What does a strong O-Level Planned Response look like for each kind of prompt?

Does SEAB publish model answers for the O-Level English oral?

No. SEAB publishes the format and the band descriptors; it publishes no past clips, no past prompts and no model answers. Every answer below was written by OralPrep. Where we say an answer is Band 5 or Band 3, that is our placement against the Response descriptors, argued from their wording, not a mark an examiner gave. The prompts are as candidates recalled them, which is reliable on substance and approximate on wording.

What are the answers being measured against?

Against the K300 Response descriptors. The Planned Response is worth 15 marks, split between Response and Delivery. Response is judged on two criteria SEAB lists as "Development and organisation of ideas" and "Expression of ideas". These are the bands we place answers in:

Each descriptor has two halves, one per criterion, and a placement has to satisfy both. In our reading, that is the trap with an answer that reads well: well-organised ideas in plain words meet the first half of Band 5 but not the second, which asks for a "wide range" of vocabulary and structures. Every band, with Delivery and Spoken Interaction alongside, is on how Response, Delivery and Spoken Interaction are marked.

What kinds of prompt does the Planned Response use?

SEAB names no prompt types. Sorting the prompts candidates recalled from 2023 onwards, we group them into three shapes; the grouping is our working taxonomy, and it describes what was asked, not what will be. All the examples below are corroborated rows (two or more independent sources) from the recalled O-Level English oral topics.

• Judge: is it good, useful, meaningful? "Are the elderly women making good use of their time? Why, or why not?" (2023). "Do you think machines like the one shown in the video are useful?" (2025). "Is this a good way to manage and collect waste? Why or why not?" (2024).
• Recommend: should there be more, would you recommend it? "Would you recommend such a recreational activity for young people? Why or why not?" (2024). "Do you think there should be more open spaces like the one shown?" (2023). "Do you think there should be such street hawkers in Singapore? Why or why not?" (2024).
• Reflect: what do you think, prefer or now believe? "Did the video change your opinion of robots?" (2023). "Would you and your friends prefer hands-on learning?" (2023).

Some recalled prompts fit none of the three neatly. "Who will be attracted to this activity? Why do you say so?" (2025) and "How do you think people's plans may be affected by the weather?" (2025) ask you to read the situation in the clip. Our guidance for those is the same: a position in the first sentence, points that are developed rather than listed, and an ending that answers the question.

What does a strong answer to a judge prompt do?

In our reading, it finds the question underneath the question. Both of our existing worked scenarios are judge prompts, and both of the answers we place in the top band do this in their first sentence. To "Do you think machines like the one shown in the video are useful?", our top-band answer opens: "Useful to whom, though?" and then takes the airport, the passenger and the cleaner in turn. The full answers, each beside a middle-band answer to the same prompt, with line-by-line notes, are on the Planned Response page. This page does not repeat them; here is the one contrast from that pair we find most instructive:

What does a strong answer to a recommend prompt do?

In our reading, it says who the recommendation is for, gives a reason beyond the obvious one, and attaches a condition. "Would you recommend…?" invites a yes or a no; our guidance is to narrow the recommendation (to whom? on what condition?) and then explain it. Both answers below are new, written by OralPrep in October 2026.

Our top-band answer: recommending skateboarding (written by OralPrep, not SEAB)

Recalled prompt (2024, corroborated by three sources): "Would you recommend such a recreational activity for young people? Why or why not?" The clip, as recalled, showed teenagers skateboarding at an outdoor skate park. What the candidate says about skaters falling is their own experience, not a recalled detail of the clip.

"I would recommend it, though not to every young person, and not for the reason people usually give.

The usual argument is exercise, and that is true as far as it goes. Skateboarding is demanding physical work, and an afternoon at a skate park is an afternoon not spent scrolling on a phone. But the same could be said of almost any sport, so exercise alone does not explain why I would choose this one.

