yyyhy/nash-arena

Overview

Provides an arena for evaluating large language models through chess and card games. Built in Python, it allows developers to test and compare LLM strategic reasoning in turn-based game scenarios.

Best for

Best for
Developers researching LLM capabilities in game-based strategic reasoning

Use cases

• Benchmarking LLM decision-making in chess and card games
• Comparing performance of different language models on structured game tasks
• Testing strategic reasoning and planning abilities of LLMs

Notes

Provides an arena for evaluating large language models through chess and card games. Built in Python, it allows developers to test and compare LLM strategic reasoning in turn-based game scenarios.

2 stars on GitHub. Last updated 2026-03-20. Licensed MIT.

Use cases

• Benchmarking LLM decision-making in chess and card games
• Comparing performance of different language models on structured game tasks
• Testing strategic reasoning and planning abilities of LLMs

Pros

• Open source and freely available
• Focused specifically on LLM game-playing evaluation
• Simple Python codebase easy to modify

Cons

• Very low community adoption (2 stars)
• Unclear documentation and setup instructions
• Limited to chess and card games only

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.