Gemini Model Tops New AI Rankings for Strategy Board Games

2026-02-04

Google's Gemini model outperforms competitors in board game benchmarks. Google Deepmind and Kaggle have expanded their "Game Arena" platform, introducing two new games: Werewolf and Poker. This platform tests AI models in strategic games, measuring various cognitive abilities—chess assesses logical thinking, Werewolf tests social skills like communication and deception, and Poker evaluates how models handle risk and incomplete information.

These games provide an objective way to measure skills such as planning and decision-making under uncertainty. Gemini 3 Pro and Gemini 3 Flash currently lead all rankings. The Werewolf benchmark also serves a dual function in safety research: it tests whether models can detect manipulative behavior without real-world consequences. According to Demis Hassabis, CEO of Google DeepMind, the AI industry requires more rigorous testing to properly evaluate the latest models.

RECENT AI NEWS

RECENT AI TOOLS