
There is no published cutoff for SOVA or TestGorilla, and the pass rate estimates in circulation range from 30 percent to 70 percent. Here is why they disagree, what each platform actually measures, and the bar worth planning against.
Bain does not publish a passing score for either assessment platform, does not release your result, and does not confirm a cutoff to candidates or to prep providers. Every number you will read on this topic, including the ones on this page, is inferred from candidate reports and coaching data rather than taken from Bain.
What can be said with more confidence is the shape of the thing. Your result is not compared against a fixed mark you either clear or miss. On SOVA, cognitive sections are ranked in percentiles against a norm group of other applicants, so the effective bar moves with the pool. On TestGorilla, section level results are assembled into a picture that recruiters read alongside your CV rather than as a standalone gate.
Search this question and you will find confident figures that contradict each other by more than thirty percentage points. They are not all wrong. They are mostly measuring different things.
| Figure reported | Source and date | What it appears to measure |
|---|---|---|
| 50% to 70% pass | Hacking the Case Interview, Bain acceptance rate breakdown, 2026 | Share of everyone who sits the assessment and advances, across all offices and candidate types |
| 30% to 35% advance | Hacking the Case Interview, Bain aptitude test guide, 2026 | Share of test takers reaching interviews, which folds in screening that happens after the test |
| 70% to 75% cutoff | MConsultingPrep, Bain online test overview | A score threshold candidates report, not a pass rate. A different quantity entirely |
| 85% to 90% eliminated | PrepPartner, MBB recruitment odds, 2025 | The resume screen plus assessment combined, which is why it looks so much harsher |
Three things are being conflated across those rows, and separating them is most of the work.
Which assessment you sit changes what "a good score" even means. Switch between them below.
Your raw accuracy on the cognitive sections is converted into a percentile against other applicants. A section that felt brutal may still score well, because everyone else found it brutal too.
Because the three cognitive sections are read together, an even performance beats a spike. A collapse in one section drags the composite down further than a strong section lifts it. Full detail in the SOVA guide.
No published section scoring and no disclosed cutoff. What is consistently reported is that accuracy is weighted more heavily than speed, and that recruiters review the result as one input rather than as a gate.
Since nothing is percentile ranked back to you, the only target you can act on is operational: finish every block, keep accuracy intact across all of them, and do not leave an item blank. A candidate who answers all twelve numerical items at 80 percent accuracy is in better shape than one who answers eight at 95 percent, because the four unreached items count against them.
Timing and block structure are broken down in the TestGorilla guide.
For most candidates the assessment is not where the application dies. The resume screen in front of it is far more selective, and in several offices the assessment is read together with the screen rather than after it.
That reframing is worth holding onto. A strong assessment result does not rescue a weak screen, and a merely adequate one is often survivable when the rest of the file is strong. The assessment is a filter on the already filtered.
Stage by stage figures are reported by coaching providers rather than published by Bain. See the sources at the foot of this page.
The reason a single passing score does not exist is that the assessment is one reading among several, positioned differently by office. Seeing the whole funnel makes the stage easier to size.
Two implications follow. The assessment is a smaller filter than the screen that precedes it, so a candidate who received a link is already through the hardest cut. And because several offices read the assessment alongside the first round rather than strictly before it, a middling result does not always end the application.
Since you cannot see your result, cannot learn the cutoff, and cannot influence the cohort you are ranked against, the only useful targets are ones you can verify yourself before test day.
Evenness matters more than a peak. On a scored mock, your weakest section is the one deciding your composite, so that is where preparation time belongs.
On TestGorilla an unreached item scores as wrong. On SOVA a slow crawl costs you on the speed component. Finishing is not a nice to have, it is part of the score.
Around 45 to 60 seconds per numerical item and 30 to 40 per logical item. Pick the pace in advance and practice at it, rather than discovering it during the real sitting.
The work style block runs a consistency check across repeated themes. A constructed profile fails it, and the failure is not recoverable in the room later.
Confirm from the invitation email whether you are sitting SOVA, TestGorilla, or a video component, before you spend any preparation time at all.
Bain recruits office by office, and a large office receiving thousands of applications applies a harder cut than a smaller one. The same result can pass in one city and not another in the same cycle.
Percentile ranking means you are measured against whoever applied when you did. A strong year raises the bar without anything about the test changing.
Some offices use it as a hard filter before interviews. Others read it alongside the resume and the first round, which makes a borderline result survivable.
Bain is reported to run a points based screen across academics, experience and extracurriculars. A strong file buys tolerance for an average assessment, and a weak one does not.
Most candidates finish convinced they did badly, because there is no feedback of any kind: no score, no percentile, no section breakdown, and often a wait of one to three weeks before you hear anything. The feeling is not evidence.
Two things are worth knowing. You cannot retake the assessment within the same application, and Bain does not release your result on request. If the application is unsuccessful you would reapply in a later cycle, with the waiting period set locally by the office, so confirm it with the recruiter rather than assuming a standard.
Every figure on this page is reported by coaching providers or candidates rather than published by Bain. Dates matter here, because the assessment mix has changed since 2024.
Full simulators for both Bain assessment formats, built to the real rules and timings. The first section of each is free.