Checked 28 September 2026. From the Civil Service Careers online-tests, additional-information and Gen AI pages for the 2025-26 campaign, and the 2026 recruitment data on GOV.UK. The test formats can change between campaigns; the 2027-entry pages were not yet published when this was checked.
The first real gate in the Civil Service Fast Stream is not an interview and not an essay. It is a pair of online judgement tests — the Work-based scenarios test and the Case Study assessment — and they are nothing like the verbal and numerical reasoning papers many candidates still prepare for. This guide sets out, from the official online-tests page for the 2025-26 campaign, what each test contains, how the four-point effectiveness scale works, what the one published worked example teaches, and what happens in the weeks around the tests.
Where other pages get this wrong
The Fast Stream online stage is two video-based judgement tests: 12 work scenarios and a 24-situation case study. Neither is a verbal or numerical reasoning paper.
Commonly published instead: verbal and numerical reasoning tests, an e-tray exercise and a video interview as the early sifts. Those descriptions belong to earlier versions of the process.
The official online-tests page lists exactly two tests for the 2025-26 campaign. Both show you video and audio from workplace characters. Both ask you to judge how effective different actions would be. Neither is timed: the page gives about 90 minutes as a guide, and you have 7 days to finish once the tests are released.
Civil Service Careers, online tests page (2025-26 campaign).
The two tests side by side
| Feature | Work-based scenarios | Case Study assessment |
|---|---|---|
| What it measures | Interpersonal skills: judgements in work situations involving other people | Decisions and judgements in work situations, and adapting to changing circumstances |
| Size | 12 scenarios | 7 pieces of background information, then 24 situations |
| Material | Video and audio from characters in each scenario | One evolving scenario, with video and audio from the characters involved |
| Your task | Rate how effective each possible action would be, on a four-point scale | From four responses, pick the most effective and the least effective |
| Your role | A role in each work situation (the official example puts you in charge of a team) | An imaginary role you learn about, then respond in as the situation changes |
| Timing | Both tests together: about 90 minutes as a guide, not timed | Both tests together: about 90 minutes as a guide, not timed |
| Deadline | 7 days from the moment the tests appear on your applicant homepage | 7 days from the moment the tests appear on your applicant homepage |
Everything in this table is from the official online-tests page. Note the difference in the answer format. The Work-based scenarios test asks for a rating of each action on its own, so several actions can share the same rating. The Case Study asks you to pick the most effective and the least effective of four responses.
The four-point effectiveness scale
| Rating | Official definition |
|---|---|
| Counterproductive | An unacceptable action that would make the situation worse |
| Ineffective | A poor action that would not help the situation |
| Partially effective | A useful action that would be of some help to the situation |
| Effective | A good action that would help to resolve the situation |
The scale is the whole test, so it is worth reading the definitions slowly. The line between Counterproductive and Ineffective is whether an action makes things worse or simply fails to help. The line between Partially effective and Effective is whether an action helps a little or actually moves the situation towards being resolved. Many candidates rate every sensible-sounding action as Effective. The official example below shows why that loses marks: two of its four actions sound reasonable and are still keyed Ineffective.
The official worked example, decoded
The online-tests page publishes one example with its answers and rationales. You lead a team on a National Survey project and share important IT equipment with another department. Kamala, a colleague on your team, tells you that the other department has taken on more work and is now booking far more time on the equipment. She adds that some other teams have similar equipment that is not being used right now.
| Action | Official rating | Official rationale (paraphrased) |
|---|---|---|
| A. Ask the other department's manager to find alternative equipment | Ineffective | Restricts use of shared equipment, though the other team may be able to get extra resources |
| B. Talk to the other manager about your concerns and share a contact with underused equipment | Effective | Opens a dialogue and offers a constructive alternative |
| C. Ask colleagues to tell the other manager your project takes priority | Ineffective | May free up some time, but puts the burden on colleagues |
| D. Ask colleagues to book as much equipment time as possible | Counterproductive | Limits the other team's work and opens no dialogue |
Rate it yourself: the official Work-based scenarios example
This is the one worked item Civil Service Careers publishes. You lead a team on a National Survey project and share important IT equipment with another department. Kamala, from your team, says the other department has taken on more work and is booking far more time on the equipment, and that some other teams have similar equipment that is not in use right now. Rate each action, then check against the official answers.
A. Contact the manager of the other department and request they find alternative IT equipment.
B. Speak with the manager of the other department about your concerns and share a contact who has similar IT equipment that is currently underused.
