Start free. Pay when it gets serious.

Three ways to work with us, in the order most clients take them.

free

Shortlist

Tell us your task mix, volumes, and constraints on a one-page form. We answer from the standing grid: a shortlist of models and endpoints worth your time, with the reasoning.

  • No calls required, no strings
  • Answered from live grid data
  • This is how most clients meet us
Request a shortlist
fixed scope

Bake-off

The decision, made properly. Everything in Shortlist, plus mprobe on your own sample tasks.

  • Named endpoints, reproducible cost calculations
  • Latency and concurrency testing
  • Quantization guidance from qprobe
  • A go/no-go matrix you can defend internally
  • Fixed scope and price, agreed before we start
  • One 30-minute call to set it up
Start a bake-off
ongoing

Enterprise

Stay right, continuously. For teams running open models in production.

  • gprobe monitoring configured to your stack
  • Scheduled re-evaluations as the ecosystem moves
  • Direct access when a release or a price move demands a same-week answer
  • Scoped to you
Talk to us

Where our grid coverage is thin for your case, the report says so explicitly. We would rather tell you what we do not know than sell you certainty we do not have.