Why PTE Scores Dropped After August 2025 — and What to Do About It

If your practice scores stopped matching your real scores some time after August 2025, you are not imagining it, and it is probably not bad luck. Two specific things changed on 7 August 2025, and between them they punish exactly the preparation style that used to work.

What actually changed

Pearson made two changes to the test on that date.

Two new speaking tasks. Summarize Group Discussion and Respond to a Situation joined the Speaking and Writing part, taking the test to 22 scored task types. Both are different in kind from what came before: one asks you to track several speakers and report what a group covered, the other gives you an everyday situation and asks you to respond as a person would, not as a candidate would.

Human review of Content. A human expert now also reviews the Content score on seven question types, alongside the automated scoring. Pronunciation and oral fluency remain fully machine-scored.

It is worth being precise about what did not change, because a lot of commentary gets this wrong. The score report format did not change. The price did not change. Results still arrive in about two days. And the Enabling Skills section was not removed in 2025 — Pearson replaced it with the personalised Skills Profile back in November 2021.

Why this hurt template users

For years, a large part of the PTE coaching industry sold memorised frameworks. A fixed opening for Describe Image. A fixed skeleton for Re-tell Lecture. An essay introduction you could write before reading the prompt.

Those worked because an automated Content score, scoring at speed, rewarded fluent delivery and relevant-sounding vocabulary. A confident, well-paced answer that only loosely addressed the image could still score.

A human reviewer reads for whether you answered this question. A beautifully delivered 40 seconds that describes a bar chart in general terms, while the screen shows a map of rainfall in two regions, now reads as what it is. The delivery marks survive. The content marks do not.

This is the single most common cause of the pattern we see in students who come to us after a disappointing result: oral fluency and pronunciation high, Speaking overall dragged down, and a candidate who is certain the machine made a mistake.

The two new tasks reward the opposite habit

Summarize Group Discussion cannot be templated in any useful way. You do not know in advance how many speakers there are, whether they agree, or where the disagreement sits. The marks come from tracking who said what and reporting it accurately — a listening and note-taking skill, dressed as a speaking task.

Respond to a Situation is stranger still for anyone trained on exam English. The scenario is ordinary: a colleague has asked you something, a delivery has gone wrong, a friend needs an answer. A stiff, formal, memorised register sounds worse here than plain speech. Candidates who have spent six months learning to sound like an IELTS essay find this genuinely difficult.

What to change in your preparation

1. Keep structure, drop scripts. You still need a reliable order to speak in — otherwise you stall. What you cannot keep is pre-written content. Learn a shape, fill it live.
2. Practise naming specifics out loud. For Describe Image, force yourself to say the actual axis labels, the actual highest value, the actual place names. Specifics are what a human reviewer is looking for.
3. Record and listen back. Not for pronunciation — for whether you answered the prompt. Most students are shocked at how little of their answer was about the thing on screen.
4. Train note-taking separately. Summarize Group Discussion, Re-tell Lecture and Summarize Spoken Text all rest on the same underlying skill. Work on it as a skill, not as three tasks.
5. Use mock tests that reflect the current format. A practice platform that has not added the two new tasks is preparing you for a test that no longer exists.

The same image answered two ways

Suppose the screen shows a line graph of rice production in two divisions between 2015 and 2024.

The template answer. "The given image is a line graph which illustrates information about two categories over a period of time. As can be seen clearly, there are significant variations between them. The first category shows an upward trend while the second shows fluctuation. In conclusion, the graph provides valuable insight into the data presented."

Fluent, well-paced, forty seconds filled. It also describes nothing. Swap in any line graph and the answer still fits — which is precisely the problem.

The specific answer. "This line graph compares rice production in two divisions from 2015 to 2024. Rangpur rises steadily from around two million tonnes to just over three. Khulna is flat until 2019, then falls sharply after 2021. By 2024 the gap between them is the widest on the chart."

Less polished. Noticeably better scoring, because every clause is about the thing on the screen. You do not need elegant English here. You need the axis labels, the direction, the turning point and the comparison, said out loud in plain sentences.

If your score fell between sittings

Compare your Skills Profile from both attempts before you book a third. If fluency and pronunciation held steady while the overall Speaking score dropped, you have a content problem, not an English problem — and content problems are fixable in weeks rather than months.

Note also that some of the score requirements for Australian Home Affairs visas changed from the same date. Before you conclude that you missed your target, confirm what the target currently is for your subclass and your test date on the official source.

Language Academy Bangladesh runs PTE batches of no more than 12 students in Dhanmondi, Dhaka, with mock tests built around the current 22-task format. Book a free consultation if you would like us to read your last score report with you.