Elon Musk's xAI launches Grok 4.1 with improved emotional intelligence and creative writing capabilities

Elon Musk's xAI has launched Grok 4.1, a next-generation AI model that dominates industry leaderboards in emotional intelligence and creative writing while slashing factual errors by nearly two-thirds.

The upgraded chatbot is now rolling out across grok.com, X, and iOS and Android apps.Grok 4.1 claimed the number-one spot on LMArena's Text Arena with a commanding 1,483 Elo score—31 points ahead of its nearest competitor. Even without extended reasoning, the model scored 1,465 Elo, outperforming every rival's most advanced configuration. During a two-week stealth rollout starting November 1, users chose Grok 4.1 over its predecessor nearly 65% of the time.

Grok 4.1 masters empathy while cutting misinformation rates

The chatbot scored 1,586 on EQ-Bench3, a rigorous emotional intelligence assessment, marking a dramatic leap from Grok 4's score of 1,206. xAI credits the gains to a novel training approach using advanced AI systems as evaluators during reinforcement learning, enabling the model to refine conversational nuance and empathy.Factual accuracy improved substantially, with hallucination rates plummeting from 12.09% to 4.22% on real-world queries.

On FActScore, which tests biographical precision across 500 questions, errors dropped from 9.89% to under 3%. The improvements address a persistent weakness in fast-response AI models that sacrifice depth for speed.

Creative writing performance rivals top AI systems worldwide

In creative writing benchmarks, Grok 4.1 earned a 1,708.6 Elo rating, positioning it among the world's elite AI models. The company engineered the upgrade to enhance personality consistency, conversational appeal, and collaborative responsiveness while preserving technical precision. xAI emphasized that Grok 4.1 better interprets subtle user intentions and maintains a distinctive voice throughout extended interactions.tio