Do I need to be accurate with numbers in Describe Image?

No. The AI doesn't compare your numbers to the image. As long as you sound fluent and use the right vocabulary, content scores well even with approximate numbers.

Can I use the same template for every image?

Yes — that's the recommended strategy. A memorised template you deliver fluently beats an improvised description every time.

What if the image is a map or process diagram?

Swap the trend language ('highest', 'increased') for spatial/sequential language ('starts at', 'leads to', 'passes through'). The 5-step structure stays the same.

How long should I speak?

Aim for 35–38 seconds of the 40-second window. Trailing silence at the end is penalised.

PTE guide

PTE Describe Image — tips, template & practice

Updated 29 May 2026 · 5 min read

PTE Describe Image gives you 25 seconds to study a graph, map or process diagram and 40 seconds to describe it. It's the most-feared PTE Speaking task because the visual variety is wide (bar charts, pie charts, flow diagrams, maps, photographs) and the time pressure is unforgiving.

The secret most high-scorers know: the AI doesn't grade visual accuracy. It grades fluency, pronunciation, vocabulary, and structure. So a memorised template you can deliver smoothly beats a 'correct' description full of hesitations.

Want to skip ahead?

How it's scored

Content (max 5): covers main features (trend, max, min, comparison). Doesn't need to be 100% accurate.
Oral Fluency (max 5): smooth delivery without long pauses or false starts.
Pronunciation (max 5): clear vowels, correct word stress.

A fluent, templated description with one minor numerical error scores higher than a hesitant, perfectly accurate one.

The template that works

Universal Describe Image template (40 seconds):

1. Opener (5s): "The image is a [bar chart / line graph / pie chart / map / process diagram] which illustrates [topic from title]." 2. Main feature (10s): "The most striking feature is that [highest value / clearest trend]. Specifically, [one number]." 3. Comparison (10s): "In contrast, [lowest / opposite trend]. Meanwhile, [middle category]." 4. Detail (10s): "Other notable points include [two more specifics]." 5. Closer (5s): "Overall, the image suggests [one-sentence conclusion]."

Memorise this verbatim. Plug in the image-specific details during the 25s prep window.

Top tips for higher scores

Memorise one universal template — don't try to invent structure on the spot.
Pick TWO specific numbers to mention. More than three and you'll stumble.
Start speaking immediately at the 'begin recording' beep — silence at the start tanks fluency.
Use linking words: 'specifically', 'in contrast', 'meanwhile', 'overall'. They lift the Vocabulary sub-score.
If the image is a map or process, swap 'highest/lowest' for 'starts at / ends at / passes through'.
Don't say 'I don't know' or 'I can't see' — keep generating sentences even if you're improvising.

Common mistakes to avoid

Pausing for 3+ seconds at the start trying to 'understand' the image.
Trying to mention every number — you'll run out of time and sound rushed.
Apologising or admitting confusion — the AI hears it as disfluency.
Reading the title verbatim instead of paraphrasing — looks lazy to human reviewers.
Ending early (10s+ silence at the end) — the AI marks the silence as incomplete fluency.

Frequently asked questions

Ready to put this into practice?

Start free practice with instant AI scoring — no credit card needed.