LMArena (first called Chatbot Arena) ranks AI chat models by letting people compare two anonymous answers side by side and vote for the better one.
In plain words
You type a prompt and get replies from two models without knowing which is which. You pick the one you prefer. Many thousands of these votes are turned into ratings, like chess ratings, to build a public leaderboard. Because real people ask real questions, it measures what users prefer rather than scores on a fixed test.
A quick example
You ask the arena for a short poem about rain. Model A and Model B both answer. You vote for B, and only then learn which models they were. Your vote nudges B's rating up and A's down on the leaderboard.
Where you'll meet it
- Comparing chat models on a public leaderboard.
- Launch announcements that cite arena rankings.
Related
Sources
- PaperChatbot Arena: An Open Platform for Evaluating LLMs by Human Preference, Chiang et al. (arXiv) · read 29 Sept 2026
- DocsChatbot Arena: Benchmarking LLMs in the Wild with Elo Ratings, LMSYS Org · read 29 Sept 2026
- DocsHow Arena Works, LMArena · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.