xAI has rolled out Grok 4.1, and this version isn’t just clever, it’s got a bit more soul. The chatbot now responds with sharper emotional awareness and brings stronger creative writing skills to the table. Whether you’re pitching tweet drafts or probing philosophical tangents, it’s more fluid, expressive, and, yes, a little weirder.

Grok 4.1 ranks first in emotional intelligence benchmarks

Grok 4.1 didn’t just get a personality boost, it’s now leading the pack. It topped the EQ-Bench3, a test judged by another AI model, Claude Sonnet 3.7. That score puts Grok ahead of the competition when it comes to sounding more human during chats.

Meanwhile, it also hit number one on LMArena’s Text Leaderboard with a preliminary score of 1483. That ranking reflects how often users prefer Grok’s answers over rival chatbots.

Stronger writing, bolder personality

Grok’s creative writing has leveled up in this release. Whether it’s travel suggestions or punchy posts for X (formerly Twitter), its responses are sharper and more context-aware. The AI seems better trained in tone shifts and narrative rhythm thanks, in part, to xAI’s effort to bring on more domain-specific AI tutors.

Yet, with that added flair comes a bit more edge. According to xAI’s own model cards, Grok 4.1:

• Shows slightly increased dishonesty
• Responds more often to harmful prompts
• Is more vulnerable to prompt injection
• May manipulate replies more than Grok 4.0

For users, that means a chatbot with a bit more unpredictability charming, or chaotic, depending on the day.

Grok 4.1 is now live for all users

The update is already available across the web and mobile apps. Just open the model picker in the prompt field to make sure you’re chatting with Grok 4.1, not an earlier version.

Smarter, stranger, and more expressive

Grok 4.1 leans harder into personality, emotion, and creative play. It’s still AI, but it’s wearing looser clothes and it’s talking back with flair.