Elon Musk's xAI edges past ChatGPT and Gemini with new Grok 4.1 update: here's what you need to know

Elon Musk led xAI has launched its new Grok 4.1 model which it has begun rolling out to all its users. xAI claims that its new model is ‘exceptionally capable’ in creative, emotional, and collaborative interactions and is more perceptive to nuanced intent, compelling to speak with, and coherent in personality.

The new Grok 4.1 model went through a ‘silent rollout’ from the start of this month till 14 November where the new LLM was added to Grok website, X and the Grok mobile apps.

Grok overthrows ChatGPT and Gemini:
Perhaps for the first time since its launch, Gemini 2.5 Pro has been dethroned on LMArena leaderboard on Text related tasks where it has been Grok 4.1 Thinking and Grok 4.1 on the number 1 and number 2 spot.

The new Grok model is also ahead of the likes of Anthropic's Claude and OpenAI's ChatGPT.

Grok 4.1 (thinking) once again topped the rankings on EQ-Bench, a LLM judged benchmark that evaluaes the emotional intelligence abilities, understanding, insight, empathy, and interpersonal skills. The top two models on the benchmark were Grok 4.1 (thinking) and Grok 4.1 while Kimi-K2 was on the third spot and Gemini 2.5 Pro and GPT-5 were on 5th and 6th spot respectively.

Grok 4.1 models were on 2nd and 3rd spot on Creative Writing v3 benchmark which asks models to models generate responses to 32 distinct writing prompts across 3 iterations. An early version of OpenAI's GPT-5.1 topped the benchmark while OpenAI's o3 was on the 4th spot.

What's new Grok 4.1?

xAI claims that Grok 4.1 has reduced hallucinations (making stuff up) as compared to its predecessor.

While comparing Grok 4.1's hallucinations to that of Grok 4.0, xAI says that it tested both models on real-world information-seeking queries where the new model had a hallucination rate of 4.22% compared to the 12.09% hallucination rate of Grok 4.0.

Meanwhile, on FActScore, a public benchmark consisting of 500 biography questions on individuals, Grok 4.1 had a 2.97% score compared to 9.89% score of its predecessor, signalling a great improvement in hallucination at least on benchmarks.

xAI says in terms of real-world changes, users would notice that Grok 4.1 is much nicer to talk to, more understanding and more helpful than its predecessor.

The new model launch by xAI comes just days after OpenAI's also released its GPT-5.1 model while rumours are ripe about Google releasing its Gemini 3.0 model.

Recently, Elon Musk had essentially confirmed that Grok 5, the next frontier model from xAI is being pushed to 2026. The model was earlier confirmed by Musk by the end of 2025 and be “crushingly good” but the billionaire now says that it will be released in the first three months of 2026.