India’s Homegrown AI: EKA-1, Dhir Models Enhance Digital Ecosystem

India’s technological sovereignty takes a significant leap forward with the launch of indigenous AI foundation models EKA-1, Dhir 2B, and Dhir 5B, poised to transform sectors from banking to defence with their multilingual and offline capabilities.

The IndiaAI Mission, under the Ministry of Electronics and IT (MeitY), on Tuesday released indigenous foundation models EKA-1, developed by Soket Labs Technology and Research, and Dhir 2B and Dhir 5B, developed by ZenteiQ AiTech Innovations, marking another step in the country’s push to build homegrown AI capabilities.

Advancing India’s AI Capabilities

The models were unveiled at an event attended by IT Secretary S Krishnan, senior officials from the ministry and IndiaAI, as well as the leadership of both companies.

These three indigenous models are a key milestone under the IndiaAI Mission, showcasing India’s progress in creating homegrown AI technology.

Dhir 2B and Dhir 5B enable offline-capable, document-grounded automation, knowledge retrieval and citizen or enterprise services across defence, government, industrial and educational sectors, particularly in low-connectivity or sensitive environments.

Both firms are among the teams selected under the IndiaAI Innovation Centre (Foundation Models) pillar, which aims to build indigenous AI capabilities and large language models.

The government said access to subsidised national AI compute infrastructure, along with support for ancillary research expenses, enabled the development of the models. It cited this as an example of how public AI infrastructure can speed up homegrown innovation.

EKA-1: Tailored for Indian Languages and Finance

Built for Indian languages, banking and cybersecurity, the EKA-1 has 26.3 billion parameters and was trained on about 7.5 trillion tokens. It covers all 22 scheduled Indian languages along with English. Soket describes it as an open, sparse hybrid language model designed for math, code, reasoning and agentic use.

Interestingly, EKA-1 can be used even without a mobile data or WiFi connection, as it has the capability to do certain inferencing on the local device, say a smartphone.

Soket CEO Abhishek Upperwal said the model has been tailored to India’s banking, financial services and cybersecurity landscape. It is meant to address sector-specific requirements and understand the country’s evolving digital ecosystem, he said.

Dhir Models: Secure, Multilingual, and Offline Solutions

ZenteiQ’s Dhir 2B and Dhir 5B are open-source models that support English and 10 Indian languages. They can be deployed as secure, on-premise multilingual assistants.

According to an official release, the focus is on low-connectivity or sensitive environments.

ZenteiQ CEO Prof. Sashikumar said the models can run on different computing platforms, including GPUs and TPUs. Their lightweight architecture also allows on-device deployment on mobile phones without depending on cloud connectivity, he said.

Users can therefore access AI locally while keeping data on their devices, which improves privacy, security and accessibility.

Strengthening India’s AI Autonomy

Krishnan said such indigenous models strengthen India’s technological and strategic autonomy in artificial intelligence. They also open up opportunities to improve productivity and efficiency across government, industry, research and education, he said.

Platforms such as AIKosh can make the models accessible to researchers, students, innovators and the wider public, encouraging further experimentation, he added.

Disclaimer: News content is sourced from the stated source. Headlines, summaries, section headers, and images are automatically generated or selected using AI/algorithms and may not always be fully accurate. Readers are advised to refer to the full article for complete context.