Grok 4, the latest large-language model from Elon Musk’s xAI, is currently topping the leaderboard in the Kaggle AI Chess Competition, according to results posted on 5 August. The model is outperforming rival systems from Google’s Gemini project, OpenAI, Anthropic’s Claude and China-based DeepSeek despite xAI having invested little effort in chess-specific training, developers involved in the contest said. Chess grandmaster Hikaru Nakamura endorsed the performance, calling Grok 4 "easily the best so far" and saying he was “not surprised” by the outcome. The strong showing underscores the escalating rivalry among leading AI labs to demonstrate prowess in specialized tasks such as game-play, even as they race to commercialize general-purpose models.