Sarvam launches Token Factory inference stack in India
Telborg
· November 16, 2026
· ✓ verified
Sarvam has announced the launch of Sarvam Token Factory, its inference stack for serving its own models and leading open-weight models from GPU clusters located in India.
- The launch covers Sarvam 105B, GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro, Kimi K3.1 Pro, and Gemma 4 31B through one OpenAI-compatible API.
- Pricing and performance details include ₹14 per million input tokens and ₹55 per million output tokens for DeepSeek V4.1 Flash introductory pricing, plus latency/throughput claims such as 0.40 seconds first token for Gemma 4 31B and 0.2 seconds first token for Sarvam 105B.
- The post also says Sarvam runs India’s largest B200 cluster, offers the stack on-premise for critical institutions, and is making Token Factory available now to developers and enterprises.