Gemini 3.0 is one of the worst kept secrets. It has many AI enthusiasts excited. Before it has gotten here, both OpenAI and xAI have introduced new models. Grok 4.1 has improved significantly over the previous version. As the company explains, they used:
the same large scale reinforcement learning infrastructure that powered Grok 4 and applied it to optimize the style, personality, helpfulness, and alignment of the model. In order to optimize these non-verifiable reward signals, we developed new methods that let us use frontier agentic reasoning models as reward models to autonomously evaluate and iterate on responses at scale.
In LMArena’s Text Arena, Grok 4.1 Thinking holds the #1 position with 1483 ELO, which 31 points over the highest non-xAI model. Its non-reasoning model is sitting at #2 with 1465 ELO. Grok 4.1 is incredible in creative writing, only coming second to Polaris Alpha, which is an early GPT 5.1 model. Hallucinations have been reduced as well. On Arena Expert leaderboard, Grok 4.1 thinking ranks at #1 with a score of 1510.
[HT]
