---
title: "Sarvam opens India-hosted API serving eight open models"
url: https://www.parallelquant.com/posts/sarvam-opens-india-hosted-api-serving-eight-open-models-b6471f
source_name: "MarkTechPost"
source_url: https://www.marktechpost.com/2026/11/17/sarvam-token-factory-india-inference-platform/
published: 2026-11-17T09:51:18.000Z
topics: ["llms", "open source", "products"]
publisher: "Parallel Quant"
---

# Sarvam opens India-hosted API serving eight open models

*2026-11-17 · Source: [MarkTechPost](https://www.marktechpost.com/2026/11/17/sarvam-token-factory-india-inference-platform/)*

Indian AI company Sarvam has opened Sarvam Token Factory to everyone. It serves open models including Sarvam 105B, GLM 5.3, DeepSeek V4.1 Flash, Kimi K3.1 Pro and Gemma 4 31B from GPU clusters in India through one OpenAI-compatible API.

**Why it matters:** Data-residency rules are pushing countries to want inference that never leaves their borders, and open-weight models make that possible without depending on US labs. Notably, much of the lineup is Chinese open models, showing how those releases are becoming the default substrate for sovereign AI efforts outside the US.

**Topics:** llms, open source, products

---
Read the original: https://www.marktechpost.com/2026/11/17/sarvam-token-factory-india-inference-platform/
Canonical: https://www.parallelquant.com/posts/sarvam-opens-india-hosted-api-serving-eight-open-models-b6471f
Published by Parallel Quant — https://www.parallelquant.com
