Tier

Moderate

Headquarters

Palo Alto, CA

Founded

2023

Flagship

Grok

Models tracked

22

Artificial Analysis

Top intelligence

53.8

Artificial Analysis Index

Cheapest model

$0.00

blended /1M

Fastest output

141

tokens / sec

Flagship Model

xAI’s top-scoring model on the Artificial Analysis Intelligence Index.

Most intelligent model

Grok 4.5 (high)

Released 2026-07-08

Intelligence

53.8

Coding

72.4

Math

In $/1M

$2.00

Out $/1M

$6.00

Speed t/s

55

Benchmark breakdown

GPQA Diamond
93
Humanity’s Last Exam
40
SciCode
54
τ²-Bench (agentic)
Terminal-Bench Hard
IFBench

Models & benchmarks

Every xAI model on file — intelligence, coding, price, and speed.

ModelReleasedIntelCodingMathIn $/1MOut $/1MSpeed t/sLatency s
Grok 4.5 (high)2026-07-0853.872.4$2.00$6.005510.91
Grok Build 0.1 06162026-06-1639.851.5$1.00$2.0000.00
Grok 4.3 (high)2026-04-3037.642.2$1.25$2.5000.00
Grok 4.20 0309 v2 (Reasoning)2026-04-0737.0$2.00$6.0000.00
Grok 4.20 0309 (Reasoning)2026-03-1036.5$2.00$6.0000.00
Grok 4.3 (medium)2026-04-3036.0$1.25$2.5014115.74
Grok 4.3 (low)2026-04-3035.4$1.25$2.501276.93
Grok 42025-07-1033.392.7$3.00$15.0000.00
Grok 4.1 Fast (Reasoning)2025-11-1930.689.3$0.00$0.0000.00
Grok 4 Fast (Reasoning)2025-09-1927.489.7$0.20$0.5000.00
Grok 4.3 (Non-reasoning)2026-04-3024.835.2$1.25$2.501200.88
Grok 3 mini Reasoning (high)2025-02-1922.584.7$0.30$0.5000.00
Grok 4.20 0309 (Non-reasoning)2026-03-1022.5$2.00$6.0000.00
Grok 4.20 0309 v2 (Non-reasoning)2026-04-0721.8$2.00$6.0000.00
Via Artificial Analysis

Trajectory

xAI’s model releases over time, by intelligence.

Mar 2024 · 6.0▲ +47.8 indexJul 2026 · 53.8

House strengths

Where xAI’s lineup over-indexes, averaged across its models.

Governance

xAI’s license posture and openness across its lineup.

Open weight

0

of 22 models

Proprietary

22

of 22 models

Commercial-OK

0

license permits

Modalities

Image · Text

Proprietary · 22

Via LLM Stats

Publications

xAI’s own papers, frameworks, and cards in the 44B Library.

© 2026 NYSGPT2525 LLC