← All notes

Summary

Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference (2024) describes the live, pairwise human-preference arena that has become a de facto public leaderboard for assistants. People compare two anonymous model replies, and the platform turns those votes into a ranking.

What it introduced: an open evaluation platform based on crowdsourced, side-by-side human preference, plus the methods used to turn those votes into stable model scores.

What it led to: a widely cited public ranking that labs and users actually look at, alongside older static benchmarks. Live notes on agent and assistant evaluation include ao-evals-agent-benchmarks-and-judges. For the earlier preference-training background, see this batch's pointers on reinforcement learning from human preferences and PPO.

Where to read it: https://arxiv.org/abs/2403.04132 (v1, 7 Mar 2024).

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

@misc{iil_note_0p1mhhza,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2403-04132-chatbot-arena/},
  note = {Permanent ID iil:note:0p1mhhza. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2403-04132-chatbot-arena.md.