LM Arena
LM Arena is an open evaluation platform for large language models maintained by institutions including UC Berkeley. Its core mechanism uses human preference voting with an Elo rating system, in which users conduct blind pairwise comparisons of anonymous model responses, aggregating large-scale real user evaluations to generate model leaderboards. The platform covers multimodal evaluation dimensions including text, code, and images, and is a widely referenced comprehensive model capability benchmark in both academia and industry.
Core Facts
LM Arena (formerly Chatbot Arena) is an anonymous head-to-head large model evaluation platform developed by a team at the University of California, Berkeley.
LM Arena (formerly Chatbot Arena), maintained by UC Berkeley's LMSYS team, uses an Elo rating system to measure model capabilities through user blind-test comparative voting.
LM Arena calculates the overall ranking of each model using an Elo rating system (derived from the chess ranking algorithm).
Timeline (last 90 days)
In LM Arena blind-test voting, Gemini 4 Argon ranks first as of October 13.
All Facts (4)
LM Arena (formerly Chatbot Arena) is an anonymous head-to-head large model evaluation platform developed by a team at the University of California, Berkeley.
75%VerifiedLM Arena (formerly Chatbot Arena), maintained by UC Berkeley's LMSYS team, uses an Elo rating system to measure model capabilities through user blind-test comparative voting.
70%VerifiedLM Arena calculates the overall ranking of each model using an Elo rating system (derived from the chess ranking algorithm).
70%UnverifiedIn LM Arena blind-test voting, Gemini 4 Argon ranks first as of October 13.
50%