Inception, a G42 company focused on advanced AI models and applications, has released the JAIS 70B large language model (LLM). This 70 billion parameter model is designed for developers of Arabic-based natural language processing (NLP) solutions and is aimed at accelerating the integration of generative AI services across various industries, particularly in customer service, content creation and data analysis.
“AI is now a proven value-adding force, and large language models have been at the forefront of the AI adoption spike,” said Dr Andrew Jackson, CEO of Inception. “JAIS was created to preserve Arabic heritage, culture, and language and to democratise access to AI.
“Releasing JAIS 70B and this new family of models reinforces our commitment to delivering the highest-quality AI foundation model for Arabic-speaking nations.”
In addition, the company has introduced a suite of JAIS foundation and fine-tuned models, comprising 20 models across 8 sizes, ranging from 590 million to 70 billion parameters. These models are specifically fine-tuned for chat applications and trained on up to 1.6 trillion tokens of Arabic, English and code data.
“For models up to 30 billion parameters, we successfully trained JAIS from scratch, consistently outperforming adapted models in the community,” said Neha Sengupta, Principal Applied Scientist at Inception. “However, for models with 70 billion parameters and above, the computational complexity and environmental impact of training from scratch were significant.”
The release includes tools such as the first Arabic-centric model that can run on a laptop. It offers both small, compute-efficient models for targeted applications and advanced models for enterprise use.
