BharatGen to Unveil 17 Billion AI Model in 22 Indic Languages at India AI Impact Summit

BharatGen will unveil a 17 billion parameter sovereign multilingual AI model built to work across 22 Indian languages at the upcoming India AI Impact Summit 2026, positioning it as a cornerstone of India’s push for AI self reliance.

The model, called Param2 17B, has been developed from the ground up using a Mixture of Experts architecture and trained extensively on Indian data. 

BharatGen says the system will go live at the summit and will be opened for developers, startups, researchers, and enterprises to build applications across governance, healthcare, education, finance, and other public service sectors.

Backed by the IndiaAI Mission, the initiative has access to government supported high performance computing infrastructure and a large research dataset repository called Bharat Data Sagar. The company claims the new model supports reasoning, math, and code tasks alongside multilingual text, speech, and vision capabilities.

Read: BharatGen’s ‘Recipe’ for Building a Trillion Parameters Indic Model

“These models are designed for real deployment across citizen services, finance, healthcare, and education, while strengthening India’s sovereign AI capability,” Rishi Bal, CEO of BharatGen said. “The summit showcase will provide fast go-to-market pathways, along with open access for developers, startups, researchers, and enterprises building AI solutions for India and beyond.”

At India AI Impact Summit, BharatGen will demonstrate sector specific use cases built with state and central government departments and industry partners.

In governance, a system called MahaGPT has been developed with the Government of Maharashtra and MITRA to improve efficiency and accuracy in complex urban development and revenue functions. 

Separate pilots include multimodal citizen access tools for the Department of Water and Sanitation under the central government and an AI led digital transformation effort with Goa Electronics Limited for the Government of Goa.

In healthcare, Medsum, built by Mata Amrita Technologies, uses BharatGen’s models to connect doctors and patients through structured health information and conversational interfaces.

In education, the Kotak Education Foundation is deploying a tool that evaluates spoken English fluency and communication skills, generating structured scores and performance insights for learners.

Cultural preservation is another focus. Under Gyan Bharatam, a digitisation drive by the Ministry of Culture, BharatGen’s OCR and conversational AI stack will help read and digitise ancient manuscripts across 22 languages, with plans to scale to an archive of 50 crore documents. 

Apart from BharatGen, 11 other firms are expected to announce their foundational models at the India AI Impact Summit. 

One of them, Sarvam AI, has been on an announcement spree for the last week where it has launched several AI models. Most recently, the company launched Saaras V3, its speech recognition model, particularly for mixed-language and noisy speech. The model also expanded support for all the 22 scheduled languages of India.