Talent Assessment Tools: 14 Compared on What They Really Cost

Most lists of talent assessment tools are written by companies that sell talent assessment tools. We read eight pages ranking for this keyword before writing anything. Four of them are vendor blogs that put themselves in the number one slot: Truffle, Metaview, Testlify and Xobin all rank their own product first on their own list. Only two of the eight publish a methodology section explaining how they picked.

We are a vendor too. SharpAssessment sells practical and job-knowledge tests to small teams, priced by candidate volume, and we have kept ourselves off this list rather than award ourselves first place on our own page. That does not make us neutral, so here is the safeguard instead. Every price and product claim below comes from the vendor’s own public pages, linked so you can check it. Where we quote what buyers say, we label it as a practitioner report from Reddit, because that is a different grade of evidence and one negotiated price from 2022 is not a price you can go and buy today.

Doing that turned up something none of the eight pages mention. Six of these fourteen vendors publish no price at all. And the eight that do publish use four different billing units between them, which is enough to make the tidy comparison tables you have been reading unreliable.

What we checked and what we did not

We fetched the homepage, pricing page and product or library page for each of the 14 vendors in late July 2026, and pulled every number from those pages. Where a vendor’s own pages disagreed with each other, we say so instead of picking the flattering number.

We did not run the tests, hire anyone through them, or measure whether they predict job performance. We also could not verify the hands-on testing other lists claim, including one that says it road-tested 30+ platforms. What we can verify is what a buyer can see before signing, which is exactly the thing that gets hidden.

The short version

Six of fourteen publish nothing

Criteria, eSkill, Wonderlic, Vervoe, Harver and Korn Ferry Assess show no dollar figure anywhere on their sites. Not a tier, not a range, not a “starting at”. Every path ends at a demo request.

Wonderlic is the most explicit about it. Its own FAQ says pricing depends on the package you choose, which is based on your organisation’s needs and employee count. Harver has no pricing page at all: /pricing, /plans and /get-a-quote all return 404. Vervoe’s pricing page now shows a single Enterprise plan with a “Let’s Talk” button, a real change in posture for a product that buyer guides still describe as accessible to smaller teams.

Three of the eight that do publish only show you the bottom rung. Prevue prices its Essentials plan from $195 per month and hides Standard and Premium behind “Request Pricing”, and Premium is where ATS integrations live. Bryq publishes $69 per month for companies of 1 to 15 employees and nothing above that band. The Predictive Index publishes a $10,000 floor for Talent and quotes Performance and Engagement privately.

This matters more for small teams than for enterprises. A 40-person company that shortlists five quote-only vendors needs five sales calls before it can compare a single number, and the vendors know a call converts better than a price tag. If your evaluation window is two weeks, that half of the market is effectively closed to you.

Everyone charges for a different thing

This is the part that breaks comparison. Read the unit column above again.

• Per candidate, with unlimited tests inside that candidate’s session: TestDome, HR Avatar, Evalart.
• Per completed test: EmployTest. One credit equals one finished test, so a candidate who takes three tests costs three credits. Its $199 starter pack is 5 credits, which can mean fewer than 5 candidates.
• Per credit, where the credit cost varies by what you run: TestGorilla charges 1 credit for a skills test and 2 for an AI interview, and does not refund credits when a candidate abandons mid-assessment.
• Per company headcount, regardless of how much you hire: Bryq, Prevue, The Predictive Index. Cheap if you hire constantly, expensive if you hire twice a year.

Two products can advertise similar monthly numbers and still differ sevenfold on your actual bill. Take a 30-person company making 6 hires a year and testing 20 candidates per hire: 120 candidate sessions. Evalart works out at $456 if you concentrate them into six paid subscription months, or $912 if you keep the subscription running all year. TestDome is $1,400, buying a 100-pack and a 25-pack. HR Avatar is $3,360, buying a 100-pack and two 10-packs, which is its cheapest published route to 120 credits. Bryq and Prevue are flat subscriptions that do not move with those 120 candidates at all, which is the whole point of them.

Work out your own two numbers before you open any pricing page: candidates tested per year, and tests per candidate. Without them, none of these prices mean anything.

Tools that publish real prices

Evalart

The cheapest published per-candidate rate we found. Evalart sells either a subscription ($76 a month for 20 candidates, $3.80 each) or a one-time pack ($90 for 20, valid 12 months), and both include unlimited tests per candidate. Credits are refundable if a candidate never starts. Coding tests, aptitude, DISC and Big Five are all in the library.

