IIT Bombay's BharatGen Unveils 17B Multilingual AI Model at India AI Summit

• IIT Bombay-led BharatGen launched its 17-billion-parameter multilingual foundational model, BharatGen Param2 17B MoE, at the India AI Impact Summit 2026.
• Backed by IndiaAI Mission and DST, Param2 17B is optimized for multiple Indic languages to boost AI adoption in governance, education, healthcare, agriculture, and enterprise.
• The model's training pipeline was developed with NVIDIA, utilizing NVIDIA AI Enterprise software and infrastructure, including NeMo open libraries.
• BharatGen aims to develop India-first sovereign AI systems reflecting linguistic and cultural diversity, reducing reliance on foreign models and accelerating domestic AI capability.
• The model will be open-sourced via Hugging Face to encourage ecosystem participation, allowing developers, startups, and enterprises to build India-centric AI applications.