writing responses and provide feedback on Speaking. Here is how that compares to what ETS actually does on test day.
| Aspect | Real TOEFL (ETS) | FullPracticeTests |
|---|---|---|
| Reading & Listening | Automated, rule-based (100% accurate) | Automated, rule-based (100% accurate) |
| Writing scoring | Human rater + e-rater automated system | Claude AI using official ETS rubrics |
| Writing feedback | None — only a score is returned | Paragraph-level feedback within seconds |
| Speaking scoring | Human raters + SpeechRater AI | Claude AI rubric evaluation (text-based) |
| Speaking feedback | None — only a score is returned | Detailed rubric breakdown per task |
| Score turnaround | 4–8 business days for full report | Instant on submission |
| Score accuracy | Official, certified | Closely aligned; ±2–3 points on Writing |
| Content | Proprietary, fixed question bank | AI-generated unique content per exam |
How reliable is AI writing feedback?
AI-based essay scoring has been used in standardized testing for over two decades. ETS's own e-rater engine powers scoring across multiple exams. For practice purposes, AI scoring is highly useful: it catches grammar errors, assesses structure and coherence, and evaluates vocabulary range consistently and immediately.
The main limitation is nuance. Human raters can detect subtle factors like unusual but effective writing styles or culturally specific references. For preparation purposes, treating the AI score as a reliable estimate within ±2–3 points of what a human rater would assign is a reasonable approach.
Why unlimited practice content matters
Official ETS practice materials are limited to a small number of published tests. FullPracticeTests generates unique content for every exam, meaning you can take as many exams as you need without seeing repeated passages or questions. This is the single biggest practical advantage of AI-generated practice over official ETS materials for high-volume practice.
See your TOEFL score estimate right now
Take a free full-length TOEFL practice exam. All four sections. Instant AI scoring on Writing. No account required.
Start Free Practice Exam →No sign-up required · Full 4-section exam · AI scoring available after
Ready to find out where you stand?
Take a Free TOEFL Practice Exam →No sign-up required · Full exam · AI scoring available after