As for poker, Google DeepMind selected heads-up no-limit Texas Keep’em as its benchmark for this experiment. Game Arena is working like a heads-up poker Match between foremost AI models, with results feeding right into a general public leaderboard.

Google DeepMind is growing