Bodhan AI Launches Four Open AI Models for Multilingual Education in India

Bodhan AI, the Centre of Excellence in AI for Education incubated at IIT Madras, has launched four open AI models designed to support multilingual education and AI applications across India.

The models are being released as Digital Public Goods, allowing educational institutions, startups and researchers to build and customise applications without having to develop AI models from scratch.

Four AI Models Support Indian Languages

Developed in partnership with AI4Bharat, the new models focus on four key areas:

- Speech recognition

- Speech generation

- Machine translation

- Optical character recognition (OCR)

The models support between 22 and 27 Indian languages and are available through open-weight releases and APIs hosted on sovereign digital infrastructure. This approach is intended to make advanced AI capabilities more accessible to organisations developing applications for India`s diverse linguistic ecosystem.

The models form part of the Bharat EduAI Stack, which is envisioned as sovereign digital public infrastructure for education.

AI Tools for Students and Teachers

Alongside the four AI models, Bodhan AI has also launched the Student TutorBot and Teacher Assistant Bot.

The TutorBot is designed for students in Classes 6 to 12 and supports text and voice interactions across 22 Indian languages. It uses textbooks to provide explanations, practice activities and assessments to help students with their learning.

The Teacher Assistant Bot is designed to support educators with tasks such as preparing lesson plans, worksheets, quizzes and assessments. Teachers remain in control and can review, modify or discard AI-generated content before using it.

Building AI for India`s Education Ecosystem

IIT Madras Director V. Kamakoti highlighted the importance of developing AI technology that understands India`s linguistic and educational needs.

Bodhan AI Principal Investigator Mitesh Khapra said the initiative is intended to build alongside the wider ecosystem rather than compete with it, creating a common AI layer and reducing duplication in AI development.

The models were trained and optimised using NVIDIA Nemotron open models and are expected to support a wide range of AI applications across India`s multilingual education landscape.

The launch marks another step towards building open, sovereign and multilingual AI infrastructure for education in India, with institutions, researchers and startups able to adadapt the technology for their own applications.