Leela Chess Zero
Neural Network Engine
Leela Chess Zero (Lc0) is an open-source chess engine that uses deep neural networks to evaluate positions, inspired by DeepMind's AlphaZero research. Unlike traditional engines that rely on hand-crafted evaluation functions, Lc0 learned to play chess entirely through self-play, starting from zero knowledge of the game and training through millions of games against itself.
The engine is a distributed computing project. Volunteers around the world donate GPU processing power to train the neural network, making Lc0 one of the largest collaborative AI projects in chess. The result is an engine with a distinctly human-like playing style that often finds creative, positional ideas that traditional engines miss.
Lc0 regularly competes with Stockfish at the top of computer chess championships and has won multiple seasons of the Top Chess Engine Championship (TCEC). It is free to download and use, and its training data and network weights are publicly available. For players and researchers interested in the intersection of AI and chess, Lc0 is one of the most important projects in the field.
Key Features
- Neural network chess engine
- Learned through self-play (AlphaZero-style)
- Distributed volunteer training
- Open-source with public network weights
- TCEC champion
- Human-like positional style
Links
Related Tools
instant game analysis
update existing
ai chess analysis
AI Chess Coach
find any position
free game review
AI Chess Analysis That Explains the Why
analyze & train
Need a chess website? ChessWeb.site