Candidates should compare the task each PLAB 2 product supports, how the interaction responds to their choices, what its feedback actually demonstrates and the total access cost. AI patients are not unique to iatroX. A lower price, more detailed product description or longer feature list does not establish which system gives better feedback on a particular candidate's performance.
This comparison is published by iatroX and includes iatroX. It reviews public product descriptions checked on 3 October 2026 and proposes a way to assess them; it does not report a hands-on head-to-head test or claim a measured accuracy advantage.
What PassMedicine publicly describes
The PassMedicine PLAB page, checked on 3 October 2026, describes virtual patients that candidates can interact with to practise history-taking for PLAB 2. Its combined PLAB access is advertised at £12 for four months, £15 for six, £20 for nine and £25 for 12.
Those prices make it plainly cheaper than iatroX's paid subscription. The same page's written-question total concerns PLAB 1, not the number of virtual-patient cases. The public description reviewed does not provide enough detail to establish every aspect of PLAB 2 feedback, case breadth or mock-circuit behaviour; that is a documentation limit, not proof that a feature is absent.
What iatroX publicly describes
iatroX's September 2026 simulation announcement names PLAB 2 among its examination-specific tracks. It describes voice and text interaction, coached practice, uninterrupted exam mode, complete mock circuits, feedback linked to the transcript, Tutor-led remediation and a prescribed next case.
The launch total of 1,114 clinician-reviewed cases refers to the entire 18-track simulation library, not PLAB 2 alone. Clinician review is a content-development statement, not examining-body endorsement, clinical validation or a guarantee that every automated response is correct.
One complete simulation is free. Under iatroX's product information dated 3 October 2026, ongoing simulations are bundled with question banks, Socratic Tutor, planner and CPD tools for £29 monthly or £99 paid upfront annually. The annual equivalent is £8.25 a month billed annually. The commercial question is whether you need those connected learning methods, not whether the bundle contains more unrelated subjects.
Define the same candidate problem first
A fair comparison starts with a task, not two product homepages. Choose a fictional learning problem such as recognising and responding to a patient's concern while gathering relevant information. Define what a useful session would reveal about the candidate's behaviour.
Do not compare a history-taking exercise in one product with a full mock circuit in another and attribute every difference to quality. Assess the overlapping task first, then separately consider additional functions you need.
The GMC's PLAB 2 assessment description, checked on 3 October 2026, is the appropriate source for understanding the broader domains. A product test can examine one relevant behaviour without claiming to reproduce the whole assessment.
Inspect whether the conversation changes when your approach changes
Use fictional, non-confidential scenarios for any evaluation. Ask whether the patient response is relevant to the question, whether new information remains consistent with the earlier account and whether the interaction leaves room for clarification rather than rewarding a particular string of words.
A useful comparison would include a straightforward attempt and an imperfect one. For example, the candidate initially explains the proposed process without establishing the person's main worry, then returns to clarify it. The question is whether the system responds coherently to that recovery, not whether it produces a friendly answer.
Repeat enough varied attempts to avoid judging a product from one unusually good or bad exchange. Record the product version, mode and conditions so that later changes are not mistaken for the same tested system. No numerical results from such a test are claimed here; conclusions about observed behaviour require an actual documented run.
Evaluate feedback against the transcript
Feedback should be checked against what the candidate actually said and the task they were asked to perform. An observation that a concern was never explored is useful only if the transcript supports it. A reasonable alternative explanation should not automatically be marked as a failure because it differs from a model script.
For a proposed evaluation, ask an appropriately experienced reviewer to inspect a sample of feedback independently. Separate clinical-content errors, unsupported observations, missed important omissions and differences of reasonable judgement. Do not collapse these into a single satisfaction score.
Also examine educational usefulness. Can the candidate identify one change to make, and can a later unfamiliar case test that change? A detailed report may be less useful than a short, accurate observation if the candidate cannot translate it into another attempt.
Test convenience without confusing it with competence
Try a session on the device and in the environment you intend to use. Assess whether you can start the relevant activity, control the interaction, inspect feedback and understand the next action without excessive setup. Consider your own reading, audio and pacing preferences rather than assuming a universal ideal interface.
Do not infer a native integration or automatic transfer between providers. A candidate can manually combine a low-cost history-taking resource with occasional human observation or another learning tool, but the work of moving between them should be included in the decision.
Convenience is a legitimate benefit. It is not evidence that a learner can perform a physical examination, manage a real consultation independently or pass PLAB 2. Those are different questions with different evidence requirements.
A fictional purchasing decision
Omar has access to appropriate practical teaching and a reliable partner, but wants additional solo history-taking between sessions. His immediate need is narrow. The lower-cost PassMedicine package may be sufficient if its available experience fits that task.
Priya has a different problem: after a practice case she cannot explain why her reasoning became inconsistent. She wants repeated cases linked to a discussion of the misconception and an organised return to related material. She can examine whether iatroX's simulation-to-feedback-to-Tutor workflow provides that connection in practice.
Neither scenario establishes a universal winner. Both candidates should retain human observation where needed and avoid purchasing a new system merely to resolve anxiety about an approaching examination. These examples illustrate decisions, not testimonials or measured improvements.
Verdict by reader scenario
Choose by the remaining preparation task. PassMedicine deserves consideration for inexpensive additional virtual-patient history-taking. iatroX deserves consideration when its connected simulation, feedback and guided-learning functions match a specific need that justifies the higher price.
Keep your current setup when it already provides suitable repetition, source checking, practical teaching and observed feedback. Seek an experienced tutor or practical session when the unresolved skill cannot be evaluated through dialogue. Until a documented matched evaluation exists, a feature comparison should remain a feature comparison rather than a claim that either product is more accurate or predicts examination success.
Frequently asked questions
Is PassMedicine cheaper than iatroX for this comparison?
Yes: its published PLAB package checked on 3 October 2026 costs £12 for four months or £25 for 12 months, below iatroX's paid subscription. The packages contain different learning functions, so assess whether the additional functions are relevant to you.
Has this article tested the two products against each other?
No: it compares published scope and sets out an evaluation method. It does not report observed interaction scores, feedback accuracy or pass-rate effects.
Can either product establish readiness for every PLAB 2 skill?
No: conversational performance cannot demonstrate all physical, practical and interpersonal competencies. Use appropriate human observation and official assessment information alongside digital practice.