Skip to content
← Models
Open weightsMBZUAI/10 trustMetrics-checked

NileChat

Open Egyptian-dialect LLMs that speak both Arabic script and Arabizi.

Visit official page →

About

NileChat is a family of open instruction-tuned models for Egyptian Arabic, developed by the MBZUAI France Lab in Paris as part of the Jais initiative. It handles both scripts Egyptians actually use: standard Arabic script and Latin-based Arabizi. The family spans a compact 4B model, mixture-of-experts variants with 6B active parameters, and a dense 12B model for complex multi-turn conversations, covering question answering, translation, and transliteration.

Egyptian Arabic is spoken by over 100 million people, roughly a quarter of the world's Arabic speakers, yet most models treat it as an afterthought. Note that a separate research project from UBC NLP released a 3B model under the same NileChat name for Egyptian and Moroccan dialects; the MBZUAI family is the larger, actively maintained line. Weights are open on Hugging Face with a research paper describing the training approach.

Specs

Context window

Pricing

Free quota

API access

Weights

Open weights

Released

July 1, 2025

Benchmarks

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

arabic-nlpdialect-aitranslation

Trust score — 7/10

/10 trustMetrics-checked

Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 96; Trending score 72/100 from researcher assessment.

Sources