The GRE's adaptivity happens exactly twice in the whole test - once between your two Verbal sections, once between your two Quant sections - and nowhere else. This is a materially different design from a question-by-question adaptive test: your performance on the first Verbal section decides whether the second Verbal section draws from a harder or easier question pool, and the same happens independently for Quant. Within a single section, every test-taker who sees that section sees a fixed set of questions - there is no live difficulty adjustment mid-section.
That single fact reframes a question a lot of aspirants get wrong: "does missing an early question tank my score?" On the GRE, no individual question carries that kind of weight - but the section you are handed second is shaped by how the first one went, and that has a real, bounded effect on the score ceiling available to you.
What "section-level adaptive" actually means here
Per ETS's own description of the current test structure, the GRE General Test since its September 2023 revision runs five sections in a fixed order: one Analytical Writing task first, then two Verbal Reasoning sections and two Quantitative Reasoning sections in either order. The first Verbal section a candidate sees is a fixed-difficulty section - identical in difficulty profile for everyone. Based on how many questions you get right in that first Verbal section, the second Verbal section is drawn from either a harder or an easier question pool. Quantitative Reasoning works the same way, independently.
This is a coarser mechanism than a fully adaptive test that recalculates after every single answer. There are effectively two difficulty "branches" available per measure (harder-second or easier-second), not a continuously shifting scale. It is also why a single careless mistake early on rarely moves your final score by much on its own - what matters is your accuracy rate across the whole first section, not any one answer.
Why the branch still matters for your ceiling
The practical consequence: a test-taker routed into the harder second Verbal section, who then performs well on it, can reach a higher final scaled score (130-170 per measure) than a test-taker who is routed into the easier second section and gets every question right there. The scoring algorithm accounts for the difficulty of the questions you actually answered, so getting the easier section is not simply "free points" - but it does cap how high your final scaled score in that measure can go, because the available question pool in the easier branch does not include the test's hardest items.
This is the actual reason test-prep guidance consistently tells GRE candidates to treat the first section of each measure as the one that matters most to get right, more than any specific question within it: your accuracy rate on that first section is what determines which branch - and therefore which score ceiling - you are working from for the rest of that measure.
What this does not mean
- It is not punishing to get a question wrong. There is no per-question penalty and no negative marking on the GRE at all - a wrong answer costs you nothing directly, it only contributes to the aggregate accuracy that decides the branch.
- It is not the same design as the old GMAT's question-by-question CAT algorithm, which recalculated difficulty after every single item. The GRE's adaptivity is coarser and happens only twice in the whole test - once per measure - which is a meaningfully different thing to prepare for.
- Time pressure late in a section is not a "the algorithm is testing you harder" signal. Within a section, every candidate who sees that section sees the same fixed question set - there is nothing dynamically escalating mid-section.
What this means for mock review
If you are logging full-length GRE mocks, the useful number to track by section is your accuracy on the first Verbal and first Quant section specifically, not just your overall percentage for the whole test - because that is the number the real exam's branching logic actually reads. A candidate who is strong in the second half of a section but weak and rushed in the first half is, on the real GRE, working against themselves in a way that a flat "I got 70% right overall" summary hides. Karma Yogi lets you log a full mock's sections separately and see the score trend per section over time, which is the shape of data this specific mechanic asks for - a first-section accuracy number, tracked across mocks, rather than one blended score.
Where this is uncertain
- ETS does not publish the exact scoring algorithm or the precise cutoff accuracy that determines which branch a candidate is routed into - the mechanism above (two branches, decided after the first section of each measure) is publicly documented by ETS in general terms, but the specific mathematical thresholds are proprietary.
- The exact percentile impact of landing in the harder vs. easier second-section branch is not something ETS states as a fixed number - it depends on your performance in that second section too, not just which branch you are on.
Sources
- ETS - GRE General Test Content and Structure: ets.org/gre/score-users/about/general-test/content-structure - section order, section-level adaptivity, scoring range.
- ETS - Shortened GRE Test announcement (September 2023 format change): ets.org/news/press-releases/shortened-gre-test-coming-september-2023.
Karma Yogi's GRE tracking lets you log a full mock's Verbal and Quant sections separately, so a first-section accuracy dip shows up in your own data instead of getting averaged away. Start tracking your GRE prep free.
End of essay
- Anish Guruvelli