Free Open-Source AI Chess Engine Maia 3 Released to Enhance Human Gameplay
The Maia Chess team has released a new open-source chess engine, Maia 3, trained on 250 million real human games. It reaches an Elo rating of about 1800—nearly 300 points higher than the previous version. Best of all, it is completely free and open-source, supports local deployment, and represents a major step toward making AI chess engines available to everyone.
Unique Approach: Simulating Human Decision-Making
Unlike top engines like AlphaZero and Stockfish, which aim for superhuman play with Elo over 3500, Maia focuses on understanding and mimicking human chess behavior. Its core goal is to "play like a human"—predicting the moves people actually make rather than finding mathematically optimal ones.
This human-centered training ensures Maia 3 avoids moves that feel unnatural to humans. Instead, it accurately simulates common patterns, psychological preferences, and even typical mistakes found in real games. This makes it a more educational and enjoyable companion for amateur club players, avoiding the frustration that often comes with playing against engines that are too strong.
Llama Architecture and Support for Multiple Board Games
Technically, Maia 3 is built on Meta's Llama 3.1 architecture with a decoder-only transformer design. Compared to the previous GPT-2-based version, this offers better context understanding and computational efficiency. Its training data covers skill levels from beginner to expert.
Notably, Maia 3 now supports not only chess but also Shogi, Go, and Xiangqi, making it a versatile platform for AI research into board games more broadly.
Smooth Performance on Consumer Hardware Advances AI Democratization
Maia 3 is released under the permissive Apache 2.0 license, allowing developers worldwide to freely use, modify, distribute, and even commercialize it. The full project code, training data, and model weights are already available on GitHub without restrictions.
To lower the entry barrier, the team has deeply optimized Maia 3 for consumer hardware. No high-end graphics card is needed—a regular laptop with a modern CPU can run it smoothly in local deployment. Players can already play against Maia 3 on the lichess.org platform, and the engine will provide free, high-quality support for future online game platforms and teaching tools.
Related article
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage
California AV Compliance: A New Era of Tickets, Geofences, and 1M Miles
Guident operates an AuveTech shuttle in South Florida, managing a four-mile route in West Palm Beach and a one-mile route in Boca Raton using its remote monitoring technology. | Credit: GuidentCalifornia is redefining the regulatory landscape for dri
Related Special Topic Recommendations
Comments (0)
0/500
The Maia Chess team has released a new open-source chess engine, Maia 3, trained on 250 million real human games. It reaches an Elo rating of about 1800—nearly 300 points higher than the previous version. Best of all, it is completely free and open-source, supports local deployment, and represents a major step toward making AI chess engines available to everyone.
Unique Approach: Simulating Human Decision-Making
Unlike top engines like AlphaZero and Stockfish, which aim for superhuman play with Elo over 3500, Maia focuses on understanding and mimicking human chess behavior. Its core goal is to "play like a human"—predicting the moves people actually make rather than finding mathematically optimal ones.
This human-centered training ensures Maia 3 avoids moves that feel unnatural to humans. Instead, it accurately simulates common patterns, psychological preferences, and even typical mistakes found in real games. This makes it a more educational and enjoyable companion for amateur club players, avoiding the frustration that often comes with playing against engines that are too strong.
Llama Architecture and Support for Multiple Board Games
Technically, Maia 3 is built on Meta's Llama 3.1 architecture with a decoder-only transformer design. Compared to the previous GPT-2-based version, this offers better context understanding and computational efficiency. Its training data covers skill levels from beginner to expert.
Notably, Maia 3 now supports not only chess but also Shogi, Go, and Xiangqi, making it a versatile platform for AI research into board games more broadly.
Smooth Performance on Consumer Hardware Advances AI Democratization
Maia 3 is released under the permissive Apache 2.0 license, allowing developers worldwide to freely use, modify, distribute, and even commercialize it. The full project code, training data, and model weights are already available on GitHub without restrictions.
To lower the entry barrier, the team has deeply optimized Maia 3 for consumer hardware. No high-end graphics card is needed—a regular laptop with a modern CPU can run it smoothly in local deployment. Players can already play against Maia 3 on the
DeepMind CEO Hassabis: I sleep six hours a day, usually feel energetic around 1 a.m.
Fortune recently featured an interview with Demis Hassabis, CEO of Google DeepMind, revealing his unconventional approach to rest and productivity. Hassabis disclosed that he sleeps very little, structuring his waking hours into two distinct work blo
OpenAI, Anthropic Vie for Market Share Despite Revenue Shortfalls
Despite recent reports suggesting OpenAI missed revenue targets, creating pressure on tech stocks this Tuesday, private AI lab investors remain resilient. Seasoned backers have confirmed they will not reduce investment despite negative media coverage





Home






