xAI has released a new version of its chatbot, Grok 4.1, which is better, more emotional, and more accurate than before. It’s now available on the Grok website, X, and mobile apps. The update was quietly tested two weeks before the full release.

Tests show that users often prefer Grok 4.1 over the old model. It ranks very high on major AI leaderboards in both rational and irrational modes. The xAI model’s style, personality, and helpfulness have been improved using advanced training methods.

The company has significantly reduced errors and reduced factual error rates. Grok 4.1 now provides more accurate, empathetic, and constructive responses, and is working with xAI benchmark creators to share the results with the public.