Nova: A Policy-Based, Human-Like Chess Engine
Introducing Nova: a policy-only neural network trained on 500M Lichess positions. Benchmarked head-to-head against Maia, integrated into the Nova Chess training platform, and released open for research.
Read more →





