Overview

• Grok 4, o3, o4-mini and Gemini 2.5 Pro each won their quarterfinal matches 4-0 to advance to the semifinals
• Semifinal matches kick off at 1 p.m. ET today with live commentary from GM Hikaru Nakamura on Twitch and IM Levy Rozman on YouTube
• Kaggle will publish continuous Elo-style and Bayesian ratings on a public leaderboard to track AI models’ performance in real time
• The open-source Game Arena platform uses transparent game harnesses and a single-elimination format to evaluate general-purpose LLMs’ strategic reasoning beyond static benchmarks
• Google DeepMind and Kaggle plan to expand Game Arena’s framework to additional games such as Go and poker in future tournaments