Codility vs CodeSignal

Codility and CodeSignal both filter engineering candidates before anyone on your team spends an hour in a room, and as of 2026 both publish a price you can pay with a card. That is what makes this a live decision again. The old framing — Codility is the transparent one, CodeSignal is quote-only — died when CodeSignal put its Hire plans on the pricing page at $79 and $479 a month. What separates them now is scope. Codility goes deep on engineering and sells the assessment methodology itself. CodeSignal goes wide across job functions and sells AI interviewers that run the first pass for you.

Where Codility wins

• Tasks authored from your own codebase. From the Scale tier up, Codility ships an MCP server: an MCP-compatible coding agent reads your repo, a pull request, or a job description and publishes a validated multi-file task with test cases straight into your task library. CodeSignal’s AI-Native Authoring, shipped 2026-05-19, generates questions from a prompt — it does not read your repository. For final rounds, where a leaked public task costs the most, that is the difference worth paying for.
• Environment fidelity for backend and data roles. The Custom tier runs assessments in VS Code with sidecar services — databases, caches, queues. A candidate wiring a query against a live database is a different measurement from one filling in a function signature.
• Scoring stays rules-based. The Codility Evaluation Engine scores against fixed criteria, and the AI sits on the candidate’s side of the assessment: they work with an agentic copilot, and their activity is reviewable per submission. CodeSignal puts AI grading and insights inside the evaluation itself. If your counsel is working through NYC LL 144 or Illinois HB 3773, that distinction is the first thing to get in writing for the exact configuration you buy — see AI screening bias.
• Interviewers do not consume seats. Every Codility plan includes unlimited Collaborator Users; only Platform Users are metered, at 1 on Starter and 3 on Scale. A twelve-person interview panel adds nothing to the bill.
• Cheapest per invite at the entry tier. $1,200 a year for 120 credits is $10 per candidate screened.
• Skills Intelligence points the same tasks at engineers you already employ, mapping capability gaps rather than filtering applicants. It is Custom-tier. CodeSignal covers upskilling through Learn, which is a separately priced product line rather than part of the hiring contract.

Where CodeSignal wins

• It assesses roles that are not engineering. Grow adds go-to-market assessments and AI Interviewers for sales, CS and marketing; Pro adds human resources plus finance and operations. Codility has a Business task library, but it is Custom-gated — there is no self-serve path to non-engineering assessment.
• AI Interviewers, AI phone screens and video avatars from the $79 tier. Product, design and engineering interviewers ship on Build. Codility equips your interviewer; CodeSignal removes them from the first pass. When interviewer hours are the constraint rather than question quality, that is what moves cycle time.
• ATS integration at a published price. Gem and Ashby connect on Grow. Every Codility ATS integration — Greenhouse, Lever, Ashby, Workday, SAP — sits behind a quote. On Codility Starter or Scale, someone copies results into your ATS by hand.
• Suspicion Score, and published research behind it. It scores four violation classes — copy-paste plagiarism, proxy test-taking, unauthorized AI use, identity fraud — from solution similarity, telemetry and copy-paste activity. CodeSignal’s own detection data, published 2026-02-25, put the flagged rate at 35% of 2025 assessments against 16% in 2024, and 40% at entry level against 15%. These are vendor figures measured on the vendor’s own platform, so treat them as a floor on the problem rather than an industry rate. The operational finding underneath them is the useful part: unproctored assessments showed score increases four times larger than proctored ones.
• Burst capacity. Grow’s 420 annual credits carry no monthly cap. Codility’s Scale allowance is 300 credits capped at 25 per month, so a January req spike cannot draw against the rest of the year.
• A score with published benchmarks. Assessment Score runs 200 to 600 across every certified assessment, recalibrated for consistency between them and validated by in-house I-O psychologists, with pass-rate benchmarks by role and region.

Pricing reality

Codility publishes two self-serve tiers. Starter is $1,200 a year, annual billing only, for 120 invite credits and 1 platform user — $10 per invite. Scale is $6,000 a year, or $600 a month, for 300 credits capped at 25 monthly and 3 platform users — $20 per invite. Custom is quote-only.

CodeSignal publishes two as well. Build is $79 a month billed annually for 60 annual credits, or $99 monthly for 5 a month — $15.80 per credit on the annual plan. Grow is $479 a month billed annually for 420 credits, or $599 monthly for 35 — $13.69 per credit. Pro is quote-only. Overages on both plans bill at $20 per credit on the next cycle.

The crossover is the number to carry into the decision. At the entry tier Codility is 37% cheaper per invite, $10 against $15.80, though $252 more in absolute annual spend. At the growth tier it inverts hard: CodeSignal Grow is $5,748 a year for 420 credits against Codility Scale’s $6,000 for 300 — 4% less money for 40% more volume, plus the ATS integration and AI interviewers Codility gates behind a quote. Screening 100 to 150 candidates a year, Codility Starter is the cheapest defensible programme you can buy. Above roughly 300, CodeSignal Grow wins on arithmetic before either feature set enters the argument.

Both quote-only tiers price in the same compliance layer — SSO, audit logs, enterprise ATS, ID verification, and validation studies from staff psychologists — so a like-for-like enterprise quote is closer than the self-serve ladders suggest.

Implementation effort

Codility Starter and Scale go from signup to first assessment the same day. What you do not get until Custom is ATS write-back, SSO, the full API, weighted scoring, ID verification and the desktop app that flags unauthorized applications. Results live in Codility until then. A Custom rollout is a procurement cycle plus programme design with the assessment science team.

CodeSignal Build and Grow are same-day too, and Grow lands results in Gem or Ashby without a manual copy step. Pro brings SSO, SCIM, RBAC, enterprise ATS connectors for Workday, iCIMS and Oracle, custom DPA and a security review.

One migration cost is specific to CodeSignal. Assessment Score replaced the historical Coding Score thresholds in a recalibration documented in April 2026. Historical scores convert automatically; your cut scores do not. The organization re-derives its own thresholds against the 200-to-600 scale. If a pass mark is written into a hiring policy or a published bias-audit filing, budget the re-derivation and the re-approval, not just the platform swap.

Verdict

• Pick Codility when hiring is engineering-only, tasks need to come from your own codebase, environment fidelity matters for backend or data roles, or the scoring has to stay rules-based and auditable for a bias-audit posture. The programme case is strongest at Custom, where Skills Intelligence turns the same methodology on the engineers you already employ.
• Pick CodeSignal when you assess beyond engineering, interviewer hours are the bottleneck rather than question quality, you need ATS write-back on a self-serve budget, or hiring arrives in bursts that a 25-per-month cap would throttle.
• Pick neither when scheduling rather than screening is the constraint — buy GoodTime — or when you need many non-engineering role types cheaply, where TestGorilla covers more ground. If you want the AI to conduct the whole interview rather than score an assessment, micro1 is the different purchase. Below roughly 10 engineering hires a year, a structured take-home reviewed by two engineers costs less than either subscription. Above roughly 50, HackerRank belongs in the set; HackerRank vs CodeSignal settles that pair.
• Choosing in a vacuum, pick CodeSignal Grow. It delivers more assessment volume, ATS integration and AI interviewers for less money than Codility Scale, and nothing about it forecloses the switch. Move to Codility — Custom, not Scale — once own-codebase tasks, sidecar-service environments or rules-based scoring become requirements rather than preferences. Credits are annual on both sides, so renewal is the natural switching point.