Chess demands deep reasoning, Werewolf tests social deduction, and poker introduces a powerful new challenge: risk management. Like Werewolf, poker is a game of imperfect information, but instead of forming alliances, players must navigate uncertainty by estimating opponents’ hands and adjusting their strategies accordingly. Success hinges not on luck, but on the ability to assess probabilities, read behavior, and make optimal decisions under pressure. To rigorously evaluate these complex cognitive skills, we are launching a new AI benchmark focused on Heads-Up No-Limit Texas Hold’em, one of the most challenging variants of poker. Alongside this, we’re hosting an AI poker tournament where top-performing models will compete head-to-head. The final leaderboard will be unveiled on Kaggle.com/game-arena on Wednesday, February 4, after the tournament concludes. For a deeper look at how we assess model performance in poker, including metrics for strategic reasoning and adaptability, visit the Kaggle blog. Watch the action To mark the launch of these new benchmarks, we’ve partnered with world-class experts: Chess Grandmaster Hikaru Nakamura and poker legends Nick Schulman, Doug Polk, and Liv Boeree. Together, they’ll host three live-streamed events featuring expert commentary and in-depth analysis across all three games—chess, Werewolf, and poker. Tune in daily at 9:30 AM PT on Kaggle.com/game-arena to witness the competition unfold. Explore the arena Whether it’s uncovering a brilliant checkmate, negotiating a fragile truce in Werewolf, or making a bold all-in move at the poker table, Kaggle Game Arena is where we test the true limits of AI reasoning, strategy, and adaptability. Discover what these models can do—join the experience at Kaggle.com/game-arena.