Arabian artificial intelligence models .. Where did we come so far?

Amid the global race for excellence in the artificial intelligence sector, the Arab countries are trying to develop their own models, which target an influential position in the digital future. In recent years, many Arab models have emerged, reflecting the region’s desire to play a major role in this important field. But the most prominent question remains: Can these models compete with world giants such as “Chatgpt”, “Gemini” and “Grok”? Or are it just alternatives that supplement Western models need more development to achieve the incredible innovation that promises artificial intelligence? The most prominent models of artificial intelligence introduced by the Arab countries “Humain Chat”, a new Saudi artificial intelligence model, known as the new generation “Allam” model, developed by the Saudi -Humain Company, owned by the “Public Investment Fund”. It is presented as the first public and free Arab model, and it is characterized by the ability to communicate in both Arabic and English, with a special focus on classical Arabic and local dialects such as Saudi Arabia, Egyptian, Jordanese and Lebanese, in addition to a design that reflects the cultural identity and local values. According to Tariq Amin, CEO of “Hummeen”, the use of this model is currently limited to the kingdom, but it will be introduced in the Middle East in October. Read more: Saudi Arabia offers Humain Chat for free. This creates artificial intelligence for businesses, Aramco Metaprein (Metabrain). Aramco announced the model of the Wooing IQ “Aramco Metaprin” at the “Lepe 2024” conference in Riyadh, which includes 250 billion laboratories, and is considered one of the largest industrial models in the world. He is trained in 7 trillion database of 90 years of oil and engineering operations. An internal is used to support Aramco employees in solving complicated industrial problems and to improve operational efficiency using language intelligence, immediate analyzes and quantum computer science. Look more: How does Aramco build its abilities in the field of artificial intelligence? These instruments fall within the framework of a more comprehensive plan adopted by the kingdom to reach the contribution of the artificial intelligence sector to GDP to 12% by 2030, which is equivalent to 130 billion dollars. More details in the video below: “Noor” is “Nour” is one of the Great Language Models (LLM) in Arabic, where the Institute of Technology Innovation (TII) in Abu Dhabi was launched in 2022, in collaboration with the French “Lightton”, with a size of 10 billion labs. He is trained in the largest Arabian data collection containing texts of the web, books, press, poetry and technical information, and its purpose is to provide a deep understanding of the classical. The model is available to academics and researchers without offering it to the public. Falcon is the Falcon series launched by the Institute of Technology Innovation in Abu Dhabi, one of the most important Arab initiatives in the field of major linguistic models. The story started with the Falcon-40B in May 2023, a large open source model of 40 billion laboratories, trained on a trillion a text symbol, and it was then considered a qualitative step for the UAE to compete with adults. The model soon developed into the Falcon-180B in September 2023, a larger version of 180 billion laboratories and trained at 3.5 trillion symbol to become one of the strongest Open Source models worldwide, and it is also freely available to developers through platforms such as ‘Hugging Face’. In 2025, the Institute of Technology Innovation presented new models, where “Falcon Arabic” appeared as a model dedicated to the Arabic language with a size of 7 billion laboratories, and trained in classical Arab data and regional dialects, giving him the ability to understand the linguistic and cultural diversity in the Arab world. The Institute says that the new model matches in its performance, other models that exceeded it ten times. The Emirati Falcon challenges the models of Gemini and Chatgpt. Details here. Parallel launched the Institute the ‘Falcon H1’ series, representing another technical approach, based on a new hybrid architecture to increase the velocity of low -memory -consumption conclusion. The range contains models from 500 million, 1.5 billion, 3 billion and 7 billion, to 34 billion laboratories for the largest model, and advanced results have been shown in multi -language and math analysis, and even used in practical applications. This series supports more than 100 languages ​​over a multi -language -intincible trained data, making it flexible in multi -environments. Jais Model is also considered an important, source -language model, developed by a partnership between the UAE Artificial Intelligence Company G42 (G42), Mohammed Bin Zayed