The three models have been developed by domestic companies under the IndiaAI Mission and were launched by the Ministry of Electronics and Information Technology (MeitY). The initiative aims to strengthen India’s technological capabilities and provide AI solutions tailored to the country’s requirements.
EKA-1 to Support More Than 100 Languages
EKA-1 has been developed by Bengaluru-based Saket AI. It is a 26-billion-parameter model trained on 7.5 trillion tokens.
The model supports more than 100 Indian and international languages, along with programming languages. It is designed to assist users with mathematical problem-solving, coding and understanding complex tasks.
Key Features of Dhruv 2B and Dhruv 5B
Dhruv 2B and Dhruv 5B have been developed by JenteAIQ. Both models are specifically designed to run on mobile phones without an internet connection.
They support English and 10 Indian languages. The models can also help users retrieve information from documents, answer questions based on document content and locate relevant information.
Google Cloud’s Tensor Processing Units (TPUs) were used in the development of these AI models.
Focus on India’s Technological Self-Reliance
According to S. Krishnan, Secretary of the Ministry of Electronics and Information Technology, indigenous AI models will strengthen India’s strategic and technological self-reliance.
All three models have been made available on the AIKosh platform for researchers, students, startups, businesses and the general public.
The launch is expected to encourage the adoption of AI across several sectors, including education, research, industry and cybersecurity. It also represents an important step towards making AI technology more accessible to users across India, including those with limited internet connectivity.
