OpenAI's o3 model has become the champion in an international chess tournament featuring leading artificial intelligence systems, including Google and xAI. By defeating Elon Musk's xAI-owned Grok 4 model in the final, o3 has further intensified the competition among technology giants.
The editorial office reports that a total of eight large language models from OpenAI, xAI, Google, Anthropic, as well as Chinese companies DeepSeek and Moonshot AI, competed in the three-day event held on Google's Kaggle platform. In the final standings, OpenAI's o3 model placed first, Grok 4 second, and Google's Gemini model, after defeating another OpenAI model, secured third place.
Although the final match began tensely, Grok 4's consecutive serious errors caused the game to shift in OpenAI's favor. Chess.com writer Pedro Pinhata noted that Grok appeared unrivaled until the semifinals but lost its advantage in the final. Renowned grandmaster Hikaru Nakamura commented live, “Grok made many mistakes, while OpenAI made no mistakes.”
Experts emphasize that strategy games with complex rules, such as chess, are an important tool for evaluating artificial intelligence's problem-solving, planning, and outcome-oriented thinking skills, and this competition has once again highlighted its relevance.
Farid Alizade