Avataar launched Varya, an indigenous video-generation AI model backed by the IndiaAI Mission, in New Delhi on 12 June 2026, in the presence of MeitY Secretary S. Krishnan.
Varya is a 14-billion-parameter distilled model that generates video from text prompts and images.
Distillation cuts generation from 50 steps to 4, at about ₹0.48 per second of video.
It is designed to understand Indian cultural contexts, for education, small-business advertising, public information and storytelling.
A large 'teacher' model's behaviour is transferred to a faster 'student'. For diffusion video models, the student learns to reach in a few denoising steps what the teacher needs many for, cutting compute and cost for similar output.
Simple Analogy: An expert's 50-step recipe taught as a 4-step shortcut.
Build India's AI ecosystem: compute, indigenous models, datasets, skills, startups and safe AI
Key: Approved 7 March 2024 with ₹10,371.92 crore; seven pillars: compute (10,000+ GPUs), Innovation Centre for foundation models, Datasets Platform, Application Development, FutureSkills, Startup Financing, Safe & Trusted AI
Nodal ministry for the IndiaAI Mission
GS Paper III > Science & Technology > AI and indigenisation
General Awareness > Science & Technology
With the present state of development, Artificial Intelligence can effectively do which of the following? 1. Bring down electricity consumption in industrial units 2. Create meaningful short stories and songs 3. Disease diagnosis 4. Text-to-Speech Conversion 5. Wireless transmission of electrical energy Select the correct answer using the code given below:
Answer: 1, 3 and 4 only
Training a smaller or faster model to reproduce a larger model's outputs.
A generative AI model that produces video clips from written prompts.
A large model trained on broad data and adapted to many tasks.