Sarvam AI Opens Token Factory To Run Global AI Models On Indian Servers

AI unicorn Sarvam AI has opened Token Factory to developers and enterprises, giving them access to its own and global AI models through a single application programming interface (API). The company runs the models on GPU clusters in India.
Token Factory hosts Sarvam 105B alongside global open-weight models including Google’s Gemma 4 31B, DeepSeek V4.1 Flash, GLM 5.3, and Kimi K3.1 Pro. Developers can switch between supported models through the OpenAI-compatible API without setting up their own computing infrastructure.
The platform charges customers based on the number of tokens they send to a model and the number it generates in response. Sarvam’s catalogue lists Sarvam 105B at ₹15 per Mn input tokens and ₹60 per Mn output tokens. Cached input tokens cost ₹5 per Mn.
Sarvam claimed Token Factory matches or beats leading global inference providers on speed and cost for most of the models it hosts. In the company’s tests, Sarvam 105B returned its first token in 200 milliseconds and generated 330 tokens per second for a single user.
The company said it serves more than 450 Bn tokens a day on DeepSeek V4.1 Flash, with 95.7% of prompt tokens served from cache. The three-tier cache reduces the computation needed when parts of a prompt repeat.
Sarvam first outlined its plans to run its own and global open-weight models on India-based infrastructure at its Sarvam Epoch event earlier this year. It announced general availability for Token Factory earlier this month.
Competition In Local Inference
Krutrim began hosting DeepSeek models on Indian servers in January 2025. OpenAI introduced India-based API inference for eligible GPT-6 Luna workloads on November 12, while Anthropic offers in-country inference for select Claude models through Amazon Bedrock.
In the broader inference market, Sarvam will compete with Together AI and Fireworks AI, which provide managed infrastructure for running open-weight models, and OpenRouter, which gives developers access to hundreds of models through one API. Cloud providers such as Amazon Web Services, Microsoft Azure, and Google Cloud also offer model hosting and inference services.
Sarvam’s pitch to enterprises is that they can switch between its own model and several global open-weight models through one API while keeping prompt processing in India. Sarvam says it runs those models on its own GPUs, so requests stay on its infrastructure without being routed to the model developers’ overseas APIs.
For banks and other financial institutions that need to process certain workloads in India, local hosting could help address data residency requirements. Token Factory also offers dedicated GPU capacity, and Sarvam deploys its inference software on premises for critical institutions.
The post Sarvam AI Opens Token Factory To Run Global AI Models On Indian Servers appeared first on Inc42 Media.


Superadmin 










