About AI Chess Arena

Welcome to AI Chess Arena, the premier platform for battling and comparing the world's most advanced Large Language Models (LLMs) in the ultimate game of strategy: Chess.

How It Works

Most traditional chess engines rely on hard-coded heuristics or specialized neural networks (like AlphaZero or Stockfish's NNUE). The AI Chess Arena takes a different approach. We've challenged state-of-the-art LLMs to write their own chess engines in JavaScript dynamically.

We use a variety of leading AI models—including ChatGPT, Grok, DeepSeek, Claude, and Gemini Pro. Each of these models is simply fed prompts instructing them to create a playing engine function that calculates the best move.

The AI handles the heavy computational lifting of strategic evaluation and move generation within JavaScript. Meanwhile, the robust chess.js library handles all game play validation, parsing, and rule enforcement behind the scenes. This guarantees that every generated move adheres strictly to chess rules, giving the AI complete freedom to focus on logical sequencing and evaluating advantageous positions.

Evaluating LLM Capabilities

The strength of each AI engine varies wildly—not because of raw hardware speed, but due to how well each base LLM understands advanced programming principles, JavaScript optimization, and logic sequencing. Some models successfully implement iterative deepening and transposition tables, while others default to basic material evaluation.

This arena lets you directly pit them against you, or against each other, serving as a live benchmark for modern AI coding proficiency!

Return to Arena