Skip to content

announced a strategic investment and partnership with Arabic.ai.

Read the announcement
Talk to our team

LLM-X ranks #1 on Stanford HELM Arabic.

A 91.9% mean score, first of 41 models evaluated, ahead of the leading frontier models.

See the benchmarks

91.9%

LLM-X mean score on Stanford HELM Arabic

#1 of 41 models

  1. LLM-XArabic.ai91.9
  2. Best frontier modelHELM Arabic84.1

41 models evaluated. LLM-X is first.

One family for text, speech and vision.

Three language models, two speech models and vision models, each with its own role.

For questions that need careful reasoning, synthesis and a developed draft: research briefs, document synthesis and substantial drafting.

Works in Atelier

Efficient, high-volume processing for clearly bounded tasks that recur across an organization.

Works in Sadu

Translation where meaning, terminology and the audience all matter.

Works in Atelier Translation Studio

Text-to-speech in natural Arabic: six voices across five dialects, in under a second.

Works in Atelier Voice Studio

Speech-to-text for calls and recordings, delivered as editable transcripts.

Works in Atelier Voice Studio

Vision models for Arabic images and documents.

Enterprise results, with their scope.

Stanford CRFM HELM Arabic Enterprise, v1.1.0, published 25 June 2026. Macro-averaged mean score across Arabic content generation, financial reasoning and UAE legal question answering.

  • LLM-XArabic.AI LLM-X (2026-06-07)0.825
  • LLM-SArabic.AI LLM-S0.724

Benchmark results only. They do not measure production reliability or autonomous decision-making.

Run each model where your policy requires.

Model deployment is scoped to the model you choose. Compute, model version and any disconnected requirements are agreed together.

See deployment options
A private server hall

Make your next ambition operational.

Bring intelligence into your knowledge, people and operations. Talk to our team about the work you want to transform.

This page does not set marketing or analytics cookies, so there is nothing to change here.