What persuades me is something I noticed watching skaters at the park near my home: how often they fall, and how little fuss anyone makes about it. A skater misjudges a jump, gets up, walks back and tries again. Nobody is grading them, and nobody laughs. For students my age, whose whole week is measured in marks, that is rare: a place where failing in public is ordinary, and where the only route to getting better runs through failing many times first. I would argue that kind of resilience is worth more than the exercise.

It also suits a particular kind of young person. Someone who thrives on a coach, a fixed schedule and teammates depending on them would probably be happier in a team sport. Skateboarding rewards the ones who prefer to teach themselves, progressing at their own pace, with friends nearby but not relying on them.

My one condition is safety. A helmet and basic protective gear should be non-negotiable, and a purpose-built skate park is far safer than a car park or a void deck, where skaters have to dodge cars and people walking past.

So yes, I would recommend it, chiefly because it teaches young people to fall, get up and try again, which is a lesson school rarely has time for."

How it meets the Response descriptors (our annotations)

• "I would recommend it, though not to every young person, and not for the reason people usually give." Organisation: the answer comes first, and the sentence then promises two things the response must deliver (who it is for, and the less obvious reason). The rest of the answer keeps both promises.
• "But the same could be said of almost any sport, so exercise alone does not explain why I would choose this one." Well-considered: the obvious reason is granted and then shown to be insufficient, so the candidate's own reason arrives as an argument rather than a list item.
• "What persuades me is…" / "the only route to getting better runs through failing many times first" Range of structures and vocabulary: a cleft sentence that puts the reason in focus, and a figurative "route… runs through" instead of "you must fail to improve". Also "misjudges", "in public", "resilience": words chosen for the exact meaning.
• "For students my age, whose whole week is measured in marks" Development: the reason is tied to the candidate's own life, which is something only they could say. The relative clause carries that in a single sentence.
• "Someone who thrives on a coach, a fixed schedule and teammates depending on them would probably be happier in a team sport." Organisation and range: this keeps the first promise (who it is for), and does it with a hypothetical rather than a flat claim about team sports.
• "non-negotiable", "purpose-built", "dodge cars and people walking past" Well-chosen vocabulary: the condition is specific, and the reason the void deck is worse is given rather than assumed.
• "So yes, I would recommend it, chiefly because it teaches young people to fall, get up and try again" Organisation: the ending answers the prompt again with the reason the whole answer has argued for, rather than repeating the opening.

Our placement: Band 5 on Response, and why. First half of the descriptor: we read it as "well-considered" because it narrows the recommendation and argues for a reason it has first shown the obvious one cannot carry, and "well-developed and organised" because every paragraph keeps a promise the opening made. Second half: "a wide range of well-chosen vocabulary and structures" is the strand an examiner could most reasonably argue down, since the register is a teenager speaking rather than an essay. We still read it as wide rather than Band 4's "range of largely appropriate" because it varies its sentence shapes (cleft, relative clause, hypothetical, condition, a three-verb close) and picks words for precise meaning rather than reaching for the nearest general one. If an examiner judged the range "largely appropriate" rather than "wide", the same answer would sit at the top of Band 4. It runs to about 290 words, by our count. On Delivery this text says nothing.

What does a strong answer to a reflect prompt do?

In our reading, it says what changed and what did not. A prompt such as "Did the video change your opinion?" can be answered with a flat yes; our guidance is to separate the parts of an opinion, because a view can move on one point and hold on another, and then to deliver both parts. Both answers below are new, written by OralPrep in October 2026.

Our top-band answer: reflecting on robots (written by OralPrep, not SEAB)

Recalled prompt (2023, corroborated by two sources): "Did the video change your opinion of robots?" The clip, as recalled, showed a robot exhibition with robots singing and drawing portraits. What visitors did at the exhibition is the candidate's guess, and the answer says so.

"Partly. It changed what I think robots can do, but not what I think they are for.

Before watching, my picture of a robot was something in a factory: a mechanical arm on an assembly line, repeating one movement thousands of times a day. So seeing robots sing and draw portraits genuinely surprised me. Drawing a face is not a repetitive task, because every face is different. If a machine can manage that, I have to admit my idea of what robots are capable of was out of date.

