A PLAB 2 mock should run as a 16-station circuit with eight minutes per station, timed transitions and feedback across the three scoring domains. In a six-week preparation plan, placing it near the start of week 5 leaves roughly 14 days to respond. No mock total predicts a GMC result because station cut scores and pass requirements are set for each diet, as the PLAB 2 scoring guide explains.
The report should identify repeated behaviour early enough for you to change it. Several loosely run mocks do not provide that information simply because there are more of them.
Why 16 stations matters
The GMC describes PLAB 2 as 16 eight-minute scenarios, with at least two rest stations and about three hours at the assessment centre. The circuit requires you to keep selecting relevant information, making safe decisions and communicating clearly well beyond one consultation.
Fatigue can change performance. A focused history early in the circuit may become a list of every possible question later on. Explanations can shorten and lose clarity, while safety-netting may disappear. A single eight-minute station cannot show what happens after sustained concentration.
A full mock also tests recovery. One poor station ends when the bell sounds, even if the diagnosis was missed or the closing sentence unfinished. The next transition is for the next task. The report and your own notes should show whether attention to the previous case affected later stations.
Single-station work remains useful for learning a technique or changing one behaviour. A full circuit adds the roughly three-hour duration, later-station concentration and recovery between tasks.
Reading a mock like an examiner
The GMC assesses each scenario across three domains: data gathering, technical and assessment skills; clinical management skills; and interpersonal skills. Across 16 stations, a mock can record 48 domain observations before any written comments are considered. An overall percentage is hard to act on when it hides where the same problem returned.
Read the domain comments alongside the station results. A management concern spread across several stations calls for different work from one poor station with no recurring pattern. The GMC results guidance explains that the real assessment combines numerical domain scores with global judgements for standard setting. A mock provider cannot know the cut score for a future GMC diet.
The following report is fictional. It illustrates how to read a pattern and does not reproduce GMC cut scores.
| Domain | Stations with a concern | Repeated comments |
|---|---|---|
| Data gathering, technical and assessment | 3 of 16 | One missed red flag, two unfocused histories |
| Clinical management | 9 of 16 | Delayed escalation, vague follow-up, missing safety-net |
| Interpersonal skills | 2 of 16 | Jargon and no check of understanding |
In this fictional report, management is the repeated problem. The remaining practice should place more emphasis on choosing and stating a safe plan than on collecting extra history questions. Review whether the immediate action, escalation threshold, follow-up interval and safety net appear where the case calls for them.
A data-gathering pattern needs a different response. Start by stating the problem and the red flags that would change the decision, then cut questions that do not alter the assessment. In an examination or procedure station, work on the missed step instead of adding a longer spoken script. For an interpersonal pattern, practise a plain-English explanation, check understanding and involve the patient in the decision. Then see whether the behaviour changes across different station types.
Academy mocks vs online mocks vs AI mocks
An in-person academy circuit can provide human role-players and physical room changes through 16 timed stations. It also requires a fixed date, travel and a fee. Ask how role-players are briefed, who provides feedback and whether all 16 stations run under the stated timing. An online peer-organised mock is easier to schedule across countries and can cost nothing. It can test eight-minute pacing, spoken structure and recovery across a long video call. Its limitation is variability: role-players may know the case too well, friends may avoid direct criticism, and one assessor's domain mark may mean something different from another's. The result can still guide practice, but it is not calibrated to a future GMC diet.
An AI mock can run on demand and apply the same feedback structure across repeated stations. It may be easier to schedule around shifts or time zones. It cannot reproduce hands-on examination, physical room movement or every nuance of a human role-player and examiner.
The formats offer different conditions. Academy circuits can add physical room changes and a cohort. Online peer mocks are easier to organise without travel. AI mocks offer availability and a consistent report structure. The broader academy and self-study comparison places the choice inside the preparation budget.
Free options and their limits
Peer circuits and community mock days are worth taking when they reproduce 16 eight-minute stations rather than a loose sequence of discussions. A free circuit can expose late-station fatigue, poor transitions and explanations that repeatedly overrun. It also forces 16 spoken performances into one sitting, which is difficult to reproduce through silent notes.
Peer marking is not calibrated, and community organisers cannot know a future diet's cut scores. Record observations by station and GMC domain. A specific note about missed escalation is easier to use than an unsupported percentage.
The platform's mock
As listed on 4 August 2026, GK's PLAB 2 Study Centre's AI mock exam runs an 8-station format for £5 and a 16-station format for £10, scores the same three domains and provides a per-station report immediately after completion. The 16-station option mirrors the real exam's scenario count.
Using the report
Use the report to name the most frequent unsafe or ineffective behaviour. Make the change observable, such as escalating an unwell patient before discussing routine follow-up, instead of writing “improve management”.
Practise that behaviour on its own, then put it back into mixed eight-minute stations. Change the case order so the next task is not predictable. After each station, record whether the target behaviour appeared and note what prevented it when it did not. Longer blocks can then test whether the change survives task switching and falling concentration. Keep the day before the exam for sleep and logistics instead of introducing a new system.
A second 16-station mock can check a specific change from the first. Use different cases with the same eight-minute timing, transitions and domain report. Compare whether the recurring concern is still present under circuit pressure. If nothing changed between mocks, another score will not explain why.
Questions candidates actually ask
When should you take a PLAB 2 mock exam?
In a six-week plan, early in week 5 leaves close to 14 days to identify a repeated weakness and change the practice routine. If your preparation is longer or shorter, keep the same principle: sit the mock early enough to use its feedback.
Are free PLAB 2 mocks worth it?
Yes, when they provide 16 timed stations, realistic transitions and specific peer feedback. They can test pacing, stamina and spoken performance, but their marks should not be treated as calibrated GMC scores because assessors and standards vary.
How many mocks before PLAB 2?
One 16-station mock run to the correct timing can be useful if its report changes your remaining practice. A second can test whether a specific weakness found in the first mock has improved, but repeating mocks without changing the practice plan adds little.