Elon Musk’s company xAI has started releasing Grok 4.1, an upgraded version of its earlier Grok 4 model.
The company has launched two versions of the update. One is called Grok 4.1 and the other is Grok 4.1 Thinking. Both models are free to use, while paid users get higher usage limits.
xAI says Grok 4.1 is 3 times less likely to hallucinate than its older models. This makes it one of the strongest releases from the company so far. It is still unclear how well the model performs against competitors like GPT 5.1, which recently introduced improvements in overall performance and emotional intelligence.
However, early results from LMArena’s Text Arena give a clearer picture. Text Arena is an open-source platform that lets people compare large language models through blind, randomized tests. In the latest benchmarks, both versions of Grok 4.1 performed very well.
According to the leaderboard, Grok 4.1 Thinking reached the number one position with a score of 1510, while Grok 4.1 secured the nineteenth position with a score of 1437. The platform noted that this marks an improvement of more than forty points compared to Grok 4 Fast, which appeared only two months earlier.
Although Grok 4.1 seems like a strong upgrade, it may not end up being the most powerful model of the year. Google is reportedly preparing Gemini 3.0, which is expected to be its most advanced model yet.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
