India Unveils BharatGen: First Sovereign Multilingual, Multimodal AI Model
This integration allows BharatGen to mimic natural communication patterns of India’s multilingual population and ensures inclusive access to AI.
- Country:
- India
Union Minister of State (Independent Charge) for Science & Technology, Earth Sciences, Personnel, Public Grievances, Pensions, Atomic Energy, and Space, Dr. Jitendra Singh, hailed BharatGen as a major leap for India’s technological sovereignty during his visit to IIT Bombay. Positioned as India’s first sovereign, multilingual, multimodal AI-driven Large Language Model (LLM), BharatGen represents a transformative national effort to build an AI system grounded fully in India’s linguistic, cultural, and technological identity.
The Minister reviewed ongoing work on the project, engaged with its core development team, and received a detailed presentation on its architecture, capabilities, and future applications. Senior leaders from IIT Bombay and the Department of Science and Technology (DST) were also present, highlighting the significance of the initiative within India’s larger scientific ecosystem.
A National Mission to Build India’s Own AI Future
Prof. Ganesh Ramakrishnan, Professor-in-Charge of BharatGen, briefed Dr. Singh on the model’s core objectives—developing a sovereign AI ecosystem that captures India’s linguistic richness, reflects India’s lived realities, and supports governance, industry, healthcare, education, and rural development.
Designed to support more than 22 Indian languages, BharatGen integrates three advanced modalities:
-
Text processing and generation
-
Speech recognition and synthesis
-
Document vision, enabling AI to understand Indian-format documents
This integration allows BharatGen to mimic natural communication patterns of India’s multilingual population and ensures inclusive access to AI.
The Minister emphasised that BharatGen aligns with Prime Minister Narendra Modi’s consistent call to develop technologies that are rooted in Indian strengths, address Indian needs, and contribute to global innovation through an Indian lens.
National Collaboration at an Unprecedented Scale
BharatGen is supported under the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS) of DST, with ₹235 crore channelled through the Technology Innovation Hub at IIT Bombay. The consortium includes an extensive network of top institutions:
-
IIT Bombay (Lead Institution)
-
IIT Madras
-
IIT Kanpur
-
IIIT Hyderabad
-
IIT Mandi
-
IIT Hyderabad
-
IIM Indore
-
IIT Kharagpur
-
IIIT Delhi
Dr. Singh observed that this multi-institutional collaboration marks a new era of mission-driven research and strengthens India’s deep-tech development capabilities.
Bharat Data Sagar: India’s Ambitious Data Sovereignty Initiative
A key highlight of the briefing was Bharat Data Sagar, an initiative to ensure India’s complete ownership over its digital knowledge assets. It aims to build one of the largest India-centric datasets through:
-
Large-scale data collection
-
Data curation for linguistic, cultural, and social diversity
-
Collaboration with institutions, state governments, and organisations
This ensures that BharatGen’s outputs reflect real-world Indian contexts and strengthens India’s long-term digital sovereignty.
BharatGen Models Released so Far: Building a Complete AI Stack for India
Dr. Singh reviewed the major models developed under the BharatGen initiative:
1. Param-1 – Foundational Text Model
-
2.9 billion parameters
-
Trained on 7.5 trillion tokens
-
Over one-third Indian content
-
Designed for multilingual understanding, translation, reasoning, and generative tasks
2. Shrutam – Automatic Speech Recognition (ASR)
-
30-million-parameter model
-
Tailored to India’s wide range of accents, dialects, and linguistic patterns
3. Sooktam – Text-to-Speech Model
-
150 million parameters
-
Supports nine Indian languages with natural, human-like speech generation
4. Patram – India’s First Document Vision Model
-
7 billion parameters, trained on 2.5 billion tokens
-
Designed to interpret Indian documents such as certificates, government forms, and bills
Together, these models form India’s first complete indigenous AI stack—text, speech, and vision—capable of powering a wide range of national applications.
Demonstrated Use-Cases: AI for Citizens, Farmers, and Small Businesses
The BharatGen team demonstrated real-world applications built using the models:
Krishi Sathi
A voice-enabled WhatsApp advisory tool allowing farmers to:
-
Ask questions in their local language
-
Receive instant, accurate responses
-
Access agricultural guidance with no literacy barrier
e-VikrAI
A tool enabling small sellers to upload a product image and automatically generate:
-
Product descriptions
-
Keywords
-
Listings for e-commerce platforms
This supports MSMEs and rural entrepreneurs entering digital marketplaces.
Docbodh
A document Q&A platform using Patram to help citizens understand:
-
Legal documents
-
Government circulars
-
Policies and certificates
Dr. Singh noted that these applications show how BharatGen can empower ordinary citizens and make public services more accessible.
Collaboration with Industry, Ministries, and State Governments
The team highlighted deepening partnerships with:
-
IBM
-
Zoho
-
NASSCOM
-
Ministries including Water and Sanitation (WASH)
-
State governments such as Maharashtra
These partnerships bring domain expertise, sector-specific data, and deployment opportunities—critical for scaling BharatGen into a national AI ecosystem.
Expanded Government Support Under the India AI Mission
The project recently received ₹1,058 crore from the Ministry of Electronics and IT (MeitY) under the India AI Mission. Combined with DST support, this funding signals the Government’s commitment to:
-
Building India’s sovereign AI stack
-
Strengthening domestic research capacity
-
Developing globally competitive AI models and infrastructure
Dr. Singh said this reflects India’s readiness to drive global digital transformation and lead in frontier technologies including quantum, cyber-physical systems, and space sciences.
BharatGen: A Turning Point in India’s Technological Self-Reliance
Dr. Jitendra Singh described BharatGen as a historic milestone in India’s digital journey. He emphasised:
-
BharatGen is not merely a technological project—it is a national mission.
-
The model ensures India’s future AI reflects Indian languages, values, and lived experiences.
-
It embodies the Prime Minister’s vision of “Technology for People, by People, and of People.”
He encouraged the team to continue developing AI systems that are:
-
Globally benchmarked
-
Uniquely Indian
-
Scalable and accessible
-
Meaningful and inclusive
Paving the Way for India’s Digital Decade
Closing his address, Dr. Singh said BharatGen will play a defining role in shaping India’s digital decade, creating AI solutions that are:
-
Powerful enough for global deployment
-
Inclusive enough for last-mile delivery
-
Secure enough for national sovereignty
-
Transformative enough to uplift every citizen
With BharatGen, India is poised to emerge as a global contributor and leader in AI, while ensuring its digital future remains firmly rooted in India’s cultural and linguistic identity.

