---
title: "LLM as online judge"
url: https://perrotta.dev/2025/09/llm-as-online-judge/
last_updated: 2026-01-03
---


In 2025 it is quite comfortable to use LLMs (GenAI) to grade your code
submissions (from LeetCode, HackerRank, etc).

Not only do they give you near-instant feedback (for performance, readability,
consistency) and code analysis, but they can also spot mistakes that would
otherwise take you precious tens of minutes of time to debug[^1].

[^1]: While it's important to spend some time debugging the code yourself,
    there's a lost opportunity cost when getting stuck, yielding lost time that
    could have been spent resolving more (diverse, interesting) problems.