Science

First superhuman Stratego AI, published in Nature

Samuel Sokota and colleagues built the first superhuman Stratego AI using general techniques they developed for reinforcement learning and test-time compute under imperfect information.

Role

Researcher, OpenAI

Posted

30 September 2026

Samuel Sokota@ssokota

In our Nature paper, we introduce the first superhuman Stratego AI, which we built using general techniques that we developed for RL & test-time compute under imperfect information. 1/N

30 Sep 2026View on X