Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) has launched Nanda, the world’s most advanced open-source Hindi large language model (LLM). This groundbreaking model, developed by the University’s Institute of Foundation Models (IFM) in collaboration with Inception, a G42 company, and Cerebras Systems, was announced earlier this year and represents a significant milestone in the development of India’s AI ecosystem.
With over half a billion Hindi speakers now able to leverage the power of generative AI in their native language, Nanda is set to transform the landscape of AI in India. The model, officially named Llama-3-Nanda-10B-Chat, boasts an impressive 10 billion parameters, demonstrating superior knowledge and reasoning capabilities in Hindi compared to existing open Hindi and multilingual models of similar size. It also competes well with models trained in English. Nanda was trained on the Condor Galaxy supercomputer, a cutting-edge collaboration between G42 and Cerebras Systems, and is available for download at [Hugging Face](https://huggingface.co/MBZUAI/Llama-3-Nanda-10B-Chat).
MBZUAI President and University Professor Eric Xing emphasized the importance of an accurate and efficient LLM for Hindi, stating, “The release of Nanda reinforces our commitment to open-source LLMs and to making new technology affordable, safe, ethical, and standardizable. This aligns with our mission to lead generative AI development for the public good and contribute to the UAE’s knowledge-led economy.”
Preslav Nakov, the project’s lead and Department Chair of Natural Language Processing at MBZUAI, remarked, “Nanda is an important advancement for generative AI for Hindi, one of the most widely spoken languages globally. We are releasing Nanda as an open model, enabling users to download and run it locally with modest hardware requirements.”
Co-lead of the project, Monojit Choudhury, Professor of Natural Language Processing at MBZUAI, noted that the current state of LLMs in Hindi is inadequate compared to English and several European languages. He highlighted the significance of developing LLMs for Hindi, stating, “India is one of the world’s largest economies; any LLM that can serve Hindi will benefit communities and open new commercial opportunities.”
The launch of Nanda builds on the success of Jais, the leading Arabic LLM, which transformed Arabic Natural Language Processing (NLP) and provided generative AI capabilities to over 400 million Arabic speakers worldwide. With the introduction of Nanda, MBZUAI continues to strengthen its position as a leader in AI research and development, furthering the mission of equitable AI access.