Codility Test: What to Expect and How the Scoring Actually Works

If a recruiter has sent you a Codility link, you are about to take a timed coding assessment that grades your submissions against hidden test cases. This guide explains what the format looks like, how the automated scoring actually works, and how to prepare without wasting time.

Quick answer: Codility is a technical skills-screening platform. You get an email invite with a link and solve several short coding tasks in a browser-based IDE, usually 30 minutes to 2 hours in one sitting. A whole-test timer runs the entire time, and your code is graded automatically by hidden test cases that check correctness and, for some tasks, efficiency.

Codility is a coding and technical-skills test, not a cognitive-aptitude battery, a personality quiz, or a game-based assessment. Employers use it to screen candidates before their engineers get pulled into live interviews, which means it usually sits early in a hiring funnel.

There are two main products you might encounter. CodeCheck is the asynchronous test you complete on your own time. CodeLive is a real-IDE interview done live with an interviewer. This guide focuses on the async CodeCheck test, which is the one most candidates receive as a link.

Codility says it has sent more than 12 million assessments through its platform, and candidates report that large employers use it in real hiring funnels — for example, candidates on Teamblind describe Amazon using a Codility online assessment for engineering roles.

The assessment consists of several tasks. Each asks you to solve a technical problem by writing or modifying a short program — usually 10 to 20 lines — or by answering a technical question. Task difficulty varies within a single test.

A few structural details matter more than most candidates expect:

• The timer covers the whole assessment, not each task. The exact limit is set by the employer, depends on the number of tasks, and is shown on the intro page before you start. Codility says assessments typically last between half an hour and two hours.
• You cannot pause or stop the timer once it starts. You must complete the test in one sitting.
• When time runs out, whatever is in the IDE is auto-submitted and the assessment ends.
• Time taken does not factor into your automated score. Finishing fast earns you nothing directly — the score is about what your code does, not how quickly you wrote it.

You are given a task description plus one example test case. That example does not count toward your score. It is there to show you the expected input and output shape.

This is the part that trips people up, so it is worth slowing down on.

Every task is scored by test cases. Each task has at least six test cases, and often more. Your score on a task is the percentage of assessed test cases your solution passes. If a task has 10 assessed cases and your code fails 4, you score 60 percent on that task.

There are two categories of test case:

• Correctness test cases apply to every task. They check that your code returns the right answer for a range of inputs, including edge cases you cannot see.
• Performance test cases apply only to fundamental algorithmic tasks. They check efficiency — effectively your Big-O behavior — by feeding your code large inputs and seeing whether it finishes in time.

Here is a worked example of why both matter. Suppose a task asks you to find duplicates in a list. A brute-force solution that compares every pair of elements might be correct on small inputs and pass every correctness case. But on a large performance case with, say, a million elements, that quadratic approach times out and fails. So you can write code that is logically right and still lose points because it is too slow. A more efficient approach — sorting first, or using a hash set — passes both the correctness and performance cases.

The single most common surprise, documented repeatedly in developer Q and A, is this: seeing OK when you run your code against the example does not mean you passed. The example run only checks the one visible case. Your real score depends on the hidden assessed cases, which you never see. Reports later show which cases passed or failed and whether a failure was a wrong answer or a timeout.

QA tasks are scored differently. There, you write your own test suite against acceptance criteria, and your suite must pass the first "perfect page" case. Fail that one and the task scores zero regardless of anything else.

Whether you ever see your own results depends on the employer. Some hiring teams share results, others keep them confidential. If you want to understand where automated coding tests sit in the wider hiring process, our overview of skills assessments covers the landscape.

Codility ships an anti-fraud suite, but employers turn features on or off per assessment. Depending on the employer, you may encounter:

• Photo ID verification — you scan a photo ID plus a selfie, validated with facial recognition.
• Network IP detection — checks whether the same device or location is being reused.
• Similarity and plagiarism detection — each solution is cross-checked against Codility's database of submitted solutions and known leaks, and matches are flagged for the employer's technical team to review.
• Leaked-task monitoring — Codility scans the internet for leaked tasks.

Screen or camera tracking depends entirely on the employer's policy. Some tests require you to grant camera or screen permissions, others do not. During the assessment you must work on your own and may not consult other people. Whether you can use books or websites depends on the specific test's stated policy, so read the intro screen carefully.

One practical note: Codility is not supported on phones or tablets. You need a laptop or desktop, especially if video proctoring or screen sharing is switched on.

Codility's own candidate-experience data — more than 48,000 responses collected between October 2025 and April 2026 — reports that roughly 9 in 10 candidates find the content fair and 87 percent rate the experience Good or Excellent. That data is vendor-collected and self-reported, so weigh it accordingly. On the employer side, Codility is rated 4.6 out of 5 across more than 865 G2 reviews.

The most consistent candidate complaint is time pressure. One candidate blog describes a test with only 70 minutes for 3 problems and failing after running out of time on the last one. A senior manager publicly complained that the system "offers ridiculously little time to complete 2 coding tasks." An independent 2026 review praised Codility's anti-cheating and assessment integrity but said the interface "is showing its age."

The platform is built for accessibility. It meets WCAG 2.2 AA, works with screen readers, supports keyboard navigation, and offers an Accessibility Mode with up to 400 percent zoom in the IDE. If you need extra time or accommodations, request them from the administrator before you start.

You cannot cram algorithmic fluency the night before, but you can remove avoidable failure points.

1. Take the public sample test. Codility offers a free demo test and free "Learn to Code" lessons on topics like Iterations and Stacks and Queues. Do the demo so the IDE and submission flow are not new to you on test day.
2. Practice writing edge-case-safe code. Since correctness cases include inputs you cannot see, get in the habit of handling empty inputs, single elements, and boundary values.
3. Think about complexity before you code. Performance cases punish slow solutions. If you can spot an O(n squared) approach and improve it, do that before submitting.
4. Stop trusting the example run. Treat OK on the visible case as a smoke test, not a pass. Re-read the task for constraints that hint at required efficiency.
5. Set up your environment. Use a supported browser on a laptop or desktop, and confirm any camera or screen permissions ahead of time.
6. Budget your time across tasks. The timer is global. If one task is eating your clock, bank partial credit elsewhere rather than running out on the last problem.

If you are earlier in your job search and want to understand the AI-driven side of hiring more broadly — how automated screening, scoring, and interview tools are converging — our comparison of the best AI recruiting software for 2026 is a useful next read.

How long is the Codility test?

Most Codility assessments run between 30 minutes and 2 hours. The exact limit is set by the employer, depends on how many tasks there are, and is displayed on the intro page before you begin.

Is the Codility test hard?

Difficulty varies by task and by employer, but the common complaint is time pressure rather than raw problem difficulty. Candidates report tests like 3 problems in 70 minutes, so pacing across the whole test matters as much as solving any single task.

Does OK mean I passed the Codility test?

No. Seeing OK when you run your code only checks the one visible example. Your actual score comes from hidden assessed test cases you never see, covering both correctness and, on some tasks, performance.

Can you retake a Codility test?

Retakes are controlled by the employer, not by a fixed Codility rule. There is no publicly documented candidate cooldown, so whether you get another attempt depends on the hiring team's policy.

Does time taken affect your Codility score?

No. How long you take does not factor into the automated score. The score is based only on how many assessed test cases your code passes.

Can you take the Codility test on a phone?

No. Codility is not supported on phones or tablets. You need a laptop or desktop, especially if the employer has enabled video proctoring or screen sharing.