xAI Releases Grok 4.1

ArticlexAIAnnouncement

xAI has launched Grok 4.1 across grok.com, X, and mobile applications. The update focuses on real-world conversational usability, emotional intelligence, and reduced hallucination rates.

Key points

  1. Now available on grok.com, X, iOS, and Android via Auto mode or direct selection in the model picker.
  2. Reinforcement learning workflow uses frontier agentic reasoning models as reward models to refine subjective signals like personality and tone.
  3. Achieved a 64.78% preference rate against the previous Grok production model during a two-week silent rollout on live traffic.
  4. Ranked #1 on the LMArena Text Leaderboard with 1483 Elo (Grok 4.1 Thinking) and #2 with 1465 Elo (Grok 4.1 Non-Thinking).
  5. Features lower hallucination rates on web search queries along with targeted evaluations on EQ-Bench3 and Creative Writing v3.

Why it matters

The release highlights a growing industry trend: leveraging advanced reasoning models as automated judges during reinforcement learning to systematically tune subjective conversational qualities at scale.

Analysis

With Grok 4.1, xAI addresses qualitative aspects of user interaction such as tone, empathy, and conversational flow, rather than purely focusing on synthetic benchmarks. The reported LMArena evaluation is notable because the non-thinking configuration reportedly scored higher than competing full-reasoning models. The long-term practical utility will largely depend on whether the reported reductions in hallucinations hold true for specialised professional and technical workflows.

Mentioned links

Summary generated by AI (Gemini 3.8 Flash, 27 September 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • Article:x.ai

    SpaceXAI Introduces Grok 4.6

    SpaceXAI has released Grok 4.6, placing special emphasis on long-running agents and interactive visual development. The model is available now via API, Cursor, and Grok Build.

    AI & LLMs· Announcement

  • Article:xAI

    xAI Introduces Grok 4.20 Multi-Agent Beta

    xAI has launched the Grok 4.20 Multi-Agent Beta model, designed for deep research tasks via multiple collaborating agents working in parallel.

    AI & LLMs· Announcement

  • Article:SpaceXAI

    SpaceXAI Introduces Grok Bot

    SpaceXAI has announced Grok Bot, a system of always-on AI agents designed to function as digital colleagues. Operating on their own cloud-hosted computers, the bots navigate existing apps and tools 24/7, completing multi-step tasks independently or coordinating among themselves.

    AI & LLMs· Announcement

  • Article:Anthropic

    Anthropic Announces Claude Opus 4.1

    Anthropic has released Claude Opus 4.1, offering upgrades in agentic workflows, real-world coding, and reasoning at the same pricing as Opus 4.

    AI & LLMs· Announcement

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.