Sam Altman’s AI model has defeated Elon Musk’s Grok in the first-of-its-kind chess tournament between leading artificial intelligence systems.
OpenAI’s o3 model swept xAI’s Grok 4 with a flawless 4–0 victory in the final of the Kaggle AI Exhibition Tournament, held on Thursday. Both models advanced to the final after defeating AI competitors from Anthropic, DeepSeek, Google, and Moonshot AI.
The matchup carried personal significance, as Mr Altman and Mr Musk co-founded OpenAI a decade ago before Musk left to launch his own rival AI company, xAI.
Their relationship has since broken down, with Mr Altman, who is now the chief executive officer of OpenAI, labelling his former friend a “bully” after Mr Musk made an unsuccessful attempt to buy the ChatGPT creator.
In response to Grok 4’s performance in the chess tournament, Mr Musk said that the AI model’s ability was a “side effect”.
He posted on X: “xAI spent almost no effort on chess.”
Computers have been able to defeat the best humans at chess since IBM’s Deep Blue supercomputer defeated the then-world champion Garry Kasparov in a six-game match in 1997.
Since then, Google’s DeepMind has developed AI programs capable of teaching itself chess and other games like Go at a superhuman level.
This is the first major tournament testing general-purpose large language models (LLMs) that have become popular since the launch of OpenAI’s ChatGPT in 2022.
“Kasparov vs. Deep Blue was a huge spectacle in the ‘90s... What’s important from this tournament is to understand more about how these products think and reason,” Mats André Kristiansen, CEO and co-founder of Take Take Take, told The Independent.
In July, current world number one chess player Magnus Carlsen beat ChatGPT in an online chess match without losing a single piece.
Commentating on the final between Grok and o3, Mr Carlsen estimated that Grok had a rating of around 800, while OpenAI’s model was at around 1200. By comparison, Mr Carlsen’s peak rating is 2882. When asked directly by a user on X, Grok estimated its rating at around 1600-1800.
