# Codility vs DevSkiller: which do AI models recommend for technical hiring assessm, October 2026

HR AI Recommendation Index, October 2026 Edition, Technical hiring assessments. Three of fourteen models named Codility first on the direct prompt; one named DevSkiller. Page: https://hr-ai-index.com/talent/technical-hiring-assessments/codility-vs-devskiller/

| | First-choice share | Rank | Negative rate | Labels | Models naming it |
|---|---|---|---|---|---|
| Codility | 9% | #6 of 9 | 25% | 53 | 14 of 14 |
| DevSkiller | 2% | #8 of 9 | 0% | 10 | 9 of 14 |

## The direct prompt, model by model

- DeepSeek V4 Flash: codility first (first choices: Codility) (alternatives: CodeSignal, HackerRank)
- GLM 4.7 FlashX: codility first (first choices: Codility, TestGorilla) (alternatives: CodeSignal, Coderbyte, HackerRank)
- MiniMax M2.5: codility first (first choices: Codility) (alternatives: AssessHub, CoderPad, HackerRank)
- Muse Glimmer 30B: devskiller first (first choices: CodeSignal, DevSkiller, HackerRank) (alternatives: Codility)
- GPT-5.4 mini: neither first, one named (first choices: HackerRank) (alternatives: CodeSignal, Codility)
- Perplexity Sonar: neither first, one named (first choices: CodeSignal) (alternatives: Coderbyte, DevSkiller)
- Grok 4.1 Fast: neither first, one named (first choices: HackerRank) (alternatives: CodeSignal, CoderPad, Coderbyte, Codility)
- Qwen 3.7 Flash: neither first, one named (first choices: CodeSignal) (alternatives: CoderPad, Codility, TestGorilla, Voomer)
- Claude Haiku 4.5: neither named (first choices: HackerEarth, TestGorilla) (alternatives: CodeSignal)
- Gemini 3.5 Flash: neither named (first choices: CoderPad, Woven) (alternatives: CodeSignal)
- Mistral Small: neither named (first choices: Goodfit) (alternatives: HackerRank)
- Llama 4 Maverick: neither named (first choices: Goodfit) (alternatives: CodeSignal, Coderbyte, HackerRank)
- Kimi K2: neither named (first choices: Coderbyte) (alternatives: HackerRank, TestGorilla)
- GPT-6 Luna: neither named (first choices: CodeSignal Hire Grow) (alternatives: CoderPad, HackerRank)

## What the models said about Codility

- "Avoid if budget is the main constraint... Codility tends to start higher and is more enterprise-oriented" (GPT-5.4 mini, budget prompt, hard negative)
- "(e.g., HackerRank, Codility, CodeSignal): ... often criticized for not reflecting real-world job tasks" (Mistral Small, negative prompt, hard negative)
- "High negative sentiment (e.g., labeled "hostile candidate experience" in AI model analyses)" (Grok 4.1 Fast, negative prompt, hard negative)
- "### 1. Codility - Best for: Mid-market to enterprise companies needing secure, scalable assessments" (MiniMax M2.5, direct prompt, first choice)
- "Choose HackerRank or Codility if you are an enterprise needing rigorous automated screening" (Gemini 3.5 Flash, comparative prompt, first choice)

## What the models said about DevSkiller

- "CodeSignal / DevSkiller are commonly recommended for 50-500 hires mid-market teams" (Muse Glimmer 30B, direct prompt, first choice)
- "CodeSignal or DevSkiller is explicitly listed for *mid-market (50–500 hires)*" (Perplexity Sonar, direct prompt, alternative)
- "Uses the RealLifeTesting" methodology... realistic code scenarios" (GLM 4.7 FlashX, comparative prompt, alternative)

Share is the count of first choices across the direct, paraphrase, budget and scale prompts over all fourteen models, for a mid-market B2B company; rank is within the category. Comparisons are drawn for the top eight products in each category. Published under CC BY 4.0; the output is the models' output, and nothing here is a recommendation by the index.