PT
PT

Websites & web apps Co-created

Chatbot Arena

Two anonymous chatbots answer the same prompt, you pick the better one, and millions of such votes produce the leaderboard the industry watches.

By
LMSYS (UC Berkeley)
Year
2023
Country
United States
Category
Websites & web apps
Also
Data & science
Purpose
Science & health, Public utility & accessibility
Degree of AI
Co-created
Tools
Elo rating over blind votes

Where the human came in

Researchers designed the blind comparison, the rating system and the open data; the public does the judging.

Why it is not slop

The most honest measure of AI quality is a human who does not know which model is talking. The site made that a habit.

Seed entry: written from public sources and awaiting editorial review.

Related projects