Pick a bundle. Pay once. Get a receipt.

Every bundle is one OCA run against private, held-out challenges. You submit your model endpoint, we score and seal the result. Volume discount: 10+ runs in a calendar month → 20% off.

Single Probe

1 module

$50/run

Any one module (RFNN, HVM, QRS, CLA, PDN)
Up to 8 challenges per run
Signed receipt + module breakdown
HTTP eval against your model endpoint

Reasoning Stack

6 modules · causal + memory + reasoning

$200/run

RFNN, ICA, QRS, CPR, DSS, RIG
Up to 24 challenges per run
Per-module mean + std + min/max
Signed receipt with HMAC-SHA256

Frontier Audit

14 modules · release-grade

$500/run

RFNN, HVM, ICA, NSCM, GWN, QRS, TGNN, UME, CMT, DSS, HCP, RIG, EQR, CLA
Up to 56 challenges per run
Suitable as a release-gate result for a model card
Optional citation appendix for the paper

Full Vault

All 22 modules · comprehensive

$1,200/run

Every cognitive axis we score
Up to 88 challenges per run
Includes meta-skill curation + self-reflection
Priority worker queue

Volume discount

Run 10 or more bundles in a calendar month and we automatically apply a 20% discount on each. The discount stacks across customers within the same organization; ask for an invoice at the end of the month.