McKinsey Solve in 2026: Redrock, Sea Wolf & What Changed

If you were invited to the McKinsey Solve assessment and your research keeps surfacing the Ecosystem game — building food chains from 39 species — much of what you are reading is out of date. Candidates in 2025–2026 consistently report a different lineup: the Redrock Study (data-analysis mini-cases) and Sea Wolf (a constraint-optimization puzzle), with a third module, the Sustainable Futures Lab, appearing on longer invitations. The older Ecosystem Building game has been phased out of the standard rotation, though some regional or role variants may still include it.

McKinsey itself publishes very little about the format. Its official careers page describes Solve only as "a gamified assessment created to showcase your problem-solving abilities" and says there is "no need to prepare — approach it with curiosity and creativity." Everything more specific below comes from candidate reports and preparation sources reviewed in July 2026, and is labeled accordingly. Your invitation email is the final word on which modules and time limit you will get.

The short answer

• Most invitations in 2025–2026: Redrock Study + Sea Wolf, about 65 minutes total.
• Longer invitations: Redrock + Sea Wolf + Sustainable Futures Lab, about 85 minutes.
• Ecosystem Building: retired from the standard rotation; still reported occasionally in some regions and variants.
• Scoring: candidates and prep sources consistently describe two layers — a product score (your answers) and a process score (how you worked: what data you opened, pacing, revisions). McKinsey does not publish a pass mark.
• Placement: after the application screen and before first-round case interviews, for most consultant-track roles.

What is the McKinsey Solve assessment?

Solve is McKinsey's online gamified assessment, introduced in 2020 as the replacement for the paper-based Problem Solving Test (PST). McKinsey receives on the order of a million applications a year, and Solve is the main screening layer between the resume review and interviews for consultant-track candidates. It runs unproctored on your own computer (desktop or laptop only — not a phone or tablet), with untimed tutorials before each game.

Two official constraints are worth taking seriously. First, McKinsey's integrity rules prohibit using "applications, websites, AI tools, or pre-prepared notes" during the assessment, and prohibit recording or screenshotting it — violations can lead to exclusion or a rescinded offer. Second, McKinsey states results are "considered in conjunction with the rest of your application," so Solve is weighed alongside your resume rather than acting as a lone pass/fail gate.

The current games

All timings are candidate-reported patterns, not official McKinsey figures.

Redrock Study: the module candidates find hardest

Redrock is a research-study simulation — a wildlife-population scenario (wolves and elk on an island reserve is the classic theme, with variants) — followed by a set of standalone mini-cases. Candidate accounts converge on a consistent structure:

1. Investigation. You read the study brief and drag relevant data points into a "Research Journal." Reports suggest only a small fraction of the available numbers actually matter — collecting selectively beats hoarding.
2. Analysis. Timed calculations using an in-game calculator: ratios, percentages, growth rates, weighted averages, percentage points versus percent. Once you leave a section, you generally cannot return to it.
3. Report. You complete a pre-written summary with your computed numbers and choose the chart that best presents the findings.
4. Mini-cases. A set of short, self-contained quantitative questions — commonly reported as five to seven — that are thematically related but do not reuse the study data. This is where time pressure peaks: most first-person accounts describe Redrock as "a stressful math test in a game skin."

The skills are ordinary consulting numeracy under a clock: fast, accurate percentage and ratio work, reading charts, and knowing when a precise calculation matters versus when an estimate is enough. If your mental math is rusty, timed arithmetic drills are the highest-leverage preparation — that skill transfers directly, whatever the scenario skin. Our Numerosity simulator trains exactly that fast-arithmetic-under-time-pressure pattern (it simulates a HireVue game, not Redrock — but the math reflex is the same muscle).

Sea Wolf: the optimization puzzle

Sea Wolf (sometimes called the Ocean Cleanup game) rolled out widely from 2024. Candidate descriptions agree on the shape: across three ocean sites, you select a small team of organisms — commonly reported as three microbes per site from a pool of about ten — so that the team's attribute averages fall inside site-specific ranges, at least one member carries a required trait, and none carries a forbidden one. Later rounds add interactions between species, so a choice that looks locally fine can break a constraint elsewhere.

It is genuinely a constraint-satisfaction puzzle: read the requirements first, compute averages before committing, and check every constraint — one reported scoring pattern weights each satisfied constraint roughly equally, so a single overlooked forbidden trait costs as much as a badly missed average. Treat that specific scoring claim as candidate-reported, not official.

