Advent of Code entered its 2025 edition with two notable changes after ten years of annual programming challenges: fewer puzzle days and no global leaderboard. Creator Eric Wastl described both decisions on the event's official information page, while keeping the familiar December 1 opening and daily release rhythm through the middle of the month.
The project presents an Advent calendar made from short coding problems. Participants can solve them in any programming language, and the exercises are used for individual practice, interview preparation, workplace training, university courses and informal competition. Wastl says a computer-science background is unnecessary, although some programming knowledge and problem-solving ability are expected. Every puzzle is designed to have a solution that finishes within 15 seconds on hardware that is ten years old.
The reduced calendar reflects the work required to run the event. Wastl said creating puzzles takes most of the substantial personal time he commits each year. After maintaining the same schedule for a decade, he wanted a change. The 2025 puzzles would still appear every day, but the sequence would end in mid-December rather than fill the conventional longer run.
Removing the worldwide ranking addressed a separate concern. According to Wastl, the leaderboard had placed growing pressure on him, on supporting infrastructure and on participants. Some competitors treated results too seriously, while others interpreted slower completion times as evidence that they were inferior programmers. He also cited conduct extending beyond the intended spirit of the challenge, including distributed denial-of-service attacks.
Private leaderboards remain available, and organizers can share a view that others may read. Wastl specifically asked users not to combine that information into a replacement worldwide ranking. Rules for private groups can be set locally, including choices about languages, allowed tools, opening times or performance limits.
The guidance also discourages using AI systems to solve the puzzles. Wastl argues that Advent of Code is intended as an exercise for people, not as a benchmark designed around machine capabilities. Participants interested in practicing AI prompting are advised to find activities created for that purpose.
The event continues to use third-party OAuth services for sign-in. Advent of Code receives identity information supplied by the chosen provider rather than collecting that service's credentials. Together, the changes preserve the challenge's focus on learning and personal problem-solving while reducing operational burden and the pressure created by a single public ranking.
The official guidance continues to emphasize debugging habits rather than a single competitive objective. Participants who get stuck are advised to compare their programs with provided examples, construct small cases they can verify manually, ask a friend or return later. That approach is consistent with the 2025 changes: success can mean learning or practice, without measuring every solver against the fastest global result.
