Codility is one of the two default answers whenever a hiring manager asks "what should we use to test candidates' coding skills" — the other being CodeSignal. It's been in the market since 2009, which makes it ancient by AI-hiring-tool standards, and its longevity shows in both its strengths and its blind spots. This review looks at what Codility actually measures, what it costs, how it holds up on accessibility, and where its age starts to work against it.
What Codility Actually Does
Codility is a technical assessment platform, not a conversational AI interviewer — an important distinction we drew out in our 2026 platform comparison. Candidates are given coding problems (chosen from a library or custom-authored) and write solutions in a browser-based IDE supporting 40+ languages. The platform runs the code against test cases, scores correctness and, on higher tiers, time and space complexity, and produces a report that recruiters and engineering managers review.
Two product lines sit under the Codility name: CodeCheck, a self-serve async assessment tool, and CodeLive, a live pair-programming interview environment used for later-stage technical interviews. Most companies use both — CodeCheck to filter a large applicant pool down to a manageable shortlist, CodeLive to run the actual live technical round.
Accuracy: What the Score Actually Tells You
Codility's core scoring is straightforward and, within its scope, reliable: does the code pass the test cases, and how efficient is it. That's a narrow but honest claim — unlike some enterprise conversational suites we've reviewed, Codility doesn't claim to predict "culture fit" or "leadership potential" from a coding test, which keeps it largely outside the bias litigation we covered in our AI hiring bias research summary.
The real accuracy question isn't "does it grade code correctly" — it does — it's "does passing a Codility test predict on-the-job performance." Here the evidence is mixed, and this is a limitation of the entire technical-assessment category, not Codility specifically. Algorithmic puzzle-solving under time pressure correlates weakly with real-world engineering ability, which tends to look more like reading unfamiliar code, debugging production issues, and making tradeoffs under ambiguous requirements — none of which a 45-minute CodeCheck session captures well.
Codility is accurate at measuring what it measures — correctness and efficiency of code under test conditions. It is not, and doesn't claim to be, a strong predictor of broader engineering competence on its own. Treat it as a filter, not a final answer.
Pricing
Codility doesn't publish list pricing — quotes are custom based on hiring volume, seat count, and whether you need CodeLive alongside CodeCheck. Third-party and reseller data puts small-team annual contracts in the $2,500–$10,000 range and mid-market/enterprise deployments considerably higher, often $15,000–$50,000+/year depending on assessment volume. That places it above the startup-tier tools we covered in our startup AI interview tools roundup but generally below the enterprise conversational suites like HireVue.
Accessibility
This is where Codility genuinely leads the category. It has invested specifically in WCAG 2.2 compliance — screen reader support, keyboard-only navigation through the assessment interface, adjustable time extensions for candidates who request accommodations, and a reduced-motion mode. Most competitors in the technical-assessment space, and certainly most conversational AI interviewers, have not made comparable public commitments. Given the accessibility gaps we're seeing more broadly across AI hiring tools, this is a real differentiator worth weighing if accessible hiring is a compliance or values priority for your org.
The AI-Assisted-Coding Problem
The most significant recent shift in this category, which we flagged in our platform comparison: technical assessment platforms are increasingly scoring how candidates use AI coding tools during the test, not just whether the final solution works. Codility has moved in this direction with detection and disclosure features around AI-assistant usage during CodeCheck sessions, but the implementation is still maturing — it can flag that AI assistance was likely used, but drawing a confident line between "used AI as a legitimate tool" and "outsourced the entire problem to AI" remains genuinely hard to automate, and Codility doesn't claim to have solved it.
Codility vs. CodeSignal vs. HackerEarth OnScreen
| Platform | Format | Standout | Weakness |
|---|---|---|---|
| Codility | Async + live pair programming | Accessibility, mature IDE, custom problem authoring | Narrow signal on real-world engineering ability |
| CodeSignal | Async + live, plus a general cognitive assessment product | Broader assessment types beyond pure coding | Less mature accessibility tooling than Codility |
| HackerEarth OnScreen | AI-led adaptive technical conversation | Merges assessment with conversational AI screening | New (launched April 2026), track record still forming |
If your priority is a well-worn, accessible, pure-coding assessment, Codility remains the safest default. If you want a broader battery of cognitive and technical assessment types under one roof, CodeSignal is the stronger fit. If you're exploring whether AI-led adaptive technical screening can replace an early human round entirely, HackerEarth OnScreen — and tools like Ray by Diyam AI, which we tested directly against the field — represent where the category is heading, though both are newer and less proven at scale than Codility.
Codility does one thing — grading code correctness and efficiency — and does it reliably, accessibly, and without overclaiming what the score means. That restraint is rarer in this market than it should be.
This is the fourteenth piece in our AI recruitment research series. Sharingan AI evaluates recruitment technology independently, without vendor sponsorships or affiliate relationships.