What has not changed is my sense of what they are for. As I see it, a robot is a tool: it exists to do a job people need done, especially work that is dangerous, monotonous or more precise than human hands can manage. A robot that draws portraits does not overturn that. It is still doing a job someone designed it for, and at an exhibition, I would guess, that job is to impress visitors. I suspect people there spend longer watching the robot draw than looking at the drawing afterwards, which suggests the real product is the demonstration, not the portrait.

The clearest test, for me, is whether I would want to keep what it made. A birthday card from my younger sister, uneven lettering and all, means more to me than a flawless portrait from a machine, because the card was made for me, while the portrait was made to show off the machine. No improvement in the robot's skill changes that difference in purpose.

So the honest answer is yes and no: the video widened my idea of what robots can do, but it confirmed that they are tools we build for our own purposes, not makers whose work we treasure for its own sake."

How it meets the Response descriptors (our annotations)

• "Partly. It changed what I think robots can do, but not what I think they are for." Organisation: a one-word answer, then a distinction that splits the response into two halves. The second and third paragraphs deliver one half each.
• "Before watching, my picture of a robot was something in a factory" Development: a reflection needs a "before". Stating the old view concretely is what makes the change clear.
• "Drawing a face is not a repetitive task, because every face is different." Development: the reason the view changed is explained, not just asserted.
• "What has not changed is my sense of what they are for." Organisation: the second half is signposted and then argued. "Dangerous, monotonous or more precise than human hands can manage" says what robots are for in the candidate's view, so the opening's promise is kept.
• "I suspect people there spend longer watching the robot draw than looking at the drawing afterwards" Well-considered: the candidate draws an inference (the product is the demonstration) and frames it honestly as a supposition, not a claim about what the clip showed.
• "The clearest test, for me, is whether I would want to keep what it made." Range of structures and vocabulary: the candidate sets their own test and applies it. "Uneven lettering and all", "flawless", "show off the machine" are chosen for meaning, and the "because… while…" sentence holds a contrast in one line.
• "tools we build for our own purposes, not makers whose work we treasure for its own sake" Organisation: the ending closes both halves (ability widened, purpose confirmed) in one balanced sentence.

Our placement: Band 5 on Response, and why. First half: "well-considered" because it splits "opinion" into ability and purpose and argues each; "well-developed and organised" because each half has a before or a claim, a reason and an example, and the ending answers both halves the opening announced. Second half: we read the range as "wide" rather than Band 4's "range of largely appropriate" vocabulary and structures, on the evidence of the cleft opening to the second half, the conditional, the "because… while…" contrast and the balanced final sentence, with words chosen for exact meaning ("monotonous", "overturn", "flawless"). As with the skateboarding answer, range is the strand an examiner could weigh differently, and a judgement of "largely appropriate" would put it at the top of Band 4. It runs to about 300 words, by our count. Delivery cannot be read from it.

Why don't these answers show Delivery?

Because Delivery is assessed on pronunciation, fluency and intonation, and a written answer shows none of them. The same transcript could earn very different Delivery marks depending on how it was spoken. We give no written advice on how any of these answers should sound, as a rule across OralPrep: a claim about sound is one we cannot source.

How should you use model answers without memorising them?

Our guidance is to take the moves, not the sentences. Each answer we place in the top band opens with a position, narrows or questions the prompt, develops one or two points with an example only the speaker could give, and ends by answering the question in a sharper form. In our view those moves carry over to other prompts in a way the sentences cannot, but this is OralPrep's house approach, drawn from answers we wrote ourselves. The descriptors name qualities, not a structure, and a Band 5 response need not take this shape. One way to practise: take one recalled prompt of each type, write only the first and last sentence, and then speak the middle without notes. Opening and closing moves are set out in openings and endings for the Planned Response, and how the middle holds together in structuring a Planned Response.