Two honest caveats. Its own pages disagree on library size: the homepage says 500 tests, the pricing page says 250 or more. And the custom test builder, which the homepage implies exists, is labelled “Coming soon” on the features page. No named ATS integrations, only a generic API. Best fit: a small team that wants low cost per candidate and can live with the stock library.

Bryq

Bryq publishes $69 a month, $828 a year, for companies of 1 to 15 employees, with unlimited invitations. That is the flat-rate option in this list. It measures cognitive ability, personality (16PF and OCEAN), skills across a stated 140+ test library, and a newer AI fluency assessment. Self-serve ATS integrations are named outright: Workable, Greenhouse, Lever, Teamtailor, BreezyHR, Recruitee, Pinpoint, Ashby, JazzHR and Zoho Recruit among others.

Caveats, and they matter. That $69 is an annual rate: the page states a 1-year commitment, and the only button on the plan is “Book a demo”, so this is a published price rather than a self-serve checkout. The homepage advertises a “Starter” plan that does not appear on the pricing page at all. Anything above 15 employees is a custom Enterprise quote. And several capabilities including SSO and validated hiring are add-ons on top, so the $69 headline understates a loaded setup.

TestDome

TestDome has the most transparent pricing page in the category: five packs, all priced, no subscription. $20 per candidate at 5, $16 at 25, $10 at 100, $8 at 300, $7 at 600. Packs never expire and an invite that goes untaken for 6 days returns to your balance. There is a 14-day trial.

Its real strength is anti-cheating, which it treats as the product rather than a feature list: screen proctoring that detects AI tools, webcam analysis, a locked-down exam browser, questions written to resist ChatGPT, and duplicate-attempt detection. The catch is what sits behind Enterprise: ATS integrations, API access, SSO, multi-user accounts and candidate anonymisation are all excluded from the five self-serve packs. A small team that wants TestDome inside its ATS has to leave the published pricing behind.

TestGorilla

The default answer in this category and the most-discussed by a distance. TestGorilla has a genuine free tier: 10 credits a month, no card, 5 essential tests. Paid starts at $142 a month on an annual commitment ($1,704 up front) for the full library of 350+ tests, and Plus starts at $400 a month ($4,800 a year) for custom tests, coding challenges, job simulations, ATS integrations and API access.

Two things buyers report that the site will not tell you. In one detailed Reddit thread a customer said scores come back unweighted, so you cannot tell the platform that skill A matters more than skill B for a role. In the same thread a candidate described withdrawing an application rather than sit a 55-minute camera-on assessment. Both are single reports, not a survey, but they name the failure modes worth testing in a trial. Credits are also consumed the moment a candidate starts and are not returned if they drop out. We go deeper on this in our TestGorilla alternatives comparison.

EmployTest

EmployTest sells credit packs with no contract: $199 for 5, $299 for 10, $599 for 25, $1,999 for 100, valid a year, with a 30-day guarantee. Its strength is the unglamorous middle of the market, office and clerical roles: Microsoft Office tests delivered as interactive simulations rather than multiple choice, typing and data entry, attention to detail, plus industry sets for legal, healthcare, accounting and warehouse work. It states that its tests follow EEOC and Department of Labor selection guidelines.

Two gaps. Credits are per completed test, not per candidate, so budget accordingly. And no ATS integration was mentioned on any of the three pages we fetched, which is unusual in this group. Results land in a dashboard and stay there.

HR Avatar

HR Avatar prices by candidate credit: $50 for one, $350 for ten, $2,660 for a hundred, valid 12 months. One credit covers unlimited tests, video interviews and reference checks for that candidate within 90 days, which makes it good value when you want several signals on one person. The free trial gives 14 days and 2 credits with no card. Tests are selected by job title and bundle cognitive ability, personality, a scored essay, behavioural history and job knowledge into one package.

Caveats: no test-library count published anywhere, no named ATS integrations on any page we fetched, and no EEOC, adverse-impact or bias-audit language on its marketing pages. At $50 it also has the highest published per-candidate rate in this list, though the single-credit pack means the smallest possible first purchase is cheaper here than at TestDome or EmployTest.

