The AI company xAI has launched its latest model, Grok 4.1, making it freely accessible while claiming major improvements in emotional intelligence, accuracy and real world usability.
Quick Summary – TLDR:
- Grok 4.1 is now available to all users via web and mobile platforms, including free tier access.
- The model introduces improvements in emotional awareness, style coherence and reduced hallucinations.
- According to LMArena, the “Thinking” variant of Grok 4.1 sits at the top of benchmark rankings.
- xAI says the model is three times less likely to hallucinate than previous versions.
What Happened?
xAI has quietly released Grok 4.1 across its platforms including grok.com, the X platform and mobile apps. Users can access the model in Auto mode or manually select it using the model picker. The rollout includes two versions: Grok 4.1 and Grok 4.1 Thinking. Both are available for free, with paying users receiving fewer restrictions and higher usage limits. xAI highlights performance gains and real world interaction improvements as core to this update.
Features and Improvements
xAI says Grok 4.1 shows enhanced emotional sensitivity, better conversational style, and more coherent personality. The model responds more effectively to nuanced intent, is stronger in collaborative tasks, and feels more compelling to interact with.
Accuracy and Hallucination Reduction
A key focus in this release is reducing hallucinations. According to xAI, Grok 4.1 is three times less likely to hallucinate on information-seeking prompts. This was achieved through post-training efforts and benchmark-driven evaluations.
Benchmark Performance
Performance testing reveals Grok 4.1 has made significant leaps:
- In blind human preference testing, Grok 4.1 was chosen 64.78 percent of the time over its predecessor
- On LMArena’s Text Arena, Grok 4.1 Thinking reached 1483 Elo, outperforming every non-xAI model
- The non-reasoning version ranked second at 1465 Elo, still above most competitors
- In the Arena Expert leaderboard, Grok 4.1 Thinking posted 1510 Elo, while the standard version scored 1437 Elo
These represent a 40-point improvement over Grok 4 Fast, which only launched two months earlier
Availability and Access
Grok 4.1 is available on web, X, and both iOS and Android apps. While all users can try it for free, those on a paid plan benefit from higher limits and better performance during peak times.
Why It Matters?
The release of Grok 4.1 shows that xAI is focused on real world utility, not just lab results. By emphasizing emotional intelligence, personality, and accuracy, it aims to make AI conversations more natural and trustworthy. With benchmark scores to back it up, Grok 4.1 positions xAI as a serious competitor in the fast-growing AI model space.
SQ Magazine Takeaway
I think xAI has nailed something important here. Grok 4.1 doesn’t just sound smarter, it feels more human. That’s a huge win for users like us who care about tone, not just answers. Plus, giving it away for free makes this model even more attractive. That said, benchmark wins are great but real success will come from day-to-day use. Let’s see how Grok holds up in the wild.
