OpenAI’s ChatGPT defeats Elon Musk’s Grok in AI chess showdown

ChatGPT

OpenAI’s o3 model has triumphed over Elon Musk’s xAI model Grok 4 in the final of a Kaggle-hosted tournament to determine the strongest AI chess player. While chess has long been used as a benchmark for AI progress, this event differed from classic machine-vs-human matches — featuring large language models designed for general use, not specialised chess engines.

Over three days, eight AI systems from OpenAI, xAI, Google, Anthropic, and Chinese developers DeepSeek and Moonshot AI competed. According to a report by BBC, Google’s Gemini model secured third place after defeating another OpenAI entry. Despite early dominance, Grok 4 faltered in the final, committing repeated blunders — including multiple queen losses — against o3’s steady play.

“Up until the semi finals, it seemed like nothing would be able to stop Grok 4,” said Chess.com writer Pedro Pinhata. But on the last day, Grok’s “unrecognisable” play collapsed under pressure. Chess grandmaster Hikaru Nakamura, commentating live, noted: “Grok made so many mistakes in these games, but OpenAI did not.”

Musk downplayed the loss, claiming Grok’s earlier success was a “side effect” since xAI “spent almost no effort on chess.” Still, the upset adds fuel to the ongoing rivalry between Musk and OpenAI CEO Sam Altman, both former OpenAI co-founders who now lead competing AI ventures.

Chess remains a favoured test for AI developers, assessing reasoning, strategic planning, and adaptability. Past milestones include DeepMind’s AlphaGo defeating human Go champions, a feat that prompted South Korean Go master Lee Se-dol to retire, declaring, “There is an entity that cannot be defeated.”