ApX logoApX logo

Grok 4.1

Parameters

-

Context Length

2M

Modality

Multimodal

Architecture

Dense

License

Proprietary

Release Date

17 Nov 2025

Knowledge Cutoff

-

Evaluation Benchmarks

Rank

#71

BenchmarkScoreRank

0.87

18

LiveBench Average

LiveBench Average

0.76

21

Professional Knowledge

MMLU Pro

0.84

23

General Text

auto
Text Arena

1466

37

General Text

Standard
Text Arena

1459

45

Web Development

WebDev Arena

1210

91

Rankings

Overall Rank

#71

Coding Rank

#90

About Grok 4.1

Grok 4.1 brings significant improvements to real-world usability with exceptional creative, emotional, and collaborative capabilities. Optimized for style, personality, helpfulness, and alignment using frontier agentic reasoning models as reward models. Achieves #1 on LMArena Text Leaderboard with 1483 Elo (thinking mode) and #2 with 1465 Elo (non-thinking), surpassing all other models. Features 2M context window, reduced hallucination rate (12.09% → 4.22% on production queries), and state-of-the-art emotional intelligence (1586 Elo on EQ-Bench). Available in both reasoning and fast non-reasoning modes through API.

Technical Specifications

Detailed architecture specifications are limited for proprietary models.

About Grok 4

xAI's frontier intelligence models trained with reinforcement learning at unprecedented scale using the 200,000 GPU Colossus cluster. Grok 4 series demonstrates state-of-the-art performance in reasoning, coding, and multimodal understanding with native tool use capabilities. Features real-time search integration across X and the web, advanced reasoning through scaled RL training, and industry-leading performance on academic benchmarks. Designed for both immediate responses and extended thinking modes with vision capabilities.


Other Grok 4 Models
Grok 4.1: Model Specifications and Details