Prevue

Prevue starts at $195 a month for Essentials, priced by employee-count band, with unlimited candidate invites and a library of 100+ job-position benchmarks. It is grounded in I/O psychology, cites 35 years of research and 250,000+ custom benchmarks, and makes the clearest adverse-impact claim in this group: it states every assessment is monitored for adverse impact across gender, ethnicity and age, and that it is built to SIOP, APA, EEOC and Canadian Human Rights Act standards. It also ships its own applicant tracking system.

The catch is tiering. ATS and HRIS integrations plus API access require the top Premium plan, and neither Standard nor Premium shows a price. Cost also tracks your headcount rather than your hiring volume.

The Predictive Index

PI publishes what almost nobody else does at the top of the market: a real floor, $10,000 a year for the Talent plan, scaling with headcount, unlimited users and unlimited assessments. It is not a test library. It is two instruments, a 6-minute behavioural assessment and a cognitive assessment, plus a job assessment that defines the target profile.

Its compliance language is the most specific we found: it states the behavioural assessment complies with EEOC guidelines and is validated against the four-fifths adverse impact standard. Note two inconsistencies on that same science page, which we are reporting rather than resolving: it gives the independent audit cadence as both every five years and every year, and the validation-study count as both 383+ and 700+. SSO is a paid add-on at every tier.

Tools that quote only

Criteria

The most cited alternative when people leave TestGorilla, and one of the strongest science stories in the category: cognitive aptitude (the CCAT), personality, emotional intelligence, risk, skills and game-based assessments, with an exclusive licence on Harvard’s MRAB and an ISO 42001 certification claim for AI management. There is a free trial.

Three things to weigh. Pricing is entirely quote-based across all three plans. AI proctoring and AI features are separately-priced add-ons rather than included. And the site’s own scale statistics do not agree with each other: the homepage says 80M+ assessments and 20+ years of talent science, the assessments page says 60M+ and 15+. Separately, an HR practitioner writing about the CCAT described it as a poor candidate experience that forces larger pipelines to absorb refusals.

eSkill

A deep, customisable library aimed at role-specific hard skills: Microsoft Office and Excel, maths and bookkeeping, clerical and typing, mechanical aptitude. It sells three plans, Growth, Scale and Custom, and every one of them ends at “Talk to sales”. We found no free-trial offer on any eSkill page we fetched.

Its own pages give two different library sizes, 600 subjects on the homepage and 800 on the skills page, against a consistent 70,000+ question bank. It claims standard ATS integrations as a plan feature without naming a single ATS on its main pages.

Wonderlic

Wonderlic advertises “85 years of proven science”, and its design is genuinely different: rather than a catalogue, Wonderlic Select is one assessment combining cognitive ability, personality and motivation into a single score against off-the-shelf job profiles. It integrates with Greenhouse, JazzHR and iCIMS, requires no paid certification to interpret results, and states it is validated, fairness-reviewed and aligned with US employment law.

Everything else is opaque. No pricing, no test count, no self-serve signup, and implementation measured in weeks rather than the same afternoon. The pricing complaints are specific and consistent: one long-term customer described a negotiated rate near $4,500 rising to $6,000, and went shopping for alternatives, which is roughly how most small buyers meet Criteria and Prevue.

Vervoe

The strongest work-sample story here. Vervoe builds role-specific simulations, spreadsheet tasks, coding challenges, presentations and video responses, and grades them with AI it says applies human grading logic at 80% accuracy. It is ISO 27001 certified and states its AI is independently audited for fairness by Holistic AI with results openly reported, which is a meaningful commitment.

But its pricing page now shows exactly one plan, Enterprise, with no number of any kind. Its assessment library size is never stated. And its headline outcome statistics, 67% fewer interviews per hire and 90% less likely to leave among them, are given without a defined baseline or methodology link, so read them as marketing rather than evidence.

Harver

Built for volume hiring, BPOs, contact centres and retail, not for a 30-person company making six hires a year. It offers 450+ assessment options across behaviour, cognitive ability, situational judgement and simulations, plus 900+ pre-built job profiles, and owns pymetrics for game-based assessment and Checkster for reference checking.

Its compliance page is the best in this list by some margin: SOC 2, ISO 27001, GDPR and CCPA, and named independent bias audits under NYC Local Law 144 conducted by BABL AI for both the Harver platform and pymetrics, with the reports published. If you hire in New York City and use automated tools, that is the standard to compare everyone else against. Against it: no pricing page, no free trial, no self-serve path, and outcome statistics that differ between its own homepage and its assessments page.

