IndiaAI-backed Avataar launches Varya video model, claiming four-step generation at ₹0.48 per second · cho.sh
IndiaAI-backed Avataar launches Varya video model, claiming four-step generation at ₹0.48 per second
AI video model releases
IndiaAI-backed Avataar launches Varya video model, claiming four-step generation at ₹0.48 per second
Avataar.ai launched Varya, an IndiaAI Mission-supported distilled video generation model that turns text prompts or uploaded images into video clips and lets users extend them into multi-clip stories.
Varya uses model distillation to reduce generation from more than 50 denoising steps to 4 steps, while Avataar claims comparable output quality to larger teacher models.
Avataar says its internal benchmarks put Varya's inference cost at ₹0.48 per video second, up to 10 times cheaper than several unnamed global video models.
The model is intended to generate visuals grounded in Indian regions, festivals, communities, food, clothing, public spaces, and daily life; cited uses include classroom lessons, MSME advertisements, and public-information videos.
Avataar says it will publish a technical report describing Varya's architecture, distillation method, and benchmarks.
cho.sh