Maia3-79M

Maia3-79M

Part of the Maia3 family of transformer models for human chess move prediction. This is the 79M-parameter variant.

For full details (architecture details, training recipe, full evaluation, and ablations) see our paper Chessformer: A Unified Architecture for Chess Modeling (ICLR 2026).

Model summary

• Family: Maia-3, human move prediction models built on the Chessformer architecture
• Architecture: encoder-only transformer with board squares as tokens, augmented by Geometric Attention Bias (GAB), a dynamic positional encoding that adapts to the geometry of chess, and an attention-based source-destination policy head
• Parameters: 79M
• Task: predicting the move a human player of a given skill level would make from a given position
• Training data: Lichess human games, January 2023 – July 2025
• License: AGPLv3

Intended use

Maia-3 models predict human chess moves conditioned on player rating. Typical uses include:

• Research on human chess modeling and human–AI alignment
• Tools for chess education and entertainment
• Move-suggestion and analysis tools that emulate play at a chosen rating
• Mechanistic interpretability research: the square-token design makes attention patterns and activations directly attributable to board squares

Not intended for maximum playing strength. For strong engine play built on the same architecture, see the Chessformer integration into Leela Chess Zero described in the paper.

How to use

Maia3-79M is a PyTorch checkpoint trained with the code at CSSLab/maia3. Clone that repo, set up the conda environment, and load the checkpoint following the instructions in its README.

Architecture hyperparameters for this variant are defined in ablate_size.sh in the training repo.

Training

• Data: Lichess monthly game dumps, January 2023 – July 2025
• Code: CSSLab/maia3
• Config: size ablation row corresponding to 79M parameters in ablate_size.sh

Evaluation

The Maia-3 family reaches 57.1% move-matching accuracy on human moves, significantly surpassing the previous state of the art with fewer than a quarter of the parameters. Per-size accuracy curves, scaling analysis, and skill-conditioned breakdowns are reported in the paper.

Citation

@inproceedings{monroe2026chessformer,
  title={Chessformer: A Unified Architecture for Chess Modeling},
  author={Daniel Monroe and George Eilender and Philip Chalmers and Zhenwei Tang and Ashton Anderson},
  booktitle={The Fourteenth International Conference on Learning Representations},
  year={2026},
  url={https://openreview.net/forum?id=2ltBRzEHyd}
}