Sampled filings retaining a validated public USAC form link.
60 of 60Reproducible product verification
Claims you can inspect, not marketing math.
BidTerrain publishes the sample rule before calculating results, links every sampled filing back to USAC, and withholds headline metrics when the coverage gate is not met.
Latest reproducible run
What the benchmark actually measured.
Funding year 2027 · 190 eligible official rows · deterministic seed retained in the result file.
Integrity results are published as achieved metrics. Synthetic software probes and runtime checks are labeled Diagnostic. Service agreement remains Exploratory until at least 30 field-eligible filings are available.
Mapped IDs exactly equal the sampled official dataset IDs.
60 of 60Available posting, response-deadline, and allowable-start dates preserved exactly as ISO dates.
180 of 180Official documents producing supported text processing: parsed, partial, or reviewed with no structured evidence. 60 of 60 document outcomes were inspected.
55 of 60Deterministic positive same-state and negative different-state statewide probes while holding one explicit supported service constant. This is a software invariant, not field-measured precision.
100% observed · 60 of 60 probesDeterministic positive same-state and negative different-state statewide probes while holding one explicit supported service constant. This is a software invariant, not field-measured recall.
100% observed · 60 of 60 probesObserved agreement on 4 filings with available official request/service-description fields. At least 30 field-eligible filings are required before this becomes an achieved metric; it is not serviceability or human adjudication.
100% observed · 8 of 8 probesObserved agreement on 4 filings with available official request/service-description fields. At least 30 field-eligible filings are required before this becomes an achieved metric; generic language may correctly abstain.
100% observed · 8 of 8 probesRuntime diagnostic for the deterministic report-contract builder. It excludes network transfer, browser rendering, and PDF export.
0.22 ms observedRuntime diagnostic for the deterministic report-contract builder. It excludes network transfer, browser rendering, and PDF export.
0.47 ms observedDeclared before scoring
A non-cherry-picked sample rule.
- 01Freeze the universe
Use every eligible record in the latest funding year returned by USAC’s open dataset at the run time.
- 02Apply the fixed seed
Deduplicate application IDs, order states and filings with the published seed, and select 60 before calculating any metric.
- 03Enforce coverage
Require at least 10 states, multiple structured service categories, both forms, and 50 inspected official documents. Service agreement needs 30 field-eligible filings before it is promoted from Exploratory.
- 04Fail visibly
Retain unsupported, image-only, failed, and no-text document outcomes. They lower measured extraction coverage instead of disappearing.
Official-source audit trail
Sampled records stay inspectable.
First 12 shown here. The downloaded result contains the complete sample and document outcome for every filing.
voice · brief 1.78 ms
Verify official form ↗other · brief 0.38 ms
Verify official form ↗other · brief 0.34 ms
Verify official form ↗other · brief 0.27 ms
Verify official form ↗other · brief 0.26 ms
Verify official form ↗other · brief 0.3 ms
Verify official form ↗other · brief 0.25 ms
Verify official form ↗other · brief 0.64 ms
Verify official form ↗other · brief 0.47 ms
Verify official form ↗other · brief 0.26 ms
Verify official form ↗other · brief 0.23 ms
Verify official form ↗other · brief 0.22 ms
Verify official form ↗Different jobs, one source of truth
USAC record access versus BidTerrain workflow.
BidTerrain does not replace USAC or the provider’s judgment. It reduces the work between discovery and a defensible review decision.
Search and filter the official Form 461/465 public dataset.
Screens the current official feed against submitted territory and service criteria.
Open the official form and each applicant attachment manually.
Extracts supported requirements with document status and source provenance; unreadable files remain flagged.
Provider applies its own network, commercial, and timing criteria.
Produces an explainable screening brief with separate territory, service, timing, and evidence findings.
The provider verifies the controlling sources and submits through the required process.
Keeps official-source verification and provider engineering, pricing, compliance, and submission decisions in human hands.
Known limits
What these results do not prove.
- Structured-field service agreement uses 4 of 60 sampled filings because the remaining rows do not expose request/service-description fields in the public dataset. It is not a human-adjudicated network-serviceability study.
- Document extraction is measured on one deterministically selected official form or supporting attachment per filing. Image-only files are not OCRed.
- Report-generation timing measures the deterministic decision-brief builder in the benchmark runtime; it excludes network transfer, browser rendering, and PDF export.
- A passing territory or service screen is not an engineering, compliance, eligibility, pricing, or award determination.
- Official USAC records and applicant documents remain controlling and can change after the stated as-of time.
Use your real criteria
Run a source-linked territory screen.
The benchmark tests the product framework. Your report still depends on the territory and service capabilities you submit.