No note yet
A lightweight chess AI powered by a transformer neural network combined with Monte Carlo Tree Search (MCTS). Trained on 132M+ positions to evaluate and explore moves efficiently, blending modern deep learning with search-based decision making. Experimental, fast, and inspired by GPT-style architectures for chess without self-play.
FranceMember since 17 Jun 2026
Active
Time spent playing: 3 hours, 18 minutes