Jais AI - UAE's Breakthrough Arabic Language Model Reshaping the Global AI Landscape
- Alick Mouriesse
- Mar 25
- 6 min read
Jais, the world's most advanced Arabic large language model (LLM), represents a significant milestone in bringing Arabic language capabilities to the forefront of artificial intelligence.
A U365 5MTS Microlearning 5 MINUTES TO SUCCESS Presentation |
Launched in August 2023, "Jais" open-source model has quickly established itself as a game-changer for the 400+ million Arabic speakers worldwide, while simultaneously demonstrating competitive performance in English language processing.
Origins and Vision Behind Jais
Named after the UAE's highest peak, Jais emerged from a strategic collaboration between Inception (a G42 company), Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), and California-based Cerebras Systems. The project addresses a critical gap in the AI landscape – the underrepresentation of non-English languages, particularly Arabic, in mainstream AI development.
The development team recognized that existing LLMs like OpenAI's GPT-4, Google's PaLM, and Meta's LLaMA offer Arabic language capabilities, but these were often described as "heavily diluted" and lacking cultural context. Jais was specifically designed to capture the linguistic nuances of various Arabic dialects and comprehend language, context, and cultural references, making it more accurate and contextually relevant than other models.
Funding and Strategic Importance
The Jais project represents a cornerstone of the UAE's ambitious national AI strategy. G42, an Abu Dhabi-based AI company chaired by UAE National Security Adviser Sheikh Tahnoon Bin Zayed Al-Nahyan, has positioned this initiative as part of the country's broader vision to become an AI leader in the Gulf region.
The strategic partnership between G42 and Cerebras Systems included a reported $100 million deal to provide nine supercomputers to G42. This substantial investment underscores the UAE's commitment to developing sovereign AI capabilities and establishing Abu Dhabi as a hub for AI innovation, cultural preservation, and international collaboration.
Evolution and Current Status
Jais was initially launched as a 13-billion parameter model trained on a dataset of 395 billion tokens (116 billion Arabic and 279 billion English). The training process took over 21 days and utilized the Condor Galaxy 1, a multi-exaFLOP AI supercomputer built by G42 and Cerebras.
The project has evolved significantly since its launch. By early 2025, Core42 (a G42 company) introduced JAIS 30B, a more advanced version with 30 billion parameters trained on a massive dataset of 1.63 trillion tokens, including 475 billion Arabic tokens. This enhancement has substantially improved performance metrics compared to the original JAIS 13B model.
Most recently, Core42 launched JAIS Chat as a mobile application available on iOS, bringing these advanced generative AI capabilities directly to users' smartphones.. . The application is designed to understand multiple Arabic dialects and cultural nuances while facilitating translation and cross-language communication at scale.
Comparative Performance in the Global AI Landscape
Jais outperforms existing Arabic models by a significant margin and demonstrates competitive performance against English models of similar size, despite being trained on less English data. This remarkable achievement suggests that the model's Arabic and English components have a symbiotic relationship, where learning in one language enhances capabilities in the other.
Before Jais, the UAE had developed Falcon, another open-source LLM created at the state-run Technology Innovation Institute in Abu Dhabi. However, Jais represents a substantial advancement in Arabic language modeling and multilingual model development in general.
Future Perspectives
The future looks promising for Jais as Inception and MBZUAI continue to expand and refine the model based on user community feedback. Several organizations, including UAE government ministries, ADNOC, Etihad Airways, and First Abu Dhabi Bank, have already partnered with Jais, providing valuable insights to enhance the model further.
Future development plans include adding functionality for working with images, graphs, and tabular data8. The open-source nature of Jais encourages scientific, academic, and developer communities to contribute to a vibrant Arabic language AI ecosystem, potentially serving as a model for other underrepresented languages in mainstream AI.
