The Mock Exam is in public beta.

The Brightroom Mock Exam goes public beta this morning. Three modes — Full, Half, Section — share one runner that mirrors test day to the minute. A candidate sitting the Full mode tomorrow will work through the same 64 questions across the same three 45-minute sections, with the same optional ten-minute break, against the same answer-change budget the GMAT enforces at the Pearson center.

Most mock exams are practice quizzes wearing a stopwatch. The Brightroom Mock Exam is the room before test day. This is what shipped today, what is still beta about it, and what the next eight weeks will tell us.

The shape

Three modes, one runner. Full Exam mirrors the live test — 64 questions across Quant, Verbal, and Data Insights, 135 minutes of working time, an optional ten-minute break the candidate places where they want it. Half Exam is the same architecture halved — 33 questions in 69 minutes, one shorter break. Section Exams isolate a single section for targeted conditioning when the schedule cannot accommodate a full sitting.

At /mock-exam/setup, candidates drag their section order, place the break, and watch the total clock update live. The order and break choices are the exact ones a candidate makes at the test center; the screen is the same screen.

The runner

The timer counts down per section, not globally. State changes at the same thresholds the live test announces — amber under five minutes, red under one — with the same ten-, five-, and one-minute milestone calls. The bookmark-and-review screen at the end of each section carries the same three-change budget, with the same item-edit list and the same "unflag" affordance.

At the end of each section, the review screen lists every question with three states: answered, flagged, blank. The candidate can edit three answers — no more — before the section closes and the section score commits. The three-change budget is the single rule that makes test-day pacing hard for most candidates; we built the runner around it before anything else.

The score report

At submission, candidates receive a scaled total in the 205–805 band, three section bands in 60–90, the percentile equivalents against the August 2024 GMAC table, and a per-section pacing trace that locates the minute the section started running tight. The score-band model is the v1.1 engine — same ability estimator, same Fisher- information selection — so a Mock Exam result reads on the same scale as the diagnostic and the in-product session history.

The score is an estimate, not a prediction. We will say more about the confidence interval on a Mock Exam result once the beta has produced enough completions to fit one honestly. Until then, the report carries a single readable interval and no point claim.

What is still beta

Three things are explicitly not finished.

Per-cohort percentile. The score report currently uses the public GMAC August 2024 percentile table. A per-cohort percentile against Brightroom Mock Exam takers specifically is more useful for benchmarking a candidate's progress against the people sitting the same surface — but it needs roughly 1,200 attempts before the figure stops being noise. We will swap in the per-cohort column when the beta clears that threshold.

Adaptive item selection inside the runner.The beta runner serves items from a curated, fixed pool per section. The v1.1 engine's Fisher-information selection runs in practice sessions but not in mock exams; we want the mock surface to mirror the live test, and the live test is curated, not adaptive, after the section starts. We will revisit if the candidate feedback says the fixed pool feels off.

Cross-attempt analytics. A candidate who sits three Mock Exams across a six-week prep window should see the trajectory across the three — pacing improvements, section-wise score lift, the sub-skill that kept costing them points. The cross-attempt panel is wired up but the visualizations are still iterating; the beta version is a table.

Who gets it

The beta is open to every Pro 1-month and Pro 4-month subscriber today, with no per-month attempt cap during the beta window. Candidates on the free diagnostic get one full Mock Exam attempt at no cost during the beta period; subsequent attempts are bundled into Pro. A standalone Mock Exam at $79 per sitting is available for candidates who want the rehearsal without the rest of the engine surface.

The beta closes when total completed attempts cross roughly 1,500. At that point the per-cohort percentile ships, the cross-attempt visualizations replace the table, and the beta label comes off. We expect that to happen before the year ends.

What we want from the beta

Two signals, in order of priority. First, whether the runner feels like the room — whether a candidate sitting the Full mode comes out of it saying they have just sat an exam, or they have just done a long practice set. That is the only thing the surface has to get right. Second, whether the score the runner produces is calibrated against the live test — which we can measure as the beta cohort sits the actual GMAT in the autumn and reports their official score back.

Both signals are visible only at the end of the beta window. We will write a longer note on the second when the autumn results land.

— Brightroom Product