Vibeleaderboard
Index / app
Visit www.lolbench.lol
Category
Developer Tools
Rank
Platform
web
Type
APP
Date

About

A leaderboard site that has thirteen large language models explain and write jokes, then lets people cast blind votes on which model's humor lands best. It turns humor comprehension and generation into a crowd-judged benchmark for comparing LLMs.

What it can do

  • Generate joke explanations using multiple LLMs

    Joke promptLLM-generated joke explanation

  • Generate jokes written by multiple LLMs

    PromptLLM-generated joke

  • Collect blind votes from users on which LLM's humor is best

    User voteVote record

Why it made the leaderboard

Humor is a capability models are notoriously bad at benchmarking with automated metrics; blind human voting on model-generated jokes gives a lightweight, engaging way to see comedic and comprehension gaps between models.

Tags

llmbenchmarkhumorleaderboardai-modelsvotingeval

Media

LOL Bench

Comments (0)

No comments yet

Editorially curated, with community endorsements as a secondary signal. Corrections welcome.