Codility test practice: what to expect and how to score higher

Every Codility task returns two numbers when you submit: a correctness score (does your code produce the right output) and a performance score (does it stay within the required time and space complexity on very large inputs). Your final task score is the average, weighted by the recruiter’s template.

This is the single most important thing to internalise. A Python solution that is O(n²) can score 100 percent correctness and 0 percent performance, giving you 50 on a task where a friend who wrote O(n log n) got 100. Recruiters see both numbers.

Codility’s hidden test cases are grouped by input size. A brute force typically passes the small and medium groups but times out on the large and huge ones. That gives you a partial score that is often good enough to clear a screening cut-off for a services role, but not enough for a competitive product-team role.

• Read the constraints line first. If N can reach 100,000 or a million, an O(n²) approach will not finish.
• Solve the brute force in your head, submit it if you must, then rewrite for performance.
• Use Python’s collections.Counter, bisect, and prefix sums freely — they are all allowed.

Codility publishes free Lessons 1 through 17 with sample tasks in each. Recruiters overwhelmingly pull from this library or lightly reword it, so working through the lessons is the highest-return prep you can do.

• Lessons 1-4: iteration, counting, prefix sums. Every candidate should score 100 here.
• Lessons 5-8: sorting, stacks, queues, leader problem. This is where most Indian candidates lose 15-25 points.
• Lessons 9-12: maximum slice, prime and composite numbers, Sieve, Euclidean algorithm. Number-theoretic and often trickier than they look.
• Lessons 13-17: Fibonacci, greedy, caterpillar method, dynamic programming. Rarely appears in a screening OA, common in follow-up rounds.

Codility does not publish company-level cut-offs, but the industry pattern is consistent. Screening rounds for European software roles usually gate at around 120 out of 300 for junior roles and 180 to 220 for mid-level. A perfect 300 is rare and not required.

Recruiters also look at task-by-task breakdown. Two solid partial scores read better than one perfect solution and two blanks. Always attempt every task, even if you only submit a brute force in the last two minutes.

Focus on the platform’s own lessons — nothing else replicates the scoring model.

• Day 1: take the free Codility demo test end-to-end. Note your correctness vs performance split.
• Day 2: Lessons 1 to 4, all tasks, timed.
• Day 3: Lessons 5 to 8, focus on Leader and Stack tasks.
• Day 4: Lessons 9 to 12, particularly MaxCounters and prime-factor problems.
• Day 5: rewrite any Day 2-4 solutions that scored below 100 performance.
• Day 6: a fresh mock at your real test’s length.
• Day 7: revise notes, sleep early.