Indian AI company Sarvam announced on X and its blog that Sarvam Token Factory is now open to everyone. It serves Sarvam 105B plus GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro, Kimi K3, Kimi K3.1 Pro and Gemma 4 31B from GPU clusters in India through one OpenAI-compatible API. Sarvam says it matches or beats leading global inference providers on speed and cost for most models, with DeepSeek V4.1 Flash launching at an introductory INR 14 per million input tokens and INR 55 per million output tokens.

Key Takeaways

  • ✓One OpenAI-compatible API serves Sarvam 105B, GLM 5.3, DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro, Kimi K3.1 Pro, Gemma 4 31B and more from compute located in India
  • ✓DeepSeek V4.1 Flash has two endpoints: Interactive at about 200 tokens per second per user for coding agents (INR 25 input / INR 105 output per million tokens) and Standard at INR 14 input / INR 55 output
  • ✓Sarvam-measured figures: Sarvam 105B returns its first token in 0.2 seconds at 330 tokens per second for a single user; Gemma 4 31B returns its first token in 0.40 seconds on a 10K-token prompt, 5x faster than the only other provider serving it
  • ✓Kimi K3.1 Pro delivers up to 1.9x tokens per GPU versus InferenceX published results on coding tasks; DeepSeek V4.1 Flash serves 450B+ tokens a day with 95.7% of prompt tokens served from cache
  • ✓Sarvam runs India's largest B200 cluster and offers the same inference stack on-premise for enterprises, pitching data residency and no lock-in
Sarvam Token Factory Opens to Everyone: India-Hosted Open-Weight Inference with One OpenAI-Compatible API for DeepSeek V4.1 Flash, Kimi K3.1 Pro and GLM 5.3
🖼️Official Media
Click to view high-res
🧭

Turn your technical choice into a development budget

Compare 40 dev plans & simulate token costs vs $20/mo subscriptions

🔬

In-Depth Technical Analysis

Background

Open-weight models keep getting stronger, but Indian companies calling them mostly go through overseas inference providers, which raises data residency, latency and cost concerns. Sarvam posted on X, drawing more than 1,700 likes, and announced on its blog that Sarvam Token Factory is now open to everyone.

How it works

Token Factory is Sarvam's own inference stack, serving Sarvam 105B plus GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash 1017, DeepSeek V4 Pro, Kimi K3, Kimi K3.1 Pro and Gemma 4 31B from GPU clusters in India through one OpenAI-compatible API. Sarvam describes a three-tier cache with cache-aware routing and admission, and an orchestrator agent that assigns optimization experiments to sub-agents, with an engineer approving every change before deployment.

Numbers that matter

All figures are Sarvam's own measurements. Sarvam 105B returns its first token in 0.2 seconds with 330 tokens per second for a single user. Gemma 4 31B returns its first token in 0.40 seconds on a 10K-token prompt, 5x faster than the only other provider serving it. Kimi K3.1 Pro delivers up to 1.9x tokens per GPU versus InferenceX published results on coding tasks, and DeepSeek V4.1 Flash serves more than 450B tokens a day with 95.7% of prompt tokens served from cache.

What it means for developers

It is available today to developers and enterprises. DeepSeek V4.1 Flash has an Interactive endpoint at about 200 tokens per second per user for coding agents, priced at INR 25 input and INR 105 output per million tokens, and a Standard endpoint at INR 14 and INR 55. Projects already using an OpenAI SDK can switch by changing the base URL and key, and enterprises can also deploy the stack on-premise.

Action HubReady to adopt this in production?

Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.

Up Nextllama.cpp and ggml Ship v1.0.0: First Major Version in 3.5 Years Adopts Semantic Versioning, With Breaking C API Changes Now Forcing a Major Bump; Release Drops 35 Deprecated Functions and Adds FP8 KV Caches, a MoE Expert Cache and /v1/audio/speech
Read Now→