Sarvam AI is a Bengaluru-based startup that built Sarvam-30B and Sarvam-105B, open-source AI models trained entirely in India — pretraining, fine-tuning, and reinforcement learning all ran on domestic compute under the government's IndiaAI Mission.

What makes Sarvam AI different from ChatGPT or Gemini?

Sarvam's models are built specifically to support all 22 officially scheduled Indian languages, including Hindi, Tamil, Telugu, Bengali, Marathi, and Kannada, plus code-mixed inputs like Hinglish — and they are fully open-source, so anyone can download and use them.

Can I use Sarvam AI right now?

The models are released as open-weights on platforms like Hugging Face, so developers can download and run them. Sarvam AI also offers Samvaad, its own conversational agent platform built on these models, for less technical users.

Why does this matter for India?

It's one of the first large AI models built end-to-end in India — a step toward reducing dependence on foreign AI companies for core AI infrastructure, and toward better AI support for Indian languages that global models often serve less well.

Published September 5, 2026 · 4 min read

Sarvam AI: India's Own Open-Source AI Model, Explained

Most AI tools people use in India — ChatGPT, Gemini, Claude — are built by American companies. Sarvam AI, a Bengaluru startup, built something different: Sarvam-30B and Sarvam-105B, open-source AI models trained entirely in India and designed around Indian languages from the ground up. Here's what that actually means.

See it running

Video: Fahd Mirza — "India Enters the AI Race: Running Sarvam-30B Locally."

Built entirely in India

Sarvam-30B and Sarvam-105B were unveiled at the India AI Impact Summit in New Delhi and officially released as open-source on March 6, 2026. What sets them apart isn't just where the company is based — the entire training pipeline (pretraining, fine-tuning, and reinforcement learning) ran on domestic compute, supported by the government's ₹10,372-crore IndiaAI Mission, with infrastructure from Yotta and technical support from Nvidia.

Why the language support matters

The models support all 22 officially scheduled Indian languages — Hindi, Tamil, Telugu, Bengali, Marathi, Kannada, and more — plus code-mixed inputs like Hinglish, which most global AI models handle less naturally. This is a genuine opening for exactly the kind of regional-language content and tools this audience needs, well beyond what most Western AI products prioritize.

Is this something you can actually use?

If you're a developer, the models are open-weights on Hugging Face — you can download and run them yourself, or fine-tune them. If you're not technical, Sarvam AI also runs Samvaad, its own conversational agent platform built on these models, aimed at everyday users rather than developers.

Frequently asked questions

What is Sarvam AI?

A Bengaluru startup that built Sarvam-30B and Sarvam-105B, open-source AI models trained entirely on Indian infrastructure.

How is it different from ChatGPT or Gemini?

Built specifically for all 22 official Indian languages plus Hinglish, and fully open-source.

Can I use it now?

Developers can download the open-weight models from Hugging Face; everyday users can try Samvaad, Sarvam's own agent platform.

Why does this matter?

It's a step toward India having its own core AI infrastructure and better support for Indian languages.

Sources: Sarvam AI's official blog, Open Source For You, ET Bharat, and Fahd Mirza's YouTube channel (2026).