#3 · Primary category: Game AI & Behavior Trees
lc0
Open source neural network chess engine with GPU acceleration and broad hardware support.
Project last updated:05/05/26
GitHub Stars
3.2K
Forks
603
Contributors
117
License
GPL-3.0
Why we included this project
Lc0 is a UCI-compliant chess engine that plays via a neural network instead of hand-crafted search rules, and it's the engine the LeelaChessZero project uses to turn its selfplay training into a real opponent. Its networks come from that project's large-scale selfplay, in the AlphaZero style, and the engine accelerates on CUDA, cuDNN, onnxruntime, and Apple Metal backends while still running on plain CPU, so it fits whatever hardware you have. If you want to see AlphaZero-style selfplay hold up in actual competitive play, or just want a strong free engine you can drop into any UCI GUI, Lc0 does both. Builds use Meson, and the release-branch workflow makes it straightforward to generate selfplay games or run match play.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.