India Launches PARAM-2 AI, a Sovereign Model for 22 Languages

Developed under the BharatGen programme, PARAM-2 represents India's push to build its own sovereign foundational AI models designed specifically for the country's languages, governance needs and cultural realities.

BharatGen is India's national generative AI initiative, supported by the Department of Science and Technology (DST).

The programme has quietly built a strong foundation over the past few years, helping India join a small group of nations capable of developing their own large-scale AI models.

PARAM-2 supports all 22 Scheduled Indian languages and has been trained on India-focused datasets under Bharat Data Sagar.

It uses a Mixture-of-Experts architecture, allowing it to handle complex multilingual tasks efficiently.

Professor Ganesh Ramakrishnan of IIT Bombay, one of the key leaders behind BharatGen, described the launch as more than just a new model release.

He said it represents the coming together of researchers, institutions, government bodies and industry partners to ensure that India can shape its own AI future.

Unlike global consumer AI platforms such as ChatGPT or Gemini, BharatGen follows a different model.

It does not operate as a centralised business-to-consumer service. Instead, its AI models are released as national public digital goods.

This allows government departments, banks, hospitals, courts and educational institutions to deploy them locally, even in secure environments without internet access. The goal is to ensure transparency, trust and data sovereignty.

The backbone of BharatGen is the National Mission on Interdisciplinary Cyber-Physical Systems under DST.

The mission initially approved Rs 235 crore to seed the project. It is now being scaled further with support from the IndiaAI Mission under the Ministry of Electronics and Information Technology, which has allocated Rs 900 crore as part of a broader national AI strategy.

- IANS