C. Ask your colleagues to tell the manager of the other department that your project takes priority.
D. Ask your colleagues to book as much time as possible with the equipment to make sure the project is done.
Answers and rationales are paraphrased from the official online-tests page. The real test adds video and audio; this text version shows only how the scale works.
Three patterns in this key are worth carrying into the real test. They are our reading of one example, not published marking rules:
What the example rewards (our reading)
- You act yourself, in person. The Effective action is the one where you speak to the other manager directly. Sending colleagues to do it (C) or pushing the problem onto the other team (A) both fall short.
- You use the information you were given. Kamala mentions underused equipment elsewhere. The Effective action is the only one that uses that fact. Scenario details are rarely decoration.
- Winning at someone else's expense is penalised. Grabbing as much time as possible (D) would get your project done and is still the only Counterproductive action, because it harms the other team and shuts down dialogue.
- Asserting priority is not a solution. C might even work in the short term. It is still rated Ineffective, because it offloads the difficulty onto colleagues.
One worked example is thin evidence. Treat these as habits to test against the video items, not as a scoring formula.
How the Case Study assessment is different
The Case Study assessment starts by giving you 7 pieces of background information about an imaginary role. It then runs one evolving scenario across 24 situations, and for each situation shows four possible responses. You choose the most effective and the least effective. Because the scenario evolves, what you learned from the background information and from earlier situations stays relevant later on. The official page describes it as an assessment of how well you make decisions and judgements in realistic work situations and adapt to changing circumstances.
Two practical consequences follow. First, read the background information properly before starting; it is the context for every later judgement. Second, the test is not timed, but it is long, and a break in the middle means holding the storyline in your head across the gap. Our suggestion: if you can, take it in one sitting in a quiet place.
Picking the least effective response is a separate skill from picking the best one. In the Kamala example, the weakest action was the one that harmed another team, not the one that simply did nothing useful. If the Case Study follows the same logic as the rating scale, a response that makes things worse should sit below one that merely fails to help. That is our inference from the published scale, not an official rule.
What happens around the tests
| Step | What the official page says |
|---|---|
| After you submit the initial application | There is a delay before the tests unlock; they show as locked on your applicant homepage, and this may take a few days |
| Tests released | You get an email; check your homepage and inbox, including spam |
| Completion window | 7 days from the date and time the tests are made available; the deadline shows on your homepage |
| Expected length | About 90 minutes as a guide; the tests are not timed |
| Pass marks | Set after the tests close; the numbers are not published |
| Outcome | The 2025-26 campaign aimed to tell candidates by early December 2025; results go out in batches, not all at once |
| Feedback | Every candidate gets detailed feedback, whatever the outcome |
Two lines in that table matter more than they look. The page says final pass marks will be set and outcomes sent out, but it never publishes the marks or how they are set. So anyone quoting a "pass score" for these tests is guessing. And the feedback, which everyone receives, is explicitly about your performance and potential, not the selection decision. The page warns that even candidates with very positive feedback may not progress, because the process is so competitive.
The additional information stage applies to 9 schemes plus Parliament
Pass the online tests for certain schemes and a further stage appears on your homepage: written questions to complete online. This is not universal. The official additional-information page lists 10 entries that require it: 9 Fast Stream schemes and the separate House of Parliament scheme.
| Scheme needing additional information | Also note |
|---|---|
| Cyber Security | |
| Diplomatic and Development | |
| Diplomatic and Development Economist | |
| Government Economics Service | |
| Government Social Research | |
| Government Statistical Service | |
| Risk Management | |
| Project Delivery | |
| Government Operational Research Service | |
| House of Parliament | Not part of the Fast Stream, but can be one of your preferences |
| Rules for all 10 entries | 7 days to complete; the aim is to assess within 2 weeks of submission; no individual feedback, because answers are anonymised before review |
The page adds one rule that catches people out: you must complete all of the additional information questions before you can progress with any of your schemes. If one of your three preferences is on this list, the extra questions stand between you and the assessment centre for all of them.
Gen AI: allowed for preparation, not during the tests
The Fast Stream has a specific page on generative AI. You may use tools such as chatbots to research the Civil Service, to simulate practice scenarios and to get feedback on practice answers. You must not use them while completing the online tests, the assessment centre or the Final Selection Board. The page says that where AI use is suspected during the online tests and assessment centre, it may affect your application. It frames this as a matter of the honesty and integrity values in the Civil Service Code.
