Trustgent
Benchmarks

Verified index composition.

Real counts from the live index, how many AI builders are verified in each capability category, where they are headquartered, and how deep the verification goes. No fabricated numbers, no placeholder ranges.

Live · Index composition

What the verified index looks like today.

These benchmarks are computed directly from the live database. Every number is a real count, test accounts and unlisted entries excluded. Recomputed hourly.

Upcoming · Data-gated

Benchmarks that publish once the data is there.

These series require verified pricing, throughput, or outcome records that the index has not yet accumulated in sufficient volume. Each page is honest about the threshold, no fabricated numbers will be published before the sample is large enough.

Methodology

How these numbers are computed.

Composition benchmarks count only public, non-deleted, non-test provider rows. Test accounts (slugs prefixed e2e- or zz-) and unlisted entries are excluded. L2+ means cross-referenced or higher, independently corroborated beyond self-claim.

Verification levels
  • L0Listed, self-submitted, not yet verified
  • L1Claimed, profile claimed by an owner
  • L2Cross-referenced, independently corroborated
  • L3Customer-rated, verified buyer ratings on file
  • L4AI-analyzed, project analysis completed
  • L5Outcome-verified, buyer-corroborated outcomes
For analysts and journalists

Cite Trustgent benchmarks.

Every composition benchmark includes a record-count disclosure, a methodology version, and a computed-at date. Licensed for citation under the published terms.