Advent of Code Changes for 2025

calendar Nov 4, 2025 · 4 min read · Python Advent of Code  ·
Share on: linkedin copy

It’s not long to wait now, Advent of Code will soon begin. For 2025 there are changes with fewer challenges and the removal of the global leaderboard.

This is the fourth year I’ve posted about taking part in Advent of code and this year the event will be different. Eric Wastl recently posted that this year Advent of Code will run over 12 days instead of the usual 25 and there will be no global leader board. Both of these changes are ones that I support.

Fewer challenges

Last years event was the 10th year of AoC, after the event I half wondered if it would be the last one, us humans like closure after a pleasing number iterations, 5 or 10 years would be fine, 9 years would have been annoying. 10 years of AoC means:

  • 10 years of planning, starting months in advance,
  • 10 different stories to keep the continuity between challenges,
  • So many facts and useful pieces of information to weave into those stories
  • So many easter eggs sprinkled throughout the challenges
  • Creation of 490 challenges that all have to be achievable - and separeate inputs files for groups of users for each of those challenges.
  • 10 years of keeping a website up and running for 25 days straight with all the pressure this sudden surge in traffic would have.

All this considered, I think Eric has done enough for AoC over the years and I will not begrudge him any time that he wants to take back for himself and his family.

Global leaderboard

This is a big change this year, there will be no global leader board. Personally I will never be on that board, the challenges open at 5am in the UK and I’m usually asleep at that time so already it favours participants in other regions.

For the first 8 years I marvelled at some of the developers that continually made the top 100, some have YouTube channels where they live solve the challenges and give the explanation for their solution.

In 2023 there were a few days where the top of the leaderboard was clearly AI generated code but this dropped away quickly as the coding tools were not quite able to solve the problems first time or as quickly.

For the 2024 challenge, the leaderboard was dominated by participants using AI to solve the problems for them, despite the organisers of AoC asking people not to…

Can I use AI to get on the global leaderboard? Please don’t use AI / LLMs (like GPT) to automatically solve a day’s puzzles until that day’s global leaderboards are full. By “automatically”, I mean using AI to do most or all of the puzzle solving, like handing the puzzle text directly to an LLM. The leaderboards are for human competitors; if you want to compare the speed of your AI solver with others, please do so elsewhere. (If you want to use AI to help you solve puzzles, I can’t really stop you, but I feel like it’s harder to get better at programming if you ask an AI to do the programming for you.)

This was a shame as those with genuine talent were pushed down the boards, replaced by participants who can copy/paste quickly. It was sad to hear the exasperation in the voices of some YouTubers when their amazing solution to parts 1 and 2 was beaten by a solution subitted in 9 seconds (see 2024 day 1 leaderboard). “This is why we can’t have nice things” springs to mind, because the leaderboard would no longer be fair it’s been taken away. Hopefully this will encourage some of the top performers from the past to continue with the challenge and hopefully they will set up a private board they can share a view of so that we can again see the benchmark times for what good looks like.

My plan for AoC 2025

At the moment I don’t have one, I did consider using rust, zig or try C++ but I honestly don’t think I have the time or headspace at the moment to take on another language. It might be I fall back to Java, JavaScript or Python again this year but I’ll post about my choice separately in a week or so.