How many people this stage is filtering
| Measure (2026 intake, applied Oct to Nov 2025) | Number |
|---|---|
| Registrations | 78,407 |
| Applications, counting each person once by first preference | 69,046 |
| Applications counting every preference | 188,865 |
| Vacancies | 654 |
| Recommended for appointment | 703 |
| Success rate (recommended divided by first-preference applications) | 1.0% |
These figures are from the Fast Stream recruitment data 2026 on GOV.UK. The Cabinet Office does not publish how many candidates passed the online tests as a separate figure, so the stage-by-stage drop-off cannot be stated. What the totals do show is that roughly one first-preference applicant in a hundred was recommended. Where along the way the rest drop out is not published, but the online tests are the first stage everyone must pass.
A preparation plan that fits the real tests
Before the window opens (our advice)
- Learn the four definitions of the rating scale well enough to explain the difference between each neighbouring pair without looking.
- Work through the official Kamala example above until you can explain each key, not just recall it.
- Read the Civil Service Code. Its honesty, integrity, objectivity and impartiality values are the standard a situational test is most likely to reward.
- Practise spotting the useful detail in a scenario, such as the underused equipment in the Kamala example.
In the 7-day window (our advice)
- Start early in the window. The 7 days are a hard deadline; the 90 minutes are only a guide.
- Sit somewhere quiet with sound, because both tests use video and audio.
- Take the Case Study in one sitting if you can, so the background information stays fresh.
- Answer as you would actually behave. The official page quotes a fast streamer advising exactly that.
Practise it, don't just read it
There is no ProSyllabus quiz in the Work-based scenarios or Case Study format, because the real items are video-based. The closest practice is judgement against the Civil Service Code, and separating fact from inference in what a scenario tells you.
The ProSyllabus Fast Stream library has 12 quizzes: four on Written Advice Exercise skills (drawing conclusions from a brief, assumptions and inference, evaluating arguments in policy prose, weighing options against stated criteria), five on reading and working with data (tables and charts, percentages and change, multi-step problems, index numbers and real terms, scheduling and capacity), and one each on the Civil Service Code, fact versus inference versus judgement, and summarising a brief faithfully.
Frequently asked questions
Are the Fast Stream online tests timed?
No. The official page says the two tests should take about 90 minutes as a guide but are not timed. You have 7 days from the moment they are made available on your applicant homepage, and you must finish each stage before that deadline.
Is there a numerical or verbal reasoning test in the Fast Stream?
Not in the online stage of the 2025-26 campaign. The official page lists two tests: Work-based scenarios and the Case Study assessment, both situational judgement tests with video and audio. Numbers do appear later: the Written Advice Exercise at the assessment centre includes tables and graphs, and the guide advises having a calculator ready.
What is the pass mark for the Fast Stream online tests?
It is not published. The official page says only that final pass marks will be set and candidates told the outcome; it gives no figure and no method. For the 2025-26 campaign that was aimed for early December 2025.
Do I get feedback if I fail the online tests?
Yes. The official page says all candidates receive detailed feedback on their performance and potential, whatever the outcome. It also warns that positive feedback does not mean you will progress.
Does every candidate complete the additional information stage?
No. It applies to 9 Fast Stream schemes plus the separate House of Parliament scheme: Cyber Security, Diplomatic and Development, Diplomatic and Development Economist, Government Economics Service, Government Social Research, Government Statistical Service, Risk Management, Project Delivery, Government Operational Research Service, and the House of Parliament scheme. You get 7 days to complete it; the aim is to assess it within 2 weeks, and there is no individual feedback.
Can I use ChatGPT or another AI tool during the online tests?
No. The Fast Stream Gen AI guidance allows these tools for preparation, such as research, practice scenarios and feedback on practice answers, but not during the tests or assessments. Suspected AI use during the online tests or assessment centre may affect your application.
Sources
- official — Civil Service Careers — Online tests
- official — Civil Service Careers — Additional information
- official — Civil Service Careers — Using Gen AI in the Fast Stream process
- official — Civil Service Careers — Understanding the application process
- official — Civil Service Careers — Initial application stage
- official — GOV.UK — Civil Service Fast Stream recruitment data 2026
- official — GOV.UK — The Civil Service Code
- official — Civil Service Careers — Fast Stream (home)
official = a document published by the conducting body. reported = a news or coaching site we could not check against an original. Where sources disagree this page says so rather than picking one.
