Back

India’s biggest AI drop so far 🇮🇳 Introducing Sarvam-30B and Sarvam-105B frontier LLMs built from India, for India. Sarvam-30B • 30B parameters (MoE) • 1B active params/token • 32K context window • Trained on 16T tokens • Competitive with Gemma-

Indian AI launch: Sarvam-M, a 24B open-weights hybrid model built on top of Mistral Small. Sarvam-M achieves a new benchmark across a range of Indian languages, math, and programming tasks, for a model of its size. Here is a detailed technical blog

Last week: “Sarvam is problematic!!” This week: Sarvam is chillin’ right under DeepSeek on Hugging Face like👀 “sooo... y’all done yelling?” Anyway, we climbing. 🧗‍♂️✨ #SarvamAI #PlotTwist

sarvam just announced 30B and 105B models! 30B: + pre-trained on 16T tokens + context length of 32k tokens. 150B: + context length of 128k tokens + on-par with most other open closed source ones of its class Sarvam Chat app will be available fro

This is a huge leap for the Indian AI ecosystem! Sarvam-M being open-weight and built on top of Mistral is super exciting—especially for those of us working in Indic NLP and multilingual apps. The focus on coding, math, and 10 major Indian language

Excited to witness the momentum of innovation at the Bharat AI Summit 2026. One of the standout moments was Sarvam's launch of their AI glasses.. SARVAM KAZE, a step towards bringing everyday AI closer to people and transforming how we interact with

l got to know abt 5 made in India AI Products in AI Impact Summit that BLEW my mind: 1. EkaScribe: A doctor in a busy rural clinic see 5 patients without touching a keyboard. The AI handles the prescriptions, history, and filing. 2. Ottobots: Made

India's biggest AI startup, $1B Sarvam, just launched its flagship LLM. It's a 24B Mistral small post trained on Indic data with a mere 23 downloads 2 days after launch. In contrast, 2 Korean college trained an open-source model that did ~200k last

What's happening to Sarvam has happened to so many consumer apps that claimed they were made for Bharat but didn't have a better use case than global counterparts even after great funding. The world doesn't care about the Bharat sentiment, they care