GMAT Focus, day one.

The GMAT Focus Edition began administering yesterday. Eight months from announcement to first ink at a Pearson VUE center, two months of registration overlap, and the older test still running until the end of January. This is the first day of the new shape of the exam.

We have spent the last several weeks reading the published specification. This brief is what came out of it.

What the spec actually says

Three sections, each 45 minutes, each scored on a 60–90 band in 1-point increments. Twenty-one Quant, twenty-three Verbal, twenty Data Insights. Total runtime — including the optional ten-minute break a candidate can place where they want it — is two hours and fifteen minutes. The composite score is reported on a new 205–805 scale, ending in five, so that no Focus score can be confused with a legacy one.

The Analytical Writing Assessment is gone. Sentence Correction is gone. Integrated Reasoning has been replaced by Data Insights, which folds in multi-source reasoning, table analysis, graphic interpretation, two-part analysis, and data sufficiency. Data sufficiency — historically a Quant item type — now lives inside DI. That is the change with the largest downstream implications for prep, and the one most of the explainer coverage has missed.

Bookmark and review

The piece of the specification we find most interesting is the new review mechanic. Inside each section, a candidate can flag any question, navigate back to it before the section ends, and change up to three answers per section. Three is the budget. Re-selecting the same answer does not consume a slot.

This is a small rule with large effects. It rewards a candidate who can identify, in real time, which of their early-section answers were marginal and which were confident. It rewards a candidate who has practised under the rule, not against a stricter or looser version of it.

Most existing mock-exam software does not implement this mechanic at all, or implements it incorrectly — either by allowing unlimited changes or by not allowing review at all. We expect this to be the single largest source of test-day surprise for candidates moving from the legacy test, and the single largest gap between published mocks and the live experience for at least the first year.

Data Insights is the section to watch

Of the three sections, Data Insights is the one that has moved the most relative to what candidates have practised for. It is also the section the prep industry has the least calibrated material for, and the section that — in our read of the public scoring algorithm — carries genuine weight in the composite score.

We expect three things to become true about DI over the next twelve to eighteen months. First, that it will carry more weight in the composite than the early prep coverage assumes. Second, that multi-source reasoning will be the load-bearing DI sub-skill at the top of the band. Third, that retired official mocks will not yet exist in enough volume to calibrate against — making DI prep, for the first cohort, materially under-served.

We are writing those down so we can come back to them in a year and check.

What we are doing with this

We are building Brightroom against the Focus specification rather than retrofitting from a legacy stack. Three consequences of that decision.

First, the item bank is being authored on the new sub-skill taxonomy from the start. We are not converting Sentence Correction items into something else. The category is gone; the prep should reflect that.

Second, our mock-exam runner will enforce the actual bookmark-and-review rule. Three changes per section, with re-selection of the same answer not counting against the budget. The candidates we work with should arrive at the test center having already trained under the constraints they will be under.

Third, Data Insights will not be the section we ship last. Of every existing prep platform we have audited, DI is the section that loads last in the curriculum, the section with the thinnest item bank, and the section with the most recycled IR content with a relabelled section header. We intend to do the opposite.

What to do if you are sitting Focus in the next six months

We are not yet shipping a product to recommend. What we can offer is the read of the spec. Three things.

Practise on the bookmark mechanic, even badly.Mock software will catch up. In the meantime, if your mock software allows unlimited review, impose the three-change budget on yourself. If it does not allow review at all, do a timed first pass and then a second pass with the timer still running.

Spend a real week on Data Insights early.Not at the end. The section is harder to retrofit than Verbal or Quant, and most prep stacks under-serve it for the reasons above.

Re-baseline. A diagnostic taken on legacy material is not your Focus diagnostic. The score scale, the section structure, and the time budget have all changed. Anything from before November is reference, not baseline.

Closing note

We will publish a brief like this whenever the underlying shape of the exam moves enough to change how a serious candidate should prepare. The next one will be in late January, after the legacy test retires, with whatever we have learned in the first three months of Focus administrations.

— Brightroom Research