Independent practice platform. Not affiliated with, endorsed by, or sponsored by Pearson. PTE is a trademark of Pearson PLC. Scores here are practice estimates, not Pearson results.

PTE Core

How practice scoring works

The practice scorer here marks objective tasks, such as multiple choice, blanks and dictation, exactly by the rule Pearson publishes. Speaking and writing answers get an estimate: your speech is transcribed in the browser, then checked against a few of the traits Pearson publishes. This page sets out every rule and every limit in one place.

Drafted with AI writing tools, then edited and checked against Pearson and IRCC sources by the site’s operator. How pages are made

The short version

It is a practice estimate built on the traits Pearson publishes; use it to compare your own attempts over time, not to predict your Pearson score. Three kinds of task are marked in three different ways, and the difference matters when you read a result.

  • Objective tasks (the multiple-choice tasks, the blank tasks, Reorder Paragraphs, Highlight Incorrect Words and Write from Dictation) are marked exactly by the rule in Pearson's score guide. A point here is a point.
  • Speaking tasks are recorded, transcribed by your browser's speech recognition, and scored on the words and timing in that transcript. The numbers are estimates.
  • Writing tasks are scored from your typed text on simple checks for form, content and language. The form check uses a practice word range, and some traits are not estimated at all.

From your voice to a number

Pearson scores spoken answers with its own automated engine, which nobody outside Pearson can run. The practice scorer takes a simpler route that works in an ordinary browser.

How a spoken practice answer becomes an estimate Four boxes joined by arrows, left to right. 1, your voice: the browser records your answer. 2, transcript: your browser's speech recognition turns it into words. 3, traits: the words and timing are scored on the traits Pearson publishes, such as content, fluency and pronunciation. 4, estimate: the trait scores become a percentage, and Analytics turns your attempts into a practice estimate on the 10 to 90 scale. 1. Your voice recorded in the browser 2. Transcript browser speech recognition 3. Traits content, fluency, ... 4. Estimate %, then 10 to 90
Speaking tasks go through all four steps. Writing tasks skip the first two: your typed text is scored directly. Objective tasks (choices, blanks, dictation) are marked exactly, with no estimate in between.

The weak link is step 2. Browser speech recognition is built to understand speech, not to grade it, so it sometimes fixes a word you said badly and sometimes misses a word you said well. Everything after that step depends on the transcript, which is why the result screen shows it to you along with a playback of your recording: you can hear for yourself where the two disagree.

Speaking tasks: what is estimated

TaskPractice timingWhat the practice scorer checksNot estimated
Read Aloud28 s to prepare (skippable), up to 40 s to recordContent: share of the passage read in order. Pronunciation: share of the passage's words the recogniser heard. Fluency: speaking rate plus mid-phrase pauses and pitch jumpsVowels and word stress as such
Repeat SentenceSentence plays once; 15 s to record, no toneContent banded 0 to 3 on words matched (more forgiving about order than Pearson). Fluency: speaking rate alone. Pronunciation: share of your words that belong to the sentencePearson's word-sequence rule
Describe Image25 s to study, up to 40 s to recordFluency only, from speaking rate (up to 190 words a minute)Content and pronunciation; nothing checks what you said about the image
Respond to a Situation20 s to prepare, 35 s to recordFluency only, from speaking rate (up to 190 words a minute), the same as Describe ImageAppropriacy and pronunciation; nothing checks which words you used, the answer's layout, or register, tone and politeness
Answer Short QuestionOne short spoken answerThe transcript matched loosely to the expected answer or its listed alternatives: articles ignored, a one-letter slip allowedPronunciation (Pearson does not score it here either)

On every speaking task, a recording that was cut short or caught nothing is kept out of your averages, so a microphone fault does not drag your figures down.

Writing tasks: what is estimated

TaskPractice clock and aimForm in practiceOther checksNot estimated
Summarize Written Text10 minutes; aim 40 to 50 wordsOnly two sentences of 40 to 50 words get the mark (stricter than Pearson's 25 to 50)Content: how closely your sentences follow the passage. Grammar: capitals, end punctuation, repeated words. Vocabulary: range and lengthWhether you chose the main points
Write Email9 minutes; aim 80 to 120 words2 points for 80 to 120 words, 1 for 60 to 139Content against a model answer or the prompt; simple grammar, vocabulary and spelling checksEmail conventions and organization
Summarize Spoken Text10 minutes after the recording; aim 20 to 30 wordsAll or nothing: 20 to 30 wordsContent from the recording's key words; vocabulary from word variety and lengthGrammar and spelling

The practice form rules are deliberately simple and do not copy Pearson's part marks. Pearson's own form rules are on each task page; when the two disagree, Pearson's rule is the one that counts on test day, so read the word counter, not only the percentage.

Objective tasks: marked exactly

For these tasks there is nothing to estimate. The practice screen applies the same point rule that Pearson's score guide describes and shows the answer as soon as you submit.

TaskRule applied
Reading and Listening Multiple Choice, Single Answer; Select Missing Word1 point for the correct option, 0 otherwise
Reading and Listening Multiple Choice, Multiple Answers+1 per correct option, −1 per wrong option, never below 0
Reading: Fill in the Blanks, Reading & Writing: Fill in the Blanks1 point per correct blank, nothing taken off for a wrong one
Reorder Paragraphs1 point per correct adjacent pair
Highlight Incorrect Words+1 per changed word clicked, −1 per unchanged word clicked, never below 0
Listening: Fill in the Blanks, Write from Dictation1 point per correct, correctly spelled word

The rule is Pearson's; the practice items are not Pearson's, and nothing on this site sees a Pearson test or its scores.

How Analytics turns attempts into 10 to 90

Each attempt is saved as a percentage. Analytics averages your attempts and places the average on the 10 to 90 scale PTE Core reports on, so 0% maps to 10 and 100% maps to 90. Beside the overall figure it shows a rough CLB label from one simplified scale.

That label is a convenience, not the official conversion. IRCC converts each of the four skill scores separately with Pearson's table, and the cut-offs differ by skill: a score that is CLB 9 in Reading can be CLB 8 in Writing. To read a real result, or to set a per-skill target, use the official score chart and the PTE Core to CLB converter.

What the practice scorer cannot do

  • It cannot hear like Pearson's engine. Pronunciation and fluency come from the transcript and timing, not from a model of English sounds.
  • It cannot judge meaning. No practice check knows whether a Describe Image answer described the image, or whether a summary picked the main idea.
  • It cannot judge tone. Respond to a Situation and Write Email are scored partly on register in the test; the Write Email scorer only sees words, and the Respond to a Situation scorer marks fluency alone.
  • It does not know your test form. Pearson's raw points are combined into skill scores in ways it does not publish, so no practice tool can turn task scores into a Pearson score.

Used for what it is, the estimate is still useful. The same answer gets the same number every time, so a rising line in Analytics means your own answers are getting closer to what the traits reward. For a scored rehearsal on Pearson's own engine, Pearson sells an official scored practice test; the traits each task is scored on are in the PTE Core Score Guide.