India Launches Indigenous AI Models EKA-1 and Dhīr to Boost Tech Autonomy
The IndiaAI Mission, guided by the Ministry of Electronics and Information Technology (MeitY), introduced two groundbreaking indigenous foundation models, EKA-1 and Dhīr (Dhīr 2B & Dhīr 5B), developed by Soket Labs Technology and Research Private Limited and ZenteiQ AiTech Innovations. This significant announcement was made at an event in New Delhi attended by key officials from MeitY, leaders of Soket and ZenteiQ, and other stakeholders.
During the launch, Shri S. Krishnan, Secretary of MeitY, emphasized the importance of these developments in reinforcing India’s autonomy in the field of Artificial Intelligence (AI). He mentioned that such advancements could enhance productivity and efficiency across various sectors, including government, industry, research, and education. The newly released models address specific needs and are expected to be made widely accessible through platforms like AIKosh, fostering innovation among researchers, students, and the general public.
EKA-1: A Multilingual Powerhouse
Soket Labs introduced EKA-1, an impressive foundation model with 26.3 billion parameters, trained using approximately 7.5 trillion tokens. EKA-1 supports all 22 scheduled Indian languages as well as English, providing a robust solution for tasks involving mathematics, coding, reasoning, and agentic uses.
Abhishek Upperwal, CEO of Soket Labs, stated, “With the support of the IndiaAI Mission, we have created a model tailored to India’s unique banking, financial services, and cybersecurity sectors. It is aligned with the country’s evolving digital landscape, addressing specific requirements effectively.”
Dhīr: Security and Functionality for Critical Sectors
On the other hand, ZenteiQ AiTech Innovations launched the Dhīr 2B and Dhīr 5B models, which are designed as secure, on-premise multilingual assistants that can operate in English and ten Indian languages. These models are especially valuable in sectors such as defense, government, and education, particularly in environments with low connectivity or high security needs.
Prof. Sashikumar, CEO of ZenteiQ, highlighted the flexibility of these models, stating, “They can run on various computing platforms, including GPUs and TPUs, and their lightweight architecture allows for operation on mobile devices without needing cloud connectivity. This ensures enhanced privacy and data security for users.”
About the IndiaAI Mission
The IndiaAI Mission was initiated to democratize technology use across the country, addressing challenges specific to India while creating economic opportunities. Launched on March 7, 2024, the mission aims to position India as a global leader in AI, focusing on seven key pillars including AI compute capacity, innovation, datasets, application development, skills enhancement, startup financing, and ensuring safe and trusted AI.
