India's sovereign AI inference — DPDP compliant, zero retention by default.Explore models

Model Requests

Need a model we don't run yet?
Request it.

A dedicated single-model deployment can cost six figures a month — too much for any one team alone. So we pool demand: when enough teams request a model, we stand it up. New models come online fortnightly, first-come first-served.

How it works

Requesting is a signal of intent — the stronger and earlier the demand, the sooner a model ships.

01

Request from the console

Sign in, open Models, and join the pool for the model you want to run. Tell us your rough monthly volume — it takes a minute.

02

Demand pools

Your request joins others asking for the same model. A dedicated deployment is expensive to stand up; pooling demand is what makes it viable.

03

We deploy — fortnightly

We bring new dedicated models online roughly every two weeks, prioritising the models with the most demand behind them.

04

First-come, first-served

Access is granted in the order you requested. Ask early and you hold your place at the front of the line when the model goes live.

Most requested right now

Where demand stands today. Join a pool to push your model up the list.

  1. 01
    GLM-5.2
    Z.ai (Zhipu AI)
    demand
  2. 02
    Kimi K2.6
    Moonshot AI
    demand
  3. 03
    DeepSeek-V4 Pro
    DeepSeek
    demand
  4. 04
    Qwen3-235B-A22B
    Alibaba
    demand
  5. 05
    Llama 4 Maverick
    Meta
    demand
  6. 06
    MiniMax M3
    MiniMax
    demand
  7. 07
    Command A
    Cohere
    demand
  8. 08
    DeepSeek-V4 Flash
    DeepSeek
    demand