Pymetrics is now a Harver product. Harver announced the completed acquisition in August 2022, and the games are still sold: Harver’s own site lists pymetrics as its game-based assessment. So if an employer has asked you to take “the pymetrics test”, you are taking a Harver assessment that still carries the pymetrics name.
That distinction matters, because much of the advice written about pymetrics predates the acquisition and describes a standalone company that no longer exists. What has not changed is the shape of the thing: a set of short browser exercises, around 25 minutes in total, that measure behavioral traits rather than knowledge.
What happened to pymetrics?
pymetrics was founded as a standalone assessment company and built its reputation on using short, neuroscience-based games in place of a questionnaire. Harver, a hiring platform, announced the completed acquisition in August 2022, and pymetrics’ founder framed it at the time as a chance to rebuild how talent decisions are made across the hiring process.
Since then the product has sat inside Harver’s assessment range, alongside its personality questionnaire, cognitive ability testing and situational judgement tests. Harver has kept the pymetrics name for the games specifically, so both names are correct in 2026: the assessment is pymetrics, and the company behind it is Harver.
There is one practical consequence. If you go looking for pymetrics as a vendor, or for a pymetrics login or support page, you will end up at Harver. Treat the invitation email from the employer as the authoritative instruction on where to sit the assessment and by when.
What are the pymetrics games?
Harver describes the assessment as more than a dozen interactive, gamified exercises taking around 25 minutes to complete on average. They are short, played in a browser, and deliberately nonverbal: Harver’s stated design goal is that the tasks are intuitive and minimize cultural and language bias, so there is very little reading involved.
There is no syllabus and nothing to revise. The exercises are not testing what you know. They record how you behave while you play, and those behaviors are mapped onto the trait set below.
Because they record behavior rather than answers, there is no score in the usual sense and no published pass mark to aim at. What exists instead is a comparison: the employer configures what the pattern looks like for people who do the role well, and your profile is compared against that pattern. Two candidates can produce equally strong profiles and match different roles.
What the pymetrics games measure
Harver publishes the trait set the games are built to capture:
- effort
- attention
- focus
- learning
- decision making
- risk tolerance
- fairness
- generosity
- emotion
Read down that list and something should stand out. Several of these are traits where there is no universally good answer. High risk tolerance suits some roles and rules you out of others, and the same goes for how quickly you commit to a decision. That is the single most useful thing to understand before you start, and it is why trying to “ace” a pymetrics assessment is the wrong frame: there is no direction that is better in general, only a direction that matches a given role.
Harver states that the assessments are validated for fairness across gender, ethnicity and socioeconomic status.
Can you practice for pymetrics games?
Not in the way you can practice a numerical reasoning test, and it is worth being skeptical of anything that promises otherwise. There is no question bank behind the games and no technique that produces a better trait profile.
What genuinely helps is narrower:
- Remove the friction. Use a laptop or desktop rather than a phone where you can, a stable connection and an uninterrupted half hour. Timing and reaction speed carry real signal in some of the exercises, so lag and interruptions cost you accuracy on the measurement itself.
- Read each instruction screen before you start playing, not while you play. Every exercise has its own rule. Spending the first rounds working out what the exercise wants is the most common avoidable problem, and those rounds still count.
- Answer as yourself, consistently. Traits are inferred from patterns across the whole set rather than from any single exercise. Playing a character you think the employer wants tends to produce an inconsistent profile rather than a flattering one.
- Get used to being assessed on screen against a clock. That familiarity is transferable, and it is the only part of the experience you can really rehearse.
For the last of those, our cognitive ability tests and personality tests put you in the same situation: timed, unfamiliar, and on screen with someone reading the output. They will not resemble the games, but they take the novelty out of the format, which is where most of the avoidable nerves come from.
What happens after you play
Harver says every job seeker who completes the assessment receives immediate results and feedback. That is unusual at a screening stage, and it is worth reading rather than closing, because it is a description of how you approached the tasks that you can use in the interview that follows.
What the employer does with the result is the employer’s own process and it varies. Some run the assessment as an early screen before a human reads your application, others later, alongside interviews. If it matters to you where it sits, the recruiter is the right person to ask.
For more on the publisher, see our guide to pymetrics, and for a comparison with another game-based assessment you may meet in the same process, our guide to Arctic Shores.