Korn Ferry Assess

The enterprise consulting end of the category. Assess is one module in a ten-module talent suite, built around Success Profiles and Korn Ferry’s Leadership Architect competency framework, and it runs inside Workday Recruiting through the paired Korn Ferry Select product.

It ranks first in Google for this keyword, which is worth noting given how little its page actually commits to. There is no test count, no named test types, no pricing, and no EEOC, adverse-impact or Local Law 144 language anywhere on it. For a small team this is not a realistic option, and the page does not pretend otherwise.

How to choose without burning a quarter on it

1. Work out your two numbers first. Candidates tested per year, and tests per candidate. They decide whether per-candidate or flat-rate pricing wins, and nothing else in the evaluation matters as much.
2. Send the test after a human conversation, not before it. This is the most consistent piece of advice from recruiters who post about assessments, and it costs nothing. A test as the first gate is what produces the abandonment described below.
3. Cap it, and say the cap out loud. Publish the time commitment in the invitation. For high-volume and hourly roles, keep the whole thing under 15 minutes and open it on a phone before you send it to anyone.
4. Check whether scores can be weighted by role. One TestGorilla buyer reported that they could not. If skill A matters more than skill B for the role and the tool cannot express that, you are reading a number that averages away the thing you care about.
5. Ask what happens to the data. Specifically: is proctoring included or an add-on, is there a published bias audit, and can you export the raw responses.
6. Do not buy on library size. Three of these vendors publish two different library numbers on their own website: eSkill says 600 subjects on one page and 800 on another, Evalart says 500 and 250+, TestDome says 100+ skills and 130+. It is the least reliable figure in the category and the least connected to whether the tool works for the one role you actually need to fill.

If most of your hiring is for practical, job-related roles, the narrower question of skills assessment software is usually the better place to start than the whole talent-assessment category.

What none of these tools fix

Assessment vendors sell better decisions. Read what practitioners report and a narrower picture appears.

Qualified people quit the process. In our sample of 41 Reddit threads this was the joint most frequent complaint, tied with scepticism about personality testing. Candidates describe abandoning applications when a long unpaid test appears, and recruiters describe filling larger pipelines to absorb the refusals. No feature on any pricing page addresses this. Shortening the test and moving it later does.

Some buyers measure no effect. One recruiter described tracking assessment scores against hire outcomes, finding little correlation, dropping the tool and saving the money. That is a single account rather than a study, but it is the right experiment and it is worth running before you renew.

Personality testing carries a credibility problem. Among HR professionals discussing it publicly, personality instruments get called “the astrology of HR”, and MBTI in particular is widely rejected for selection, a position the Myers-Briggs Company itself takes. Job-knowledge and work-sample tests are easier to defend, scientifically and legally, which is why they are the sensible starting point. We cover the split in more detail in personality tests for hiring.

Legal exposure is yours, not the vendor’s. The constraints buyers are usually pointed at are the EEOC Uniform Guidelines on showing job-relatedness where a test produces adverse impact, NYC Local Law 144 on bias audits and candidate notice for automated employment decision tools, and the ADA where an instrument functions as a medical examination. We did not verify how any of these apply to a given employer, and this is not legal advice. What we can report is who says anything concrete: Harver publishes named third-party bias audits under Local Law 144, The Predictive Index describes independent audits against the four-fifths standard, and Vervoe says its AI is audited for fairness by Holistic AI. The rest leave the question to you.

Cheating moved faster than proctoring. Recruiters report proxy test-takers appearing on video and candidates using AI assistants during assessments. TestDome and TestGorilla publish the most detailed countermeasures. Nobody has solved it.

Where we land

If you want a defensible number to compare, start with the eight vendors that publish one. The cheapest way in is TestGorilla, because it has a real free tier. Among paid plans Bryq shows $69 a month and Evalart $76, and Evalart has the lowest published per-candidate rate; TestDome has the most complete pricing page. If you hire in volume in New York City, read Harver’s bias-audit page before you read anyone’s feature list. If your roles are clerical and office-based, EmployTest is built for exactly that and priced for it.

And if a vendor will not show you a price, it is worth asking why the number is the part that needs a sales call.

We wrote this because the small end of this market is mostly served by tools priced and sold for enterprises, which is the gap SharpAssessment is built for: practical and job-knowledge tests charged by the volume of candidates you assess, with no annual commitment. We left ourselves out of the table above rather than award ourselves first place on our own page, and on a page arguing that vendors should publish what they charge, we will tell you our number when you ask for it at signup rather than making you sit through a call for it.