Sustainable Futures Lab: the newest module

The Sustainable Futures Lab (SFL) began appearing on invitations in phased rollouts from early 2026 and is reported mainly on the longer ~85-minute invites. Unlike the other two, it is a situational-judgment module, not a math or puzzle game: one continuous project storyline — typically sustainability-themed — with around 13 questions, mostly scenario-based multiple choice plus a ranking exercise. Reports say it probes prioritization, decision-making under uncertainty, interpreting incomplete information, and balancing trade-offs.

There is no calculation to drill here. The honest preparation is understanding the format and answering consistently, the same advice that applies to any judgment-based assessment: there is no published "correct personality," and second-guessing what McKinsey wants usually produces less coherent answers than considered, consistent ones.

How Solve is scored

No official rubric exists, but two-layer scoring is one of the most consistently reported facts about Solve:

• Product score — did you get the right answers: calculations, selections, constraint satisfaction.
• Process score — how you got there: which data you opened, how long you spent reading versus acting, pacing, and revision patterns.

The practical consequence: random-clicking your way to a few lucky answers scores worse than slightly-less-accurate but systematic work. Read first, act deliberately, and do not thrash. Candidate reports also say the platform notices tab-switching and external tools — which, given McKinsey's explicit integrity rules, you should avoid anyway.

There is no published pass mark or pass rate. Prep-source estimates of the advance rate cluster around 20–30%, but treat that as folklore-grade: nobody outside McKinsey can verify it.

How Solve compares with other game-based assessments

Solve sits at one end of the game-assessment spectrum. Vendor batteries like pymetrics measure behavioral traits through many short neuroscience-style tasks with no right answers; HireVue games test cognitive skills like working memory and mental arithmetic in two-to-three-minute bursts. Solve is closer to a timed case exam wearing a game interface: it has objectively right answers, rewards calculation speed and structured reading, and runs over an hour. If you are preparing for several employers at once, our game-based assessments guide maps which vendors work which way.

We do not currently offer a Solve simulator. This guide exists because the format changed substantially in 2024–2026 and much published advice still describes the retired Ecosystem game. What transfers from practice on our platform: timed mental arithmetic (Redrock's core demand), systematic constraint checking, and calm pacing under a visible clock.

Preparation that actually helps

1. Rebuild timed arithmetic. Percentages, ratios, growth rates, weighted averages — computed quickly without a calculator crutch, since the in-game calculator still costs time. A week of short daily drills beats a weekend cram.
2. Practice reading data selectively. Redrock punishes hoarding. When you practice with any chart-heavy material, force yourself to state what question am I answering before touching the numbers.
3. Walk through the current formats. Know the Redrock section flow (you cannot go back), the Sea Wolf constraint types, and the SFL format before test day, so no attention is wasted on interface surprise.
4. Set up properly. Desktop or laptop, stable connection, ~90 minutes blocked, no interruptions. Contact McKinsey's Solve support immediately if the platform crashes mid-run.
5. Ignore stale guides. Anything centered on Ecosystem food-chains or Plant Defense describes an assessment most 2026 candidates will not receive.

FAQ

Is the Ecosystem game still part of McKinsey Solve?

Mostly no. Candidates in 2025–2026 overwhelmingly report Redrock + Sea Wolf as the standard pairing, and Ecosystem has been phased out of the normal rotation. A minority of reports still mention it in some regions or variants, so check your invitation and complete the tutorials either way.

How long does the Solve assessment take?

Reported invitations are about 65 minutes (Redrock + Sea Wolf) or about 85 minutes (plus the Sustainable Futures Lab). Block at least 90 minutes for loading, tutorials, and transitions.

Can I retake Solve if I do badly?

Not within the same application cycle. Candidates report needing to wait for a new application — commonly 12 months or more — before another attempt. McKinsey does not publish an official retake policy; ask your recruiter if it matters to your situation.

Is there a passing score for McKinsey Solve?

No published one. McKinsey says results are considered together with the rest of your application. Prep-source pass-rate estimates (~20–30%) are unverifiable.

Do I need to prepare if McKinsey says no preparation is needed?

McKinsey's "no need to prepare" is true in the sense that there is no syllabus. But Redrock is a timed numeracy exercise, and timed numeracy is trainable. Familiarity with the current formats and fast arithmetic demonstrably reduces the two biggest score-killers: interface surprise and clock panic.

Does McKinsey use HireVue or pymetrics instead of Solve for some roles?

Solve is McKinsey's own assessment and the standard for consultant-track screening. Some non-consulting or regional processes differ. McKinsey does not use pymetrics — if another employer sends you pymetrics games, that is a different assessment with different preparation.