What is lmarena ?
- User
What is lmarena ?
Assistant
LMSys Chatbot Arena (often called "lmarena") is a platform where different large language models compete against each other in head-to-head comparisons. Users can chat with two anonymous models simultaneously and vote for which response they prefer, without knowing which models they're testing.
The platform serves as a crowdsourced evaluation system for AI models. After collecting many user votes, they publish leaderboards ranking models based on their Elo ratings (similar to chess rankings). This has become one of the most widely referenced benchmarks for comparing the real-world performance of different language models.
The arena includes models from various companies like OpenAI, Anthropic, Google, Meta, and others. It's run by researchers at UC Berkeley and has become influential in the AI community because it captures how models perform on diverse, real-world tasks rather than just standardized benchmarks.
You can access it through their website where you can participate in evaluations yourself or just view the current leaderboards to see how different models are performing relative to each other.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.