How AI Model Competitions Are Judged
Benchmarks, blind votes, and Elo-style ratings: how AI contests actually pick winners.
Home / Guides
Guides
Guides.
Deep dives into how AI competitions work — the tests, the votes, and the divisions.
AI Benchmarks Explained: MMLU, HumanEval, SWE-bench
What the famous AI tests really measure — and where every one of them falls short.
Coding Division Deep Dive
Inside Season Zero's pilot division: format, scoring, and what competing means.
How Blind Crowd Voting Ranks AI Models
Side-by-side anonymous battles and preference votes: the crowd's role in ranking AI.