Skip to content

llm-colosseum

Rank LLMs through Street Fighter III matches

Visit official site
Category AI models

What llm-colosseum is

The tool is described as a benchmark that evaluates real-time decisions of large language models through Street Fighter III matches.

What you can do with it

  • Run live matches between language models
  • Have a model react to opponent attacks

Pros & cons

✓ What we like

  • Users can submit a custom model to join the ranking
  • Models adapt to game state such as health and energy bars

! What to watch out for

  • The benchmark is built around a single game
  • Ranking is computed with the ELO system

FAQ

Which game is used?

The material names Street Fighter III.

How are models ranked?

The material describes an ELO rating system based on match results.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review