This New India-Built AI Model Wants to Make Video Generation Far More Affordable

Avataar.ai has launched Varya, an indigenous AI video generation model developed with support from the IndiaAI Mission, claiming significant improvements in efficiency and affordability over existing video generation systems.

The company says Varya uses a distilled architecture that reduces the video generation process from 50 steps to just four while maintaining comparable output quality. According to Avataar's internal benchmarks, the model can generate video at ₹0.48 per second, making it up to 10 times more cost-efficient than several leading global video models.

The launch was announced at an event in New Delhi attended by S. Krishnan, Secretary, Ministry of Electronics and Information Technology (MeitY), and comes amid India's broader push to develop homegrown AI capabilities under the IndiaAI Mission.

Focus on Cost and Indian Context

Unlike many frontier video models developed overseas, Varya has been designed with an emphasis on Indian contexts and affordability. According to the company, the model has been trained to better understand and generate visuals reflecting India's regional diversity, festivals, communities, food, clothing, and everyday environments. The intended applications span education, commerce, citizen services, MSMEs and content creation.

The company argues that existing video generation models remain expensive to operate while often underrepresenting India's cultural and linguistic diversity, creating an opportunity for locally developed alternatives.

Built With Support From the IndiaAI Mission

Avataar was among the companies selected under the IndiaAI Mission to build indigenous foundational AI capabilities. According to the company, access to subsidised national AI compute infrastructure enabled the research behind Varya and demonstrates how public AI infrastructure can accelerate domestic innovation.

The model is built around a workflow that allows users to type a prompt or upload an image, generate a video and continue building longer visual narratives from additional clips. Avataar describes the experience as "Idea → Video → Story".

The company also plans to publish a technical report detailing Varya's architecture, distillation methodology and performance benchmarks.

Efficiency May Matter More Than Scale

The bigger takeaway from Varya's launch is not necessarily that India has another AI model, but that it reflects a different approach to competing in the global AI race.

Rather than chasing ever-larger models, Avataar is betting that lower inference costs and context-aware performance could make AI video generation practical at a population scale. As video becomes an increasingly important medium for education, commerce, and public services, reducing the cost of generation could prove as important as improving model quality itself.