As the global AI landscape continues to evolve, Jais stands as a testament to the UAE's commitment to ensuring that innovation is accessible to everyone, regardless of language. By advancing Arabic language understanding and generation, Jais is ushering in a more linguistically inclusive and culturally aware era in artificial intelligence.
Please Rate and Comment
How did you find The book Essential? What has your experience been like using its content? Let us know in the comments at the end of that Page!
If you enjoyed this publication, please rate it to help others discover it. Be sure to subscribe or, even better, become a U365 member for more valuable publications from University 365.
BONUS - THE U365 AI ADVANTAGE |
How to Use Jais AI ?
To start using Jais AI, follow these steps:
Access the Model:
Find Jais-30B on Hugging Face : https://huggingface.co/inceptionai/jais-13b
Alternatively, access Jais through Azure Machine Learning studio.
You can also try the model online with a web browser after creating a account for free : https://jais.inceptionai.ai/ In that case the folowing steps do not apply.
Installation:
If using Hugging Face, install the Transformers library:
text
pip install transformers
For Azure, ensure you have an Azure subscription with a valid payment method.
Set Up the Environment:
Import necessary libraries:
python
import torch from transformers import AutoTokenizer, AutoModelForCausalLM
Load the Model:
python
model_path = "core42/jais-13b" tokenizer = AutoTokenizer.from_pretrained(model_path) model = AutoModelForCausalLM.from_pretrained(model_path, device_map="auto", trust_remote_code=True)
Generate Text:
Create a function to generate responses:
python
def get_response(text, tokenizer=tokenizer, model=model): input_ids = tokenizer(text, return_tensors="pt").input_ids generate_ids = model.generate(input_ids, max_length=200, do_sample=True) response = tokenizer.batch_decode(generate_ids, skip_special_tokens=True) return response
Use the Model:
Generate text in Arabic or English:
python
arabic_text = "عاصمة دولة الإمارات العربية المتحدة ه" english_text = "The capital of UAE is" print(get_response(arabic_text)) print(get_response(english_text))
Mobile Access:
For on-the-go use, download the JAIS Chat mobile app available on iOS.
Remember to comply with the terms of use when working with Jais AI4. For enterprise-grade deployments or specific use cases, consider exploring integration options through major cloud providers or Core42's Compass platform.
Resources:
Discussions To Learn Deep Dive - Podcast
Click on the Youtube image below to start the Youtube Podcast.
Discover more Dicusssions To Learn ▶️ Visit the U365-D2L Youtube Channel
Do you have questions about that Publication? Or perhaps you want to check your understanding of it. Why not try playing for a minute while improving your memory? For all these exciting activities, consider asking U.Copilot, the University 365 AI Agent trained to help you engage with knowledge and guide you toward success. U.Copilot is always available, even while you're reading a publication, at the bottom right corner of your screen. You can Always find U.Copilot right at the bottom right corner of your screen, even while reading a Publication. Alternatively, vous can open a separate windows with U.Copilot : www.u365.me/ucopilot.
Try these prompts in U.Copilot:
I just finished reading the publication "Name of Publication", and I have some questions about it: Write your question.
I have just read the Publication "Name of Publication", and I would like your help in verifying my understanding. Please ask me five questions to assess my comprehension, and provide an evaluation out of 10, along with some guided advice to improve my knowledge.
Or try your own prompts to learn and have fun...
Are you a U365 member? Suggest a book you'd like to read in five minutes,and we’ll add it for you! |
Save a crazy amount of time with our 5 MINUTES TO SUCCESS (5MTS) formula.
5MTS is University 365's Microlearning formula to help you gain knowledge in a flash. If you would like to make a suggestion for a particular book that you would like to read in less than 5 minutes, simply let us know as a member of U365 by providing the book's details in the Human Chat located at the bottom left after you have logged in. Your request will be prioritized, and you will receive a notification as soon as the book is added to our catalogue.
NOT A MEMBER YET?
DON'T FORGET TO RATE AND COMMENT ABOUT THAT PUBLICATION
Comentarios