LMSYS Chatbot Arena
Description: Community-driven LLM ranking through pairwise comparisons with Elo system
Website: https://lmarena.ai
The LMSYS Chatbot Arena is the best-known platform for comparing Large Language Models through community voting. With over 800,000 votes and 24 million monthly visitors, it is the gold standard for LLM evaluation.
Features
- Elo ranking system: Like chess, based on direct comparisons
- 90+ models evaluated: Commercial (GPT, Claude, Gemini) and open-source (Llama, Mistral, DeepSeek)
- Community-driven: Real users rate responses in blind tests
- Transparency: Code (FastChat) and data available on GitHub
Usage
On lmarena.ai you can:
- View current rankings
- Test and rate models yourself
- Filter by categories (open-source, coding, etc.)
- Track performance trends
Highlight
Unlike automatic benchmarks, rankings are based on real user preferences in actual conversations.