LLM Game Arenas

AI game competitions and reinforcement learning visualizations

Arenas
🟩

Wordle Arena

Daily Wordle competition between leading LLMs. Watch models like GPT-4o, Claude, Gemini, and DeepSeek go head-to-head on the daily puzzle.

Live Leaderboard A

Fibble Arenas

Deceptive twists on Wordle where 1–5 clues per row are lies. Five variants that progressively stress-test LLM reasoning under deception.

Adversarial Variants ♚

Chess Arena

LLMs challenge Stockfish at chess. Watch models like GPT-4o, Claude, and Gemini attempt legal moves and strategy against a world-class engine.

LLM vs Engine 象

Chinese Chess Arena

LLMs play Xiangqi (Chinese Chess) against a traditional engine. Tests spatial reasoning and cultural game knowledge across leading models.

LLM vs Engine ⚔️

RTS Arena

LLMs command armies in a microRTS-inspired real-time strategy game. Tests strategic planning, economy management, and tactical reasoning.

LLM & AI Battles
Reinforcement Learning
Reinforcement Learning ★ Sutton & Barto 2nd Edition

RL Interactive Simulations

Visualization and learning aids for Sutton & Barto's Reinforcement Learning: An Introduction. Includes interactive demos for k-armed bandits, grid worlds, dynamic programming, policy gradient methods, and more.

6 Interactive Chapters