BharatGen Marks India’s First Sovereign Multilingual Large Language Model Push

Congratulating the BharatGen team, Dr. Singh described the initiative as a landmark in India’s technological self-reliance journey.

Lauding India’s first government-owned sovereign Large Language Model (LLM) and multilingual AI stack, Union Minister of State (Independent Charge) for Science & Technology, Dr. Jitendra Singh, today said Artificial Intelligence is no longer optional but essential across every sphere of life.

He was addressing the session titled “BharatGen Models: Vision and Technical Execution 2026” at Bharat Mandapam, New Delhi, as part of the Global AI Summit.

The session was hosted by BharatGen in association with the India AI Mission (MeitY) and the Department of Science and Technology (DST).

---

Government-Owned, Sovereign and Multilingual

Congratulating the BharatGen team, Dr. Singh described the initiative as a landmark in India’s technological self-reliance journey.

He highlighted that while LLMs are globally prevalent, BharatGen’s distinguishing feature is its sovereign, government-supported character, tailored to India’s socio-cultural and linguistic diversity.

“This represents a proactive policy intervention at an early stage of technological evolution,” he noted.

BharatGen is positioned as a multilingual and multimodal foundation AI stack, designed for India-centric applications across governance, healthcare, agriculture, education and legal systems.

---

₹10,820+ Crore Ecosystem Backing AI Sovereignty

BharatGen operates under a consortium led by the TIH Foundation for IoT and IoE at IIT Bombay, with participation from:

• IIIT Hyderabad

• IIT Hyderabad

• IIT Mandi

• IIT Kanpur

• IIT Madras

• IIM Indore

The initiative is supported by:

• ₹235 crore under DST’s National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS)

• ₹10,585 crore under the India AI Mission (MeitY)

Dr. Singh described this as a “whole-of-science, whole-of-government and whole-of-nation” model.

---

Technical Depth: Text, Speech and Document Intelligence

The Minister outlined BharatGen’s multi-modal capabilities:

• Text-based Large Language Models

• Speech technologies (speech-to-text and text-to-speech)

• Vision-language models for complex documents

He noted that BharatGen has already released domain-specific fine-tuned models:

• Ayur Param (Ayurveda)

• Agri Param (Agriculture)

• Legal Param (Indian legal domain)

Recently launched models include:

• Param-2 text foundation model (17 billion parameters) in 22 scheduled Indian languages

• Shrutam speech-to-text models in 12 Indian languages

• Sooktam text-to-speech models in 12 languages

• Patram models under the DocBodh framework for multilingual document intelligence

---

Expanding Beyond 22 Scheduled Languages

Dr. Singh emphasized that India’s linguistic diversity extends beyond the 22 scheduled languages and stressed the need to continuously expand datasets to cover regional languages and dialects.

“Scientific advancement cannot be confined by administrative boundaries,” he said, particularly highlighting applications in digital health and citizen services.

---

Institutional Scale and Model Sovereignty

The initiative has transitioned into a Section-8 company — BharatGen Technology Foundation, enabling national-scale operations while ensuring data and model sovereignty through platforms such as Bharat Data Sagar.

Dr. Singh underscored the importance of collaborative ecosystems, saying emerging technologies demand partnership across academia, industry and government rather than siloed approaches.

---

National AI Infrastructure for Viksit Bharat

The session also featured technical presentations from BharatGen leadership and academic collaborators, along with addresses by senior policymakers including the Secretaries of MeitY and DST, and the Principal Scientific Adviser.

An MoU exchange between BharatGen and the IIT Bombay Heritage Foundation was also witnessed by the Minister.

Concluding his remarks, Dr. Jitendra Singh said BharatGen represents a new phase in India’s AI journey — combining sovereign capability, collaborative execution and inclusive design — and will play a central role in building robust national AI infrastructure aligned with the vision of Viksit Bharat.