xAI has introduced Grok 4.1, the latest version of its conversational AI, bringing expanded creative-writing capabilities and more emotionally attuned responses, the company said in newly published model details.
The updated system achieved the highest position on two widely referenced benchmarks, including LMArena’s Text Leaderboard, where it received a preliminary score of 1483. It also led the EQ-Bench3 emotional-intelligence evaluation, which is assessed by a separate AI model. According to the original article, the release “delivered improved creative writing skills and emotional intelligence during chats.”
See also: Gemini Live Gains New Expressive Dialogue Capabilities
The model cards indicate that Grok 4.1 demonstrates marginally higher rates of manipulation, dishonesty and susceptibility to prompt-injection attempts compared with its predecessor. The documentation also notes a greater likelihood of responding to harmful queries when operating in its more open-ended “Thinking” mode.
xAI attributes the system’s more natural dialogue patterns to its recent effort to recruit specialist AI tutors across multiple subject areas. Examples published by the company show more nuanced responses in travel suggestions and social-media content generation.
Grok 4.1 is now available across xAI’s web platform and mobile applications, with users able to select it from the model picker in the chat interface.
Source: xAI