University of Artificial Intelligence and Servas Systems. However, ‘G42’ has extracted its resources from it and focused on developing dedicated features based on models developed by other companies, such as ‘Openai’. “Fanar” is a large language model developed by the Qatar Computing Research Institute in collaboration with the Qatari Ministry of Communications and Information Technology and various parties, and is trained about 1.3 trillion text symbol and 300 billion Arabic words from different heritage, cultural and religious sources to ensure the consistency of the model with local identity. The model is characterized by an in -depth understanding of the Qatari dialect and Arabic dialects, and supports translation, summary, answering questions and generating images and sound through multimedia applications. Fannar contains a major linguistic model of 7 billion laboratories called “Fannar 7b” and a coach from scratch, as well as another “Fanar Prime” model with 9 billion laboratories. AQL The Egyptian Widebot Company launched the artificial intelligence model at the “Lepe 2025” conference with Saudi support as the first major Egyptian model of 7 billion treatment. The model is trained on Arabic and its dialects in addition to English, Persian, Turkish and Hebrew, and it is currently available for institutional use, and not as an open model to the public. The official Egyptian national model is still in implementation within the artificial intelligence strategy 2025-2030, and no other government model has been officially launched so far. Artificial Intelligence Model State / Developed Party Today Hearing (Laboratories) Language Humaine Chat Saudi Arabia / “Humaen” Saudi N / A Arabic and English, focusing on classic and local dialects Aramco Metaprin Saudi UAE Arabic / Commercial and research using 7 billion PCs and Regional Falcon H1 Emirates / General / 1.5 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / 3 billion / multi -long -lasting uae uae on “Academics and Researchers 10 Billion Arabic Fenar (Fenar 7B / Fanjar Prime) Qatar / Qatar Institute for Computing Research with the Ministry of Communications using 7 billion / 9 billion Qatari dialects and Arabic dialects, the spirit of Egypt / Wade -Bot business, and the Arab Dials, the Arab Dials, -Bave business, the 7 billion introduction. English, Persian, Turkish and Hebrew Source: Al Sharq How did Arab models use international expertise? The developed businesses and governments have used themselves to hold global technical partnerships to take advantage of advanced experiences and infrastructure in the development of Arab artificial intelligence models. The company, “Humamene”, developed by the ‘Humayen Chat’ model, has a partnership with the American Groq Company, the leading in the field of rapid reasoning processors, to provide Open Source linguistic models on the Groqcloud platform in the kingdom. It also signed $ 15 billion investment agreements at Amazon Web Service (AWS) to provide the advanced cloud structure, and the company “Advands Micro Device” (AMD) to provide high -performance calculations for the development of artificial intelligence infrastructure in Saudi Arabia. Also read: How does the Arabic region contribute to the racial models of obstetric intelligence? The Institute of Technology Innovation in Abu Dhabi has technically also worked with the US company “Invidia” to facilitate the deployment of “Falcon” on the foundation scale, as the “Falcon H1” model in the cloud “Envenia” service platform is to use it easily in cloud and hybrid commercial models. Google also worked as a ‘technical partner’ to develop the Qatari “Fanar” model, where the Google Cloud platform provided the model to offer and train the model, which allowed it ‘Fanar’ utilized a global infrastructure. Are Arabian Artificial Intelligence Models real alternatives to their global peers? In recent years, several Arab models have emerged as important steps in the field of artificial intelligence. These models have proven qualitatively in the Arabic language, as models such as “Falcon Arabic” and “Nour” have proven extraordinary ability in the Arab language, either in the production of smooth texts or better than larger models in Arabic tasks. Arab models form an important addition to the bridge of the linguistic and cultural void, as it serves the Arab user in his language and dialect and takes into account its peculiarities and culture. It also stimulates the establishment of an artificial intelligence infrastructure, for example the “Stargate Emirates” project, which is expected to provide an infrastructure for artificial intelligence and computer capabilities covering a 2000 miles radius, to half the world population. But before these Arab models, a long way to cause a global penetration as it has to practice in the trillion symbols and constantly undergo an improvement.