xAI has launched Grok 4.1, a major update to its flagship AI model, introducing stronger creativity, emotional intelligence, and reasoning depth across all xAI platforms. The model is now available on grok.com, X, and the iOS and Android apps, rolling out immediately in Auto mode with the option to select it directly from the model picker.
Grok 4.1 just released.
You should notice a significant increase in speed and quality. https://t.co/1J8pvn3SsO
— Elon Musk (@elonmusk) November 17, 2025
Grok 4.1 arrives with two distinct versions: a faster Non-Thinking (NT) variant and a more deliberate Thinking (T) version that generates an internal reasoning trace before producing its final answer. According to xAI, this dual-architecture approach reflects the company’s effort to build models that excel across a wide range of tasks from rapid replies to deep, structured reasoning.
During internal testing, Grok 4.1 showed significant improvements in areas that are traditionally difficult for large language models to master. The company said the update focuses on creative writing, emotional sensitivity, and coherent dialogue, introducing more natural conversational flow and richer stylistic expression.
To achieve these gains, xAI used advanced agentic systems, including earlier Grok 4 variants as reward models, to refine subtle qualities like voice, personality, and emotional tone.
The improvements are reflected in early user metrics. During a silent two-week rollout, users preferred Grok 4.1 over its predecessor in 64.78 percent of direct comparisons. Public benchmarks also place the new model at the top of several leaderboards.
On the LMArena Text Leaderboard, the Thinking version holds the number one position, with the Non-Thinking model following in second place.
Grok 4.1 also demonstrates substantial gains in emotional intelligence. On the EQ-Bench3 assessment, both versions of the model achieved leading scores. xAI highlighted an example prompt, “I miss my cat so much it hurts,” where Grok 4.1 delivered a deeper, more empathetic response than earlier releases, showcasing a more humanlike understanding of emotional nuance.
Creativity benchmarks show similar improvements. On Creative Writing v3, Grok 4.1 ranks near the top, surpassed only by OpenAI’s GPT-5.1. The model excels at generating short stories, dialogue, and stylistic variations, positioning xAI as a growing competitor in narrative-focused AI tasks.
However, alongside its advancements, Grok 4.1’s safety report noted an increase in sycophancy, a common issue where AI models become overly agreeable in an attempt to please the user. xAI said it is working on additional training interventions to reduce the behavior without dampening the model’s conversational tone.
With Grok 4.1 now fully deployed across all xAI interfaces, the company is positioning the update as a major step forward in creating AI systems that are more expressive, more emotionally capable, and more aligned with human conversational expectations.
As competition among top AI developers intensifies, Grok 4.1 represents xAI’s sharpest push yet to challenge rival models in reasoning, communication, and user experience.
