New Delhi: The IndiaAI Mission on Tuesday unveiled three indigenous artificial intelligence foundation models, EKA-1, Dhīr 2B and Dhīr 5B, marking a significant step towards strengthening India’s domestic AI capabilities and technological self-reliance.
The models were developed by Soket Labs Technology and Research and ZenteiQ AiTech Innovations under the IndiaAI Innovation Centre’s Foundation Models pillar, which aims to promote the development of homegrown large language models.
The unveiling event was attended by IT Secretary S Krishnan, senior Ministry of Electronics and Information Technology (MeitY) officials and representatives of the two companies.
EKA-1, developed by Soket Labs, is a 26.3-billion-parameter model trained on approximately 7.5 trillion tokens. Designed for Indian languages, banking and cybersecurity applications, it supports all 22 scheduled Indian languages along with English.
The company said the open, sparse hybrid model is designed to handle mathematics, coding, reasoning and agentic applications. It can also perform certain tasks directly on smartphones without requiring mobile data or Wi-Fi connectivity.
Meanwhile, ZenteiQ’s Dhīr 2B and Dhīr 5B are open-source models supporting English and 10 Indian languages. They are designed for document-based automation, knowledge retrieval and multilingual assistance, particularly in sensitive or low-connectivity environments.
The models can be deployed on-premise and across computing platforms, including GPUs and TPUs, allowing users to process information locally while retaining data privacy and security.
The government said subsidised access to national AI computing infrastructure and assistance with research-related expenses had facilitated their development.
Krishnan said indigenous AI models would strengthen India’s technological and strategic autonomy while creating opportunities to enhance productivity across government, industry, research and education.
He also highlighted platforms such as AIKosh as a means of making these models accessible to researchers, students, innovators and the wider public.
