Inception unveils Jais: The Advanced Arabic Language Model for AI Enthusiasts
Mokshita P.
10X Technology
Published:

Inception unveils Jais: The Advanced Arabic Language Model for AI Enthusiasts

Collaboration between Inception, MBZUAI, and Cerebras Systems results in a game-changing model for Arabic AI Advancement, encouraging innovation and academic partnerships. It provides over 400 million Arabic speakers with access to the far-reaching potential of generative AI.

Inception, a G42 company focused on applied AI research and advancements, has released Jais, an open-source Arabic Large Language Model [LLM] that has been trained on an dataset encompassing a 395 billion tokens in both Arabic and English languages. Named after the UAE’s highest peak, Jais is a monumental achievement, boasting a 13-billion parameter design.

This creation is the result of a collaboration between Inception, the Mohamed bin Zayed University of Artificial Intelligence [MBZUAI], and Cerebras Systems. A milestone achievement, Jais was nurtured on Condor Galaxy, an multi-exaFLOP AI supercomputer developed jointly by G42 and Cerebras.

Jais carries profound implications for the Arabic-speaking world. As the first of its kind, this homegrown UAE innovation provides over 400 million Arabic speakers with access to the far-reaching potential of generative AI. By stimulating innovation, Jais reinforces Abu Dhabi's position as a global hub for AI, innovation, cultural preservation, and international cooperation.

Inception's decision to open-source Jais reflects their commitment to catalysing a vibrant Arabic language AI ecosystem, setting a model for other underrepresented languages in the AI landscape. The CEO of Inception, Andrew Jackson, emphasised that this endeavour aligns with their belief that collaboration fuels innovation. He expressed confidence that Jais will empower the Arabic language within the AI realm, promoting excellence and democratising AI's benefits.

Boasting superior performance, Jais outshines existing Arabic LLM models and competes favourably with English models of similar size, despite being trained on a comparatively smaller English dataset. This breakthrough underscores a symbiotic relationship between the Arabic and English components, ushering in a new era in LLM development.

MBZUAI President and University Professor Eric Xing highlighted the extensive research and nuanced understanding of the Arabic language that underpin Jais's development. Thanks to collaborative efforts with Inception and prominent global organisations, MBZUAI is poised to pioneer efficient, accurate, and effective LLMs.

Coinciding with the model's launch, Inception and MBZUAI forged an academic partnership granting early-release access to forthcoming Arabic LLMs for testing purposes. Reputable academic partners, including Carnegie Mellon University, Ecole Polytechnique, and NYU Abu Dhabi’s CAMeL Lab, are actively contributing to this dynamic initiative.

Multiple organisations, including the UAE Ministry of Foreign Affairs, Ministry of Industry and Advanced Technology, and leading corporations like ADNOC and Etihad Airways, are poised to leverage Jais for invaluable insights, further enhancing the model's capabilities.

Inception's transformational role at the nexus of academia, industry, and regulation positions it as a catalyst for synergistic growth, collaborative pursuits, and the accelerated commercialisation of AI across diverse sectors.