Open SourceMedium importance
Sarvam opens a “token factory” in India that serves DeepSeek, GLM, Kimi and Gemma on local GPUs, with an OpenAI-compatible API
India’s Sarvam AI launched on 11/16 Sarvam Token Factory, a service that runs open-weight models, its own and third parties’, on GPUs in India, through an OpenAI-compatible API. The initial lineup includes Sarvam 105B, GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash, DeepSeek V4 Pro, Kimi K3.1 Pro and Gemma 4 31B. According to the company, the service already processes more than 450 billion DeepSeek V4.1 Flash tokens a day, requests are not forwarded to the developers’ overseas APIs and data is not retained. The introductory price for Sarvam 105B is ₹10 per million input tokens and ₹25 per million output tokens through the end of 2026. In Sarvam’s own tests, the service is faster than competitors on some models and slower on others, such as DeepSeek V4.1 Flash.
Published in edition no. 35 of Noticiário do Joãozinho ·
Translated by Joãozinho AI from the original Brazilian Portuguese edition. Facts, numbers and sources are the same as in the original; quotes, video titles and source names stay in their original language. Read the original in Portuguese
Original source
https://www.unite.ai/sarvam-opens-india-hosted-token-factory-for-open-weight-models/