Union minister of state for commerce & industry and electronics & IT Jitin Prasada launched BharatGPT Mini, a compact small language model (SLM) built by Indian AI startup CoRover, at the VivaTech 2025 conference in Paris.

Unlike larger, cloud-dependent language models, CoRover’s SLM is built to operate offline, on-device and in multiple Indian languages. With 534 Mn parameters, BharatGPT Mini is optimised for use in areas with limited internet access or computing power.

BharatGPT supports 14 Indic languages and is trained on CoRover’s proprietary conversational datasets. The model is built in India and is optimised for edge and server implementations, allowing seamless AI performance on low-end devices.

As cloud infrastructure remains unevenly distributed and data privacy concerns rise in AI usage, smaller, task-specific models are being seen as more practical alternatives for public and enterprise use.

Talking about the impact that BharatGPT Mini can have on Indian businesses and citizens, CoRover CEO Ankush Sabharwal said, “Bharat GPT Mini can work offline, making it perfect for businesses and citizens alike. Citizens can now access healthcare, education, government services, banking, and more in 14 Indian languages, simply by speaking.”

He said that large language models (LLMs) that rely on heavy infrastructure often miss the mark in India’s diverse and decentralised ecosystem. A light-weight model that runs offline can actually help those who need AI, he added.

“It is a milestone in democratising access to intelligent conversational AI agents,” Sabharwal said.

The model is currently available in a limited release via CoRover Builder, a no-code platform that allows users to build and deploy AI assistants for web, telephony, and IoT systems.

BharatGPT Mini supports text, voice and video inputs, and is intended for use in sectors like healthcare, education, banking, and government services.

Founded by Sabharwal, Kunal Bhakhri, and Manav Gandotra, CoRover is an AI startup that provides a conversational Gen AI platform for creating human-centric, multilingual, multimodal and multichannel AI agents, copilots, and assistants.

The startup raised $4 Mn (about INR 33.4 Cr) in a funding round led by Venture Catalysts last year.

It operates in the rapidly growing Indian AI market, which is set to cross the $17 Bn mark by 2030. BharatGPT Mini could become a key part of this growth. As the market demand shifts towards smaller models that can operate securely and privately without external data dependencies, enterprises are seeking privacy-focused AI tools.

While LLMs remain important for complex, general-purpose tasks, SLMs like BharatGPT Mini are gaining traction for their adaptability, faster deployment, and better performance on low-end hardware.

According to CoRover, integrating BharatGPT Mini can make project pipelines 60% to 70% more efficient. The AI startup also estimates a fivefold increase in SLM adoption by small and mid-sized businesses by FY26.

The post CoRover Unveils BharatGPT Mini, India’s First Offline Multilingual AI Model appeared first on Inc42 Media.