AI workloads / Public reference

Model token consumption

A supplied tokenCount or tokens value of 100,000 or more.

What you need to supply

Telemetry you include. The check does not call a model provider.

If it reports a finding

Medium below 1,000,000 tokens, and high at or above that. The suggested next step is to review prompt size, routing, caching, and per-request limits.

What this does not prove

It does not decide whether the spend was worth it.

Telemetry field

{ "tokenCount": 100000 }

All review checks. Reading this page does not submit your infrastructure.