New Delhi: At the forthcoming India AI Impact Summit 2026, the government-backed AI consortium BharatGen will present Param2 17B, a 17-billion-parameter multilingual foundation model. The Mixture of Experts (MoE) model has been thoroughly trained on carefully selected Indian datasets and supports 22 Indian languages.
In line with the “Viksit Bharat 2047” vision, Param2 17B is positioned as a significant turning point in India’s drive for autonomous AI capabilities. The concept intends to enable contextual, language-inclusive AI solutions and is intended for implementation across enterprise, governance, healthcare, education, and agriculture.
The effort focuses on language accessibility, real-world adoption, and bolstering India’s domestic AI ecosystem, according to Rishi Bal, CEO of BharatGen. Along with open access channels for developers, startups, and researchers, the summit will include demonstrations with stakeholders from the government and industry.
BharatGen, which is funded by the IndiaAI Mission, has developed sophisticated MoE systems that can reason, calculate, and generate code by utilizing government-backed high-performance computing infrastructure. Additionally, using “Bharat Data Sagar,” a carefully curated research-led data repository, the consortium will present autonomous AI systems in the areas of text, speech, and vision.
Source – Business Today
