COMPARISON, read 10 Nov 2026
KEV is a public list anyone can audit; vendor ratings add coverage and speed but cannot be checked from outside
- Maintained by
- KEV: CISA. Vendor ratings: Tenable, Qualys, Nucleus Security
- Scale
- KEV: listed or not. VPR: 0 to 10. TruRisk and Nucleus Threat Rating: none found on 2026-11-10
- Current version
- KEV catalog 2026.11.10. Vendor ratings: not versioned in the pages read
- Updated
- KEV: on each addition. VPR: nightly. Nucleus Threat Rating: daily
- Answers
- KEV: whether exploitation is confirmed. Vendor ratings: the vendor's view of risk or likelihood
What differs
KEV is open data. CISA publishes the catalog as a JSON feed with a catalogVersion, and anyone can recompute a count from it. A vendor rating is an output of the vendor's platform, derived from its own data, and the vendor's pages are the only description. This page labels each statement as the vendor's.
The difference is verifiability. An auditor can check that a CVE is in KEV on a given catalogVersion. An auditor cannot check why a vendor rated a CVE 8.7 on a given night, because the pages read give no formula and no feature list.
| Property | KEV | Tenable VPR | Qualys TruRisk | Nucleus Threat Rating |
|---|---|---|---|---|
| Public per-CVE data | Yes | No | No | No |
| Scale published | Listed or not | 0 to 10 | None found on 2026-11-10 | None found; low to existential |
| Method published | Inclusion criteria | Machine learning and exploit data, no formula | 25+ intelligence sources named as a count | None found |
| Refresh | On addition | Nightly | Not stated in the pages read | Daily |
What vendors add
Vendors cover what KEV does not. Tenable states that VPR is recalculated nightly across more than 280,000 vulnerabilities, so it rates the large part of the population that CISA has not listed. Nucleus states that its rating is refreshed daily for every CVE. Qualys states that TruRisk uses 25 or more sources of real-time threat intelligence. These are the vendors' claims.
They also offer speed. KEV listing came a median 10.5 days after NVD publication in the ledger, and 144 of 318 additions (45.3 percent) within 7 days. A vendor that signals earlier than CISA lists would have value, and the claim is testable.
Qualys adds a specific capability claim: the Enterprise TruRisk Management page describes TruConfirm, exploit validation covering 1,800 or more CVEs, by Qualys's own count. Validation of that kind is evidence of a different type from a rating, since it tests whether an exploit works, and is outside this site's scope except to note that it is a vendor feature.
The claim to test
Nucleus states in a white paper that between October 2025 and March 2026 it reviewed every new KEV addition and found that 22 vulnerabilities, 18 percent, showed meaningful exploitation or exploitability signals before formal listing. By the same count, 82 percent did not. Tenable states that VPR reduced the CVEs requiring action from about 160,000 under CVSS to about 4,000 while maintaining strong exploit coverage, without a definition in the page read.
The ledger gives the means to test such claims: 318 CVEs with listing dates, NVD publication dates and the EPSS and CVSS values on those dates. Ask each vendor for its rating on the ledger's CVEs as of the day before each dateAdded and compute how many sat above its own action threshold. Nothing on this site has done that for any vendor yet.
The ledger stores the CVE identifiers, listing dates and NVD publication dates in data/kev/ledger.json on this site, and a CSV copy of the same rows. A vendor can replicate the test with its own historical ratings in an afternoon, and a buyer can run it during a proof of concept.
What the ledger sets as the benchmark
On listing day the public numbers were these: median EPSS 0.0066, 75.8 percent of 260 scored additions below 0.1, CVSS below 7.0 for 10.4 percent of the 318. A vendor rating that flags most of the 318 before listing would beat both. A rating that flags few would match the public signals.
Because the ledger lists only exploited CVEs, it measures coverage and not false positives. A vendor can reach full coverage by rating everything high. Ask also for the volume the rating sends to the action queue, as Tenable does when it quotes 4,000 against 160,000, and for the dates behind that count.
A related caution concerns dates. A vendor rating exported today for a CVE that was listed last month reflects the vendor's current view, which may already include the KEV listing. The test only works with the rating as it stood before listing, so request the historical value or a dated export, not a present snapshot.
How to use the two together
Keep KEV as the public backstop. Add a vendor rating as a ranking aid for the CVEs not in KEV, with its name and the export date recorded. If a vendor rates a KEV-listed CVE low, KEV wins; FIRST gives the same rule for EPSS.
In a platform migration, rebuild the rating from public inputs first. Kenna, for example, published a 0 to 100 risk score according to the Cisco API documentation, and the Nucleus Threat Rating has no published numeric scale, so no mapping between them can be written from public pages.
Switching from one rating to another is a data project. If the old rating was stored on tickets, those tickets keep the old value. Keep KEV and EPSS on every ticket as well, so the history stays comparable when the vendor changes.
Questions to put to a vendor
Ask for the scale, the update interval, the list of source feeds and the date range of any benchmark claim. Ask whether the per-CVE rating can be exported with its date, so that history can be audited. Ask what the rating does when KEV lists a CVE it rated low.
Record the answers with the date. A vendor statement is a source, and the program should be able to show which source ranked each item.
Verdict by situation
- Federal agencies and contractors: Rank by KEV first because it is the public input BOD 26-04 requires, and treat any vendor rating as supporting evidence recorded with its name and date.
- Large enterprises with ten or more scanners: Rank by KEV first and a vendor rating second for the unlisted population, once the vendor has shown its rating on the ledger's CVEs.
- Managed security providers: Rank by KEV first because it is the one signal common to all clients regardless of which vendor platform each runs.
- Mid-size teams replacing Kenna or a spreadsheet: Rank by KEV first and use the new platform's rating second, since no published mapping carries the old risk score across and KEV needs no migration.
- Application security and infrastructure in one queue: Rank by KEV first for CVE-bearing items and use the vendor rating for code-level findings only where the vendor documents how it scores them.
Other comparisons
Sources
- CISA, KEV catalog JSON feed, read 10 Nov 2026
- Tenable, Vulnerability Priority Rating, read 10 Nov 2026
- Qualys, VMDR, read 10 Nov 2026
- Qualys, Enterprise TruRisk Management, read 10 Nov 2026
- Nucleus Security, Nucleus Insights, read 10 Nov 2026
- FIRST, Using EPSS (KEV guidance), read 10 Nov 2026
- Kenna API documentation (risk score 0 to 100), read 10 Nov 2026
Last reviewed 